Compare commits

..
1 Commits
Author SHA1 Message Date
nobody d02373e837 This commit was manufactured by cvs2svn to create tag 'MCS_0_6'.
svn path=/tags/MCS_0_6/mcs/; revision=574
2001-08-22 23:32:44 +00:00
1859 changed files with 33160 additions and 208818 deletions

No files matched your search

-155
View File
@@ -1,158 +1,3 @@
2002-02-05 Dietmar Maurer <[email protected]>
* mono/tests/enum2.cs: new file with more enum tests.
2002-02-01 Dietmar Maurer <[email protected]>
* mono/tests/jit-long.cs: added some SHL/SHR tests
Fri Feb 1 15:32:36 CET 2002 Paolo Molaro <[email protected]>
* configure.in: allow PKG_CONFIG_PATH for the build-platform
pkg-config invocation.
Fri Feb 1 15:13:25 CET 2002 Paolo Molaro <[email protected]>
* configure.in: add some support/fixes for cross-compilation.
2002-01-14 Dietmar Maurer <[email protected]>
* mono/tests/exception6.cs: new tests for ADD_OVF, MUL_OVF
added tests for SUB_OVF, SUB_OVF_UN
2001-12-21 Dietmar Maurer <[email protected]>
* mono/tests/Makefile.am (TESTSRC): disabled stream-writer test
2001-12-21 Aaron Weber <[email protected]>
* doc/faq (A): added a linebreak that was missing and was causing
a heading to appear badly.
2001-12-20 Aaron Weber <[email protected]>
* doc/faq: added <a name=""> nav links, and moved all MSFT-related
questions into a single section.
2001-12-17 Aaron Weber <[email protected]>
* doc/faq (Q): Question-by-question, line-by-line revision.
2001-12-17 Dietmar Maurer <[email protected]>
* mono/tests/time.cs: new test
2001-11-30 Radek Doulik <[email protected]>
* configure.in: remove ACCESS_UNALIGNED="no" from ppc
2001-11-29 Radek Doulik <[email protected]>
* configure.in (ACCESS_UNALIGNED): uncommented and modified
powerpc-*-linux*
generate mono/arch/ppc/Makefile
2001-11-26 Dietmar Maurer <[email protected]>
* mono/tests/exception3.cs: modified the test to show a bug in the
current exception handling implementation.
2001-11-13 Dick Porter <[email protected]>
* configure.in: Some more tests for pthread features
(specifically, glibc 2.1 doesnt define pthread_mutex_timedlock but
still needs _GNU_SOURCE for PTHREAD_MUTEX_RECURSIVE)
2001-11-09 Dick Porter <[email protected]>
* configure.in: Try and get large file support, but it's not fatal
if it's not there (the io-layer just ignores the high word in that
case.)
Check for windows builds, and only bother to look for pthreads and
large files if we're not building for cygwin or native win32. Not
having pthread support on Posix systems is now a fatal error.
* autogen.sh: Fix ACLOCAL_FLAGS - it wasn't being passed to
aclocal
2001-10-15 Dietmar Maurer <[email protected]>
* mono/tests/Makefile.am: removed JITTESTS, simply run all tests for
target testjit.
2001-10-11 Dietmar Maurer <[email protected]>
* docs/object-layout: more documentation
Mon Oct 8 20:27:50 CEST 2001 Paolo Molaro <[email protected]>
* configure.in: define NO_UNALIGNED_ACCESS for platforms that
can't read on unaligned boundaries
2001-10-04 Dick Porter <[email protected]>
* configure.in: Do some deeper checks on pthreads for some of the
more interesting functions
2001-09-24 Dietmar Maurer <[email protected]>
* mono/tests/Makefile.am (testjit): a new target to test the JITer
* mono/tests/jit-*: added some test for the JITer
2001-09-23 Dick Porter <[email protected]>
* configure.in: Arrange to compile dummy thread support routines
if pthread isnt available
2001-09-23 Dick Porter <[email protected]>
* configure.in: Check for pthread.h, so that we can check for the
HAVE_PTHREAD_H define in config.h
2001-09-21 Dick Porter <[email protected]>
* configure.in: Check for libpthread
2001-09-19 Dietmar Maurer <[email protected]>
* mono/tests/test-ops.cs: added more tests
Mon Sep 10 20:19:00 CEST 2001 Paolo Molaro <[email protected]>
* configure.in: check for sizeof(void*) and for the architecture.
2001-09-05 Dick Porter <[email protected]>
* autogen.sh: Fixed the section that adds $ACLOCAL_FLAGS to the
aclocal invocation, so putting aclocal files in /usr/local works
2001-08-30 Dietmar Maurer <[email protected]>
* *: removed the libffi dependency, mono now compiles on cygwin
* mono/wrapper/build-dll: new file.
2001-08-28 Dietmar Maurer <[email protected]>
* mono/tests/console.cs: impl.
* mono/tests/box.cs: impl.
* mono/tests/stream-writer.cs: impl.
Mon Aug 27 20:24:26 CEST 2001 Paolo Molaro <[email protected]>
* mono/tests/*: New test cases, mostly for virtual method dispatch.
Use make testb for the slow benchmarking programs,
make test for the usual regression test checks.
2001-08-22 Rodrigo Moya <[email protected]>
* mono/cil/Makefile.am:
* mono/wrapper/Makefile.am: use correct variables for installation
directories
Tue Aug 21 18:54:06 CEST 2001 Paolo Molaro <[email protected]>
+1 -1
View File
@@ -1 +1 @@
SUBDIRS = mono doc runtime scripts man
SUBDIRS = libffi mono doc runtime
-10
View File
@@ -1,13 +1,3 @@
2002-Feb-11 Miguel de Icaza <[email protected]>
New release, functional x86-JIT, x86 interpreter, ppc interpreter
Class libraries ship.
Limited compiler ships.
Too many changes to list
2001-07-12 Miguel de Icaza <[email protected]>
New XSLT file from Sergey Chaban for CIL opcodes
+1 -11
View File
@@ -1,13 +1,5 @@
This is Mono.
Installation:
Read doc/download for installation instructions,
or read the HTML version:
http://www.go-mono.com/download.html
A directory roadmap:
doc/
@@ -45,8 +37,6 @@ A directory roadmap:
You can try the interpreter like this:
mint program.exe
mono-int program.exe
-6
View File
@@ -1,6 +0,0 @@
#undef HAVE_PTHREAD
#undef HAVE_PTHREAD_MUTEX_TIMEDLOCK
#undef HAVE_LARGE_FILE_SUPPORT
#undef PLATFORM_WIN32
#undef PLATFORM_WIN32_NATIVE
#undef HAVE_NEW_ICONV
+49 -28
View File
@@ -78,36 +78,57 @@ xlc )
am_opt=--include-deps;;
esac
for coin in `find $srcdir -name configure.in -print`
do
dr=`dirname $coin`
if test -f $dr/NO-AUTO-GEN; then
echo skipping $dr -- flagged as no auto-gen
else
echo processing $dr
macrodirs=`sed -n -e 's,AM_ACLOCAL_INCLUDE(\(.*\)),\1,gp' < $coin`
( cd $dr
macrosdir=`find . -name macros -print`
if grep "^AM_GNU_GETTEXT" configure.in >/dev/null; then
if grep "sed.*POTFILES" configure.in >/dev/null; then
: do nothing -- we still have an old unmodified configure.in
else
echo "Creating $dr/aclocal.m4 ..."
test -r $dr/aclocal.m4 || touch $dr/aclocal.m4
echo "Running gettextize... Ignore non-fatal messages."
echo "no" | gettextize --force --copy
echo "Making $dr/aclocal.m4 writable ..."
test -r $dr/aclocal.m4 && chmod u+w $dr/aclocal.m4
fi
fi
if grep "^AM_PROG_LIBTOOL" configure.in >/dev/null; then
if test -z "$NO_LIBTOOLIZE" ; then
echo "Running libtoolize..."
libtoolize --force --copy
fi
fi
echo "Running aclocal $aclocalinclude ..."
aclocal $aclocalinclude || {
echo
echo "**Error**: aclocal failed. This may mean that you have not"
echo "installed all of the packages you need, or you may need to"
echo "set ACLOCAL_FLAGS to include \"-I \$prefix/share/aclocal\""
echo "for the prefix where you installed the packages whose"
echo "macros were not found"
exit 1
}
if grep "^AM_PROG_LIBTOOL" configure.in >/dev/null; then
if test -z "$NO_LIBTOOLIZE" ; then
echo "Running libtoolize..."
libtoolize --force --copy
if grep "^AM_CONFIG_HEADER" configure.in >/dev/null; then
echo "Running autoheader..."
autoheader || { echo "**Error**: autoheader failed."; exit 1; }
fi
echo "Running automake --gnu $am_opt ..."
automake --add-missing --gnu $am_opt ||
{ echo "**Error**: automake failed."; exit 1; }
echo "Running autoconf ..."
autoconf || { echo "**Error**: autoconf failed."; exit 1; }
) || exit 1
fi
fi
echo "Running aclocal $ACLOCAL_FLAGS ..."
aclocal $ACLOCAL_FLAGS || {
echo
echo "**Error**: aclocal failed. This may mean that you have not"
echo "installed all of the packages you need, or you may need to"
echo "set ACLOCAL_FLAGS to include \"-I \$prefix/share/aclocal\""
echo "for the prefix where you installed the packages whose"
echo "macros were not found"
exit 1
}
if grep "^AM_CONFIG_HEADER" configure.in >/dev/null; then
echo "Running autoheader..."
autoheader || { echo "**Error**: autoheader failed."; exit 1; }
fi
echo "Running automake --gnu $am_opt ..."
automake --add-missing --gnu $am_opt ||
{ echo "**Error**: automake failed."; exit 1; }
echo "Running autoconf ..."
autoconf || { echo "**Error**: autoconf failed."; exit 1; }
done
conf_flags="--enable-maintainer-mode --enable-compile-warnings" #--enable-iso-c
+14
View File
@@ -0,0 +1,14 @@
/* config.h.in. Generated automatically from configure.in by autoheader 2.13. */
/* Define if you have the ANSI C header files. */
#undef STDC_HEADERS
/* Define if you have the <dlfcn.h> header file. */
#undef HAVE_DLFCN_H
/* Name of package */
#undef PACKAGE
/* Version number of package */
#undef VERSION
+14 -282
View File
@@ -1,296 +1,37 @@
AC_INIT(README)
AC_CANONICAL_SYSTEM
AM_CONFIG_HEADER(config.h)
AM_INIT_AUTOMAKE(mono, 0.8)
AM_INIT_AUTOMAKE(mono, 0.6)
AM_MAINTAINER_MODE
AC_CHECK_TOOL(CC, gcc, gcc)
AC_PROG_CC
AM_PROG_CC_STDC
AC_PROG_INSTALL
dnl may require a specific autoconf version
dnl AC_PROG_CC_FOR_BUILD
dnl CC_FOR_BUILD not automatically detected
CC_FOR_BUILD=$CC
BUILD_EXEEXT=
if test "x$cross_compiling" = "xyes"; then
CC_FOR_BUILD=cc
BUILD_EXEEXT=""
fi
AC_SUBST(CC_FOR_BUILD)
AC_SUBST(HOST_CC)
AC_SUBST(BUILD_EXEEXT)
# Set STDC_HEADERS
AC_HEADER_STDC
AC_LIBTOOL_WIN32_DLL
AM_PROG_LIBTOOL
# not 64 bit clean in cross-compile
AC_CHECK_SIZEOF(void *, 4)
dnl ***************************
dnl *** Checks for glib 1.2 ***
dnl ***************************
AM_PATH_GLIB(1.2.0,,
AC_MSG_ERROR([Cannot find GLIB: Is glib-config in path?]))
CFLAGS='-g -Wall -Wunused -Wmissing-prototypes -Wmissing-declarations'
AC_PATH_PROG(PKG_CONFIG, pkg-config, no)
if test "x$PKG_CONFIG" = "xno"; then
AC_MSG_ERROR([You need to install pkg-config])
fi
dnl for use on the build system
dnl pkg-config is stupid
BUILD_GLIB_CFLAGS=`$PKG_CONFIG --cflags glib-2.0`
BUILD_GLIB_LIBS=`$PKG_CONFIG --libs glib-2.0`
AC_SUBST(BUILD_GLIB_CFLAGS)
AC_SUBST(BUILD_GLIB_LIBS)
PKG_PATH=
AC_ARG_WITH(crosspkgdir, [ --with-crosspkgdir=/path/to/pkg-config/dir],
if test x$with_crosspkgdir = "x"; then
if test -s $PKG_CONFIG_PATH; then
PKG_PATH=$PKG_CONFIG_PATH
fi
else
PKG_PATH=$with_crosspkgdir
PKG_CONFIG_PATH=$PKG_PATH
export PKG_CONFIG_PATH
fi
)
## Versions of dependencies
GLIB_REQUIRED_VERSION=1.3.11
PKG_CHECK_MODULES(BASE_DEPENDENCIES, glib-2.0 >= $GLIB_REQUIRED_VERSION)
GLIB_CFLAGS=`$PKG_CONFIG --cflags glib-2.0`
GLIB_LIBS=`$PKG_CONFIG --libs glib-2.0`
GMODULE_CFLAGS=`$PKG_CONFIG --cflags gmodule-2.0`
GMODULE_LIBS=`$PKG_CONFIG --libs gmodule-2.0`
GLIB_CFLAGS=`glib-config --cflags glib`
GLIB_LIBS=`glib-config --libs glib`
AC_SUBST(GLIB_CFLAGS)
AC_SUBST(GLIB_LIBS)
GMODULE_CFLAGS=`glib-config --cflags gmodule`
GMODULE_LIBS=`glib-config --libs gmodule`
AC_SUBST(GMODULE_CFLAGS)
AC_SUBST(GMODULE_LIBS)
dnl ****************************************
dnl *** Check if we're building on win32 ***
dnl ****************************************
AC_MSG_CHECKING([if building for some Win32 platform])
case "$host" in
*-*-mingw*|*-*-cygwin*)
platform_win32=yes
AC_DEFINE(PLATFORM_WIN32)
;;
*)
platform_win32=no
;;
esac
AC_MSG_RESULT($platform_win32)
AM_CONDITIONAL(PLATFORM_WIN32, test x$platform_win32 = xyes)
AC_CONFIG_SUBDIRS(libffi)
if test x$platform_win32 = xyes; then
AC_MSG_CHECKING([if building for native Win32])
case "$host" in
*-*-mingw*)
platform_win32_native=yes
AC_DEFINE(PLATFORM_WIN32_NATIVE)
;;
*)
platform_win32_native=no
;;
esac
AC_MSG_RESULT($platform_win32_native)
AM_CONDITIONAL(PLATFORM_WIN32_NATIVE,
test x$platform_win32_native = xyes)
fi
if test x$platform_win32 = xno; then
dnl ******************************************************************
dnl *** Check for large file support ***
dnl *** (If we were using autoconf 2.50 we'd use AC_SYS_LARGEFILE) ***
dnl ******************************************************************
# Check that off_t can represent 2**63 - 1 correctly, working around
# potential compiler bugs. Defines LARGE_FILE_SUPPORT, adds $1 to
# CFLAGS and sets $large_offt to yes if the test succeeds
large_offt=no
AC_DEFUN(LARGE_FILES, [
large_CPPFLAGS=$CPPFLAGS
CPPFLAGS="$CPPFLAGS $1"
AC_TRY_RUN([
#include <sys/types.h>
#define BIG_OFF_T (((off_t)1<<62)-1+((off_t)1<<62))
int main(void) {
int big_off_t=((BIG_OFF_T%2147483629==721) &&
(BIG_OFF_T%2147483647==1));
if(big_off_t) {
exit(0);
} else {
exit(1);
}
}
], [
AC_MSG_RESULT(ok)
AC_DEFINE(HAVE_LARGE_FILE_SUPPORT)
CFLAGS="$CFLAGS $1"
large_offt=yes
], [
AC_MSG_RESULT(no)
], "")
CPPFLAGS=$large_CPPFLAGS
])
AC_MSG_CHECKING(if off_t is 64 bits wide)
LARGE_FILES("")
if test $large_offt = no; then
AC_MSG_CHECKING(if _FILE_OFFSET_BITS=64 gives 64 bit off_t)
LARGE_FILES("-D_FILE_OFFSET_BITS=64")
fi
if test $large_offt = no; then
AC_MSG_WARN([No 64 bit file size support available])
fi
dnl *****************************
dnl *** Checks for libpthread ***
dnl *****************************
AC_SEARCH_LIBS(pthread_create, pthread, [
AM_CONDITIONAL(THREADS_PTHREAD, true)
AC_DEFINE(HAVE_PTHREAD)
# Need pthread_mutex_timedlock
pthread_CFLAGS=""
orig_CPPFLAGS=$CPPFLAGS
# This is a gcc-specific error, but we already set
# gcc-specific options in CFLAGS
CPPFLAGS="$CPPFLAGS -Werror-implicit-function-declaration"
AC_MSG_CHECKING(for pthread_mutex_timedlock)
AC_TRY_COMPILE([ #include <pthread.h>], [
pthread_mutex_t mut=PTHREAD_MUTEX_INITIALIZER;
pthread_mutex_timedlock(&mut, NULL);
], [
# Works!
AC_MSG_RESULT(ok)
AC_DEFINE(HAVE_PTHREAD_MUTEX_TIMEDLOCK)
], [
AC_MSG_RESULT(no)
# glibc requires -D_GNU_SOURCE before it will declare
# this function
AC_MSG_CHECKING(whether _GNU_SOURCE is needed for pthread_mutex_timedlock)
CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"
AC_TRY_COMPILE([ #include <pthread.h>], [
pthread_mutex_t mut=PTHREAD_MUTEX_INITIALIZER;
pthread_mutex_timedlock(&mut, NULL);
], [
AC_MSG_RESULT(ok)
pthread_CFLAGS="-D_GNU_SOURCE"
AC_DEFINE(HAVE_PTHREAD_MUTEX_TIMEDLOCK)
], [
AC_MSG_RESULT(no)
dnl Add other variants here
AC_MSG_WARN(Working around pthread_mutex_timedlock)
])
])
CPPFLAGS=$orig_CPPFLAGS
CFLAGS="$CFLAGS $pthread_CFLAGS"
# Need PTHREAD_MUTEX_RECURSIVE
pthread_CFLAGS=""
orig_CPPFLAGS=$CPPFLAGS
# This is a gcc-specific error, but we already set
# gcc-specific options in CFLAGS
CPPFLAGS="$CPPFLAGS -Werror-implicit-function-declaration"
AC_MSG_CHECKING(for PTHREAD_MUTEX_RECURSIVE)
AC_TRY_COMPILE([ #include <pthread.h>], [
pthread_mutexattr_t attr;
pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
], [
# Works!
AC_MSG_RESULT(ok)
], [
AC_MSG_RESULT(no)
# glibc requires -D_GNU_SOURCE before it will declare
# this macro
AC_MSG_CHECKING(whether _GNU_SOURCE is needed for PTHREAD_MUTEX_RECURSIVE)
CPPFLAGS="$CPPFLAGS -D_GNU_SOURCE"
AC_TRY_COMPILE([ #include <pthread.h>], [
pthread_mutexattr_t attr;
pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
], [
AC_MSG_RESULT(ok)
pthread_CFLAGS="-D_GNU_SOURCE"
], [
AC_MSG_RESULT(no)
dnl Add other variants here
AC_MSG_ERROR([PTHREAD_MUTEX_RECURSIVE not found])
])
])
CPPFLAGS=$orig_CPPFLAGS
CFLAGS="$CFLAGS $pthread_CFLAGS"
], [
AC_MSG_ERROR([libpthread is required on non-win32 hosts])
])
fi
if test "x$cross_compiling" = "xno"; then
# check for new iconv version
AC_MSG_CHECKING(for new iconv)
AC_CACHE_VAL(new_iconv,[
AC_TRY_RUN([#include <stdio.h>
#include <iconv.h>
int main()
{
exit (iconv_open ("UTF-16le", "UTF-8") == (iconv_t)-1);
}
],
new_iconv=yes,new_iconv=)])
if test -n "$new_iconv"; then
AC_MSG_RESULT(yes)
AC_DEFINE(HAVE_NEW_ICONV)
else
AC_MSG_RESULT(no)
fi
fi
TARGET="unknown"
ACCESS_UNALIGNED="yes"
case "$host" in
#mips-sgi-irix5.* | mips-sgi-irix6.*) TARGET=MIPS; ACCESS_UNALIGNED="no";;
i*86-*-*) TARGET=X86; arch_target=x86;;
sparc*-*-*) TARGET=SPARC; ACCESS_UNALIGNED="no";;
#alpha*-*-linux* | alpha*-*-osf*) TARGET=ALPHA; ACCESS_UNALIGNED="no";;
#m68k-*-linux*) TARGET=M68K;;
powerpc-*-linux* | powerpc-*-sysv*) TARGET=POWERPC; arch_target=ppc;;
#arm-*-linux-*) TARGET=ARM; ACCESS_UNALIGNED="no";;
esac
if test ${TARGET} = unknown; then
CFLAGS="$CFLAGS -DNO_PORT"
AC_MSG_WARN("mono has not been ported to $host: some things may not work.")
fi
if test ${ACCESS_UNALIGNED} = no; then
CPPFLAGS="$CPPFLAGS -DNO_UNALIGNED_ACCESS"
fi
AM_CONDITIONAL(MIPS_GCC, test ${TARGET}${ac_cv_prog_gcc} = MIPSyes)
AM_CONDITIONAL(MIPS_SGI, test ${TARGET}${ac_cv_prog_gcc} = MIPSno)
AM_CONDITIONAL(SPARC, test x$TARGET = xSPARC)
AM_CONDITIONAL(X86, test x$TARGET = xX86)
AM_CONDITIONAL(ALPHA, test x$TARGET = xALPHA)
AM_CONDITIONAL(M68K, test x$TARGET = xM68K)
AM_CONDITIONAL(POWERPC, test x$TARGET = xPOWERPC)
AM_CONDITIONAL(ARM, test x$TARGET = xARM)
AC_SUBST(arch_target)
CFLAGS='-g -Wall -Wunused -Wmissing-prototypes -Wmissing-declarations'
AC_SUBST(CFLAGS)
AC_OUTPUT([
@@ -299,18 +40,9 @@ mono/Makefile
mono/metadata/Makefile
mono/dis/Makefile
mono/cil/Makefile
mono/arch/Makefile
mono/arch/x86/Makefile
mono/arch/ppc/Makefile
mono/arch/sparc/Makefile
mono/interpreter/Makefile
mono/tests/Makefile
mono/wrapper/Makefile
mono/monoburg/Makefile
mono/jit/Makefile
mono/io-layer/Makefile
runtime/Makefile
scripts/Makefile
man/Makefile
doc/Makefile
])
+2 -10
View File
@@ -1,7 +1,7 @@
WEB_FILES= \
c-sharp class-library contact contributing documentation download \
faq gcc-frontend ideas index passport rationale resources \
roadmap runtime status team testing thanks tools porting anoncvs
roadmap runtime status team testing thanks tools
OTHERS= pending resources-pending todo
@@ -12,15 +12,7 @@ all-docs: $(WEB_FILES)
runtime documentation download faq contributing \
resources status > all-docs
class-status: class-status.in pending-classes Makefile.am
cat class-status.in pending-classes > class-status
pending-classes: pending-classes.in Makefile.am
echo "This is a list of pending classes, and links to their documentation on MSDN.Microsoft.com<p><p>" > pending-classes
for i in `cat pending-classes.in`; do x=`echo $$i | sed s/\\\.//g`; link=`echo http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrf@@@ClassTopic.asp | sed "s/@@@/$$x/"`; echo " <a href=\"$$link\">$$i</a><br>" >> pending-classes; done
webit: class-status
webit:
(cd web; make && make push)
push-notes:
-54
View File
@@ -1,54 +0,0 @@
* Anonymous CVS access
Anonymous access to the Mono is available currently from the
HispaLinux CVS server (later I will provide a nice
anoncvs.go-mono.com alias).
Sources are currently being updated every hour.
The server is currently located in Spain. If you want to
become an AnonCVS mirror, go <a href="#mirror">here</a>
** Checking out the sources
To check out the sources for the first time from the
repository, use this command:
<pre>
export CVSROOT=:pserver:[email protected]:/mono
cvs -z3 co mcs mono gtk-sharp
</pre>
That will get you the compiler and class libraries (mcs), the
interpreter and JITer (mono) and the Gtk# bindings.
The server is located in Spain.
** Updating the sources
To update your sources every day, you use this command:
<pre>
export CVSROOT=:pserver:[email protected]:/mono
cvs -z3 update -dP mcs mono gtk-sharp
</pre>
** Thanks
Thanks to <a href="http://www.hispalinux.es">HispaLinux</a>
for providing this Anon CVS server. The server is located in
Spain.
Special thanks to Jesus Climent, Rodrigo Moya, Juantomás
García and Ismael Olea for making this happen.
<a name="#mirror">
** Becoming an AnonCVS mirror
To become an AnonCVS mirror, you currently need about 20
megabytes of free disk space and bandwidth that you do not
mind sharing (you probably want enough to spare).
I need to be able to rsync to your machine from the main CVS
server. The preferred method to rsync is to setup ssh in your
machine.
+45 -78
View File
@@ -1,44 +1,19 @@
* MCS: The Ximian C# compiler
MCS is currently able to compile itself and many more C#
programs (there is a test suite included that you can use).
MCS began as an experiment to learn the features of C# by
writing a large C# program. MCS is currently able to parse C#
programs and create an internal tree representation of the
program. MCS can parse itself.
We are in feature completion mode right now. There are still
a couple of areas that are not covered by the Mono compiler, but
they are very very few at this point.
MCS now does type checking at the class, interface and struct
levels and can resolve the class hierarchy and as of last week
can generate interface code.
MCS was able to parse itself on April 2001, MCS compiled itself
for the first time on December 28 2001. MCS became self hosting
on January 3rd, 2002.
A test suite is maintained to track the progress of
the compiler and various programs are routinely compiled and
ran.
** Obtaining MCS
The Mono C# compiler is part of the `mcs' module in the Mono CVS
you can get it from our <a href="anoncvs.html">Anonymous CVS</a> server,
or you can get nightly <a href="download.html">download page</a>.
** Running MCS
MCS is written in C# and uses heavily the .NET APIs. At this point
MCS only runs on Windows with the .NET Framework SDK installed. Work
is progressing rapidly in our JIT engine and our class libraries to
allow MCS to be ran in non-Windows systems.
** Reporting Bugs in MCS
When you report a bug, try to provide a small test case that would
show the error so we can include this as part of the Mono C# regression
test suite.
If the bug is an error or a warning that we do not flag, write
a sample program called `csXXXX.cs' where XXXX is the code number
that is used by the Microsoft C# compiler that illustrates the
problem. That way we can also do regression tests on the invalid
input.
Work is progressing quickly on various fronts in the C#
compiler. Recently I started using the System.Reflection API
to load system type definitions and avoid self-population of
types in the compiler and dropped my internal Type
representation in favor of using the CLI's System.Type.
** Phases of the compiler
@@ -71,62 +46,58 @@
have to postpone this decision until the above steps
are finished.
* Code generation: The code generation is done through
the System.Reflection.Emit API.
* Code generation: The compiler recently started generating IL
executables that contain interfaces. Work is
progressing in other areas.
The code generation is done through the System.Reflection.Emit API.
</ul>
** CIL Optimizations.
The compiler performs a number of simple optimizations on its input:
constant folding (this is required by the C# language spec) and
can perform dead code elimination.
Other more interesting optimizations like hoisting are not possible
at this point since the compiler output at this point does not
generate an intermediate representation that is suitable to
perform basic block computation.
Adding an intermediate layer to enable the basic block
computation to the compiler should be a simple task, but we
are considering having a generic CIL optimizer. Since all the
information that is required to perform basic block-based
optimizations is available at the CIL level, we might just skip
this step altogether and have just a generic IL optimizer that
would perform hoisting on arbitrary CIL programs, not only
those produced by MCS.
If this tool is further expanded to perform constant folding
(not needed for our C# compiler, as it is already in there)
and dead code elimination, other compiler authors might be
able to use this generic CIL optimizer in their projects
reducing their time to develop a production compiler.
<a name="tasks">
** Current pending tasks
Simple tasks:
<ul>
* Redo the way we deal with built-in operators.
</ul>
* Array declarations are currently being ignored,
Larger tasks:
<ul>
* PInvoke declarations are not supported.
* Pre-processing is not supported.
* Attribute declarations and passing currently ignored.
* Compiler does not pass around line/col information from tokenizer for error reporting.
* Jay does not work correctly with `error'
productions, making parser errors hard to point. It
would be best to port the Bison-To-Java compiler to
become Bison-to-C# compiler.
Nick Drochak has started a project on SourceForge for this.
You can find the project at: <a href="http://sourceforge.net/projects/jb2csharp/">
http://sourceforge.net/projects/jb2csharp/</a>
become Bison-to-C# compiler ([email protected]
might have more information)
</ul>
Interesting and Fun hacks to the compiler:
<ul>
* Finishing the JB port from Java to C#. If you are
interested in working on this, please contact Brian
Jepson (bjepson at oreilly d-o-t com).
More on JB at: <a href="http://www.cs.colorado.edu/~dennis/software/jb.html">
http://www.cs.colorado.edu/~dennis/software/jb.html</a>
JB will allow us to move from the Berkeley Yacc
based Jay to a Bison-based compiler (better error
reporting and recovery).
* Semantic Analysis: Return path coverage and
initialization before use coverage are two great
features of C# that help reduce the number of bugs
in applications. It is one interesting hack.
* Enum resolutions: it is another fun hack, as enums can be defined
in terms of themselves (<tt>enum X { a = b + 1, b = 5 }</tt>).
</ul>
** Questions and Answers
@@ -155,10 +126,6 @@ Q: If your C# compiler is written in C#, how do you plan on getting
We will do this through an implementation of the CLI Virtual
Execution System for Unix (our JIT engine).
Our JIT engine is working for the purposes of using the compiler.
The supporting class libraries are being worked on to fully support
the compiler.
Q: Do you use Bison?
A: No, currently I am using Jay which is a port of Berkeley Yacc to
+3 -29
View File
@@ -1,10 +1,7 @@
* CVS Access
If you are an active Mono developer, you can get a CVS account
that hosts the Mono source code.
If you are not a developer, but want to track the development, please
see the <a href="anoncvs.html">AnonCVS</a> instructions.
that hosts the Mono source code.
Send an e-mail to miguel with your public SSH key for this
purpose. Please specify if the key was generated with SSH1 or SSH2.
@@ -67,25 +64,6 @@
mail [email protected] < file.diff
</pre>
** Keeping track of changes.
We provide two e-mail based mechanisms to keep track of
changes to the code base:
<ul>
* <a href="mailto:[email protected]">
[email protected]</a>: This mailing list receives
in patch form all the changes that are being made to the
CVS.
* <a href="mailto:[email protected]">
[email protected]</a>: This mailing list only
receives the CVS commit logs with a list of files
modified.
</ul>
We hope to offer LXR and Bonsai in the future as well.
** Commiting your work
Once you get approval to commit to the CVS, or if you are
@@ -119,13 +97,9 @@
** Recommendations
Please do not commit code that would break the compile to the
CVS, because that normally wastes everybody's time. Two things
are important in this step: trying to build your sources and making
sure that you add all the new files before you do a commit.
CVS, because that normally wastes everybody's time.
To build the sources, most of the type trying the `make' command
is enough. In some cases (the class libraries) we use nant, so
you need to run nant manually.
Make sure that you add all the files before you do a commit.
Use ChangeLog entries so we can keep textual descriptions of
your work, and use the contents of your ChangeLog file as the
+28 -35
View File
@@ -30,18 +30,9 @@
You can also track live the activities of the Mono CVS module
by subscribing to the <a
href="http://lists.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list</a>
href="http://mail.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list</a>
** Missing features
Our class libraries miss some features, for example, most classes
do not implement the serialization bits at all, it would be a good
contribution to add this to each class.
This is a simple task, but it needs to be done in a compatible way
with the Microsoft.NET classes: using the same arguments to serialize
and reincarnate data.
** Layout
@@ -106,11 +97,33 @@
** Licensing
The class library is being licensed under the terms of the
<a
href="http://www.opensource.org/licenses/mit-license.html">MIT
license.</a> This is the same license used by the X11 window
system.
The class library will be licensed under the terms of the GNU
LGPL. Some people have pointed out that the plain LGPL is
troublesome for embedded use of the Mono class libraries. So
we are considering to use the GPL with a special exception
(like the <a
href="http://www.gnu.org/software/classpath/classpath.html">GNU
Classpath</a> project did.
The exception to the GPL would be:
<i>The library is distributed under the terms of the GNU General
Public License with the following exception:
If you link this library against your own program, then you do not
need to release the source code for that program. However, any
changes that you make to the library itself, or to any native
methods upon which the library relies, must be re-distributed in
accordance with the terms of the GPL.</i>
If you are going to contribute, please keep in mind that we
might require you to agree that Ximian might adjust the
license to enable the use of the class libraries on embedded
systems or to develop proprietary applications using Mono.
We suggest that you assign the copyright of your work to the
GNOME Foundation or the Free Software Foundation to simplify
defending the code in case it is used inappropiately.
** Class Library testing
@@ -156,26 +169,6 @@
In those cases, use the keyword "LAMESPEC".
** Namespaces
We are using a number of namespaces in the class libraries for
private classes when we need them, you are encouraged to use these:
<pre>
Mono
.Languages // General Compiler classes
.CSharp // Specific C# parsing/tokenizing classes
.MonoBASIC // Specific VB.NET/MonoBASIC parsing/tokenizing classes
.Python // Specific Python parsing/tokenizing classes
.Runtime
.LateBind // General latebinding-support for MonoBASIC/Python
.MonoBASIC // Specific runtime classes for MonoBASIC programs
.Python // Specific runtime classes for Python programs
.Web
.UI
.Utils // Gaurav utility classes
</pre>
** FAQ
Frequently asked questions about the class library:
-391
View File
@@ -13,395 +13,4 @@
You can also download the master <a
href="class-status/class.xml">Class Status XML</a> file.
* List of not-implemented classes
These classes are only classes pending from the CORLIB assembly
other assemblies are not yet listed. We are going to be working
on an improved status display for pending classes.
Please bear with us in the meantime.
This is a list of pending classes, and links to their documentation on MSDN.Microsoft.com<p><p>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfMicrosoftWin32Registry␍ClassTopic.asp">Microsoft.Win32.Registry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfMicrosoftWin32RegistryHive␍ClassTopic.asp">Microsoft.Win32.RegistryHive␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfMicrosoftWin32RegistryKey␍ClassTopic.asp">Microsoft.Win32.RegistryKey␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystem__ComObject␍ClassTopic.asp">System.__ComObject␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemActivator␍ClassTopic.asp">System.Activator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemArgIterator␍ClassTopic.asp">System.ArgIterator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemAssemblyLoadEventHandler␍ClassTopic.asp">System.AssemblyLoadEventHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemAsyncCallback␍ClassTopic.asp">System.AsyncCallback␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemCrossAppDomainDelegate␍ClassTopic.asp">System.CrossAppDomainDelegate␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemDBNull␍ClassTopic.asp">System.DBNull␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemDiagnosticsSymbolStoreSymDocumentType␍ClassTopic.asp">System.Diagnostics.SymbolStore.SymDocumentType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemDiagnosticsSymbolStoreSymLanguageType␍ClassTopic.asp">System.Diagnostics.SymbolStore.SymLanguageType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemDiagnosticsSymbolStoreSymLanguageVendor␍ClassTopic.asp">System.Diagnostics.SymbolStore.SymLanguageVendor␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemEventHandler␍ClassTopic.asp">System.EventHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationCompareInfo␍ClassTopic.asp">System.Globalization.CompareInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationHebrewCalendar␍ClassTopic.asp">System.Globalization.HebrewCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationHijriCalendar␍ClassTopic.asp">System.Globalization.HijriCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationJapaneseCalendar␍ClassTopic.asp">System.Globalization.JapaneseCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationKoreanCalendar␍ClassTopic.asp">System.Globalization.KoreanCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationRegionInfo␍ClassTopic.asp">System.Globalization.RegionInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationSortKey␍ClassTopic.asp">System.Globalization.SortKey␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationStringInfo␍ClassTopic.asp">System.Globalization.StringInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationTaiwanCalendar␍ClassTopic.asp">System.Globalization.TaiwanCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationTextElementEnumerator␍ClassTopic.asp">System.Globalization.TextElementEnumerator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationTextInfo␍ClassTopic.asp">System.Globalization.TextInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemGlobalizationThaiBuddhistCalendar␍ClassTopic.asp">System.Globalization.ThaiBuddhistCalendar␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemIOEndOfStreamException␍ClassTopic.asp">System.IO.EndOfStreamException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemIOIsolatedStorageIsolatedStorage␍ClassTopic.asp">System.IO.IsolatedStorage.IsolatedStorage␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemIOIsolatedStorageIsolatedStorageException␍ClassTopic.asp">System.IO.IsolatedStorage.IsolatedStorageException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemIOIsolatedStorageIsolatedStorageFile␍ClassTopic.asp">System.IO.IsolatedStorage.IsolatedStorageFile␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemIOPathTooLongException␍ClassTopic.asp">System.IO.PathTooLongException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionEmitMethodRental␍ClassTopic.asp">System.Reflection.Emit.MethodRental␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionMemberFilter␍ClassTopic.asp">System.Reflection.MemberFilter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionMissing␍ClassTopic.asp">System.Reflection.Missing␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionModuleResolveEventHandler␍ClassTopic.asp">System.Reflection.ModuleResolveEventHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionPointer␍ClassTopic.asp">System.Reflection.Pointer␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionStrongNameKeyPair␍ClassTopic.asp">System.Reflection.StrongNameKeyPair␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionTargetException␍ClassTopic.asp">System.Reflection.TargetException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemReflectionTypeFilter␍ClassTopic.asp">System.Reflection.TypeFilter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemResolveEventHandler␍ClassTopic.asp">System.ResolveEventHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesAccessedThroughPropertyAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.AccessedThroughPropertyAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCallConvCdecl␍ClassTopic.asp">System.Runtime.CompilerServices.CallConvCdecl␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCallConvFastcall␍ClassTopic.asp">System.Runtime.CompilerServices.CallConvFastcall␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCallConvStdcall␍ClassTopic.asp">System.Runtime.CompilerServices.CallConvStdcall␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCallConvThiscall␍ClassTopic.asp">System.Runtime.CompilerServices.CallConvThiscall␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCompilationRelaxationsAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.CompilationRelaxationsAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCompilerGlobalScopeAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.CompilerGlobalScopeAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesCustomConstantAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.CustomConstantAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesDateTimeConstantAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.DateTimeConstantAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesDecimalConstantAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.DecimalConstantAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesDiscardableAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.DiscardableAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesIDispatchConstantAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.IDispatchConstantAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesIsVolatile␍ClassTopic.asp">System.Runtime.CompilerServices.IsVolatile␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesIUnknownConstantAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.IUnknownConstantAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeCompilerServicesRequiredAttributeAttribute␍ClassTopic.asp">System.Runtime.CompilerServices.RequiredAttributeAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesArrayWithOffset␍ClassTopic.asp">System.Runtime.InteropServices.ArrayWithOffset␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesAssemblyRegistrationFlags␍ClassTopic.asp">System.Runtime.InteropServices.AssemblyRegistrationFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesAutomationProxyAttribute␍ClassTopic.asp">System.Runtime.InteropServices.AutomationProxyAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesBIND_OPTS␍ClassTopic.asp">System.Runtime.InteropServices.BIND_OPTS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesBINDPTR␍ClassTopic.asp">System.Runtime.InteropServices.BINDPTR␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesCALLCONV␍ClassTopic.asp">System.Runtime.InteropServices.CALLCONV␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesClassInterfaceAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ClassInterfaceAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesClassInterfaceType␍ClassTopic.asp">System.Runtime.InteropServices.ClassInterfaceType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesCoClassAttribute␍ClassTopic.asp">System.Runtime.InteropServices.CoClassAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComAliasNameAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComAliasNameAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComConversionLossAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComConversionLossAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComEventInterfaceAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComEventInterfaceAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesCOMException␍ClassTopic.asp">System.Runtime.InteropServices.COMException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComImportAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComImportAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComInterfaceType␍ClassTopic.asp">System.Runtime.InteropServices.ComInterfaceType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComMemberType␍ClassTopic.asp">System.Runtime.InteropServices.ComMemberType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComRegisterFunctionAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComRegisterFunctionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComSourceInterfacesAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComSourceInterfacesAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComUnregisterFunctionAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComUnregisterFunctionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesComVisibleAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ComVisibleAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesCONNECTDATA␍ClassTopic.asp">System.Runtime.InteropServices.CONNECTDATA␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesCurrencyWrapper␍ClassTopic.asp">System.Runtime.InteropServices.CurrencyWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesDESCKIND␍ClassTopic.asp">System.Runtime.InteropServices.DESCKIND␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesDispatchWrapper␍ClassTopic.asp">System.Runtime.InteropServices.DispatchWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesDispIdAttribute␍ClassTopic.asp">System.Runtime.InteropServices.DispIdAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesDISPPARAMS␍ClassTopic.asp">System.Runtime.InteropServices.DISPPARAMS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesELEMDESC␍ClassTopic.asp">System.Runtime.InteropServices.ELEMDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesErrorWrapper␍ClassTopic.asp">System.Runtime.InteropServices.ErrorWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesEXCEPINFO␍ClassTopic.asp">System.Runtime.InteropServices.EXCEPINFO␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesExpandoIExpando␍ClassTopic.asp">System.Runtime.InteropServices.Expando.IExpando␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesExporterEventKind␍ClassTopic.asp">System.Runtime.InteropServices.ExporterEventKind␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesExtensibleClassFactory␍ClassTopic.asp">System.Runtime.InteropServices.ExtensibleClassFactory␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesFILETIME␍ClassTopic.asp">System.Runtime.InteropServices.FILETIME␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesFUNCDESC␍ClassTopic.asp">System.Runtime.InteropServices.FUNCDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesFUNCFLAGS␍ClassTopic.asp">System.Runtime.InteropServices.FUNCFLAGS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesFUNCKIND␍ClassTopic.asp">System.Runtime.InteropServices.FUNCKIND␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesGCHandle␍ClassTopic.asp">System.Runtime.InteropServices.GCHandle␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesGuidAttribute␍ClassTopic.asp">System.Runtime.InteropServices.GuidAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesHandleRef␍ClassTopic.asp">System.Runtime.InteropServices.HandleRef␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesICustomAdapter␍ClassTopic.asp">System.Runtime.InteropServices.ICustomAdapter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesICustomFactory␍ClassTopic.asp">System.Runtime.InteropServices.ICustomFactory␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesICustomMarshaler␍ClassTopic.asp">System.Runtime.InteropServices.ICustomMarshaler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIDispatchImplAttribute␍ClassTopic.asp">System.Runtime.InteropServices.IDispatchImplAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIDispatchImplType␍ClassTopic.asp">System.Runtime.InteropServices.IDispatchImplType␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIDLDESC␍ClassTopic.asp">System.Runtime.InteropServices.IDLDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIDLFLAG␍ClassTopic.asp">System.Runtime.InteropServices.IDLFLAG␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIMPLTYPEFLAGS␍ClassTopic.asp">System.Runtime.InteropServices.IMPLTYPEFLAGS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesImportedFromTypeLibAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ImportedFromTypeLibAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesImporterEventKind␍ClassTopic.asp">System.Runtime.InteropServices.ImporterEventKind␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesInterfaceTypeAttribute␍ClassTopic.asp">System.Runtime.InteropServices.InterfaceTypeAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesInvalidComObjectException␍ClassTopic.asp">System.Runtime.InteropServices.InvalidComObjectException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesInvalidOleVariantTypeException␍ClassTopic.asp">System.Runtime.InteropServices.InvalidOleVariantTypeException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesINVOKEKIND␍ClassTopic.asp">System.Runtime.InteropServices.INVOKEKIND␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesIRegistrationServices␍ClassTopic.asp">System.Runtime.InteropServices.IRegistrationServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesITypeLibConverter␍ClassTopic.asp">System.Runtime.InteropServices.ITypeLibConverter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesITypeLibExporterNameProvider␍ClassTopic.asp">System.Runtime.InteropServices.ITypeLibExporterNameProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesITypeLibExporterNotifySink␍ClassTopic.asp">System.Runtime.InteropServices.ITypeLibExporterNotifySink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesITypeLibImporterNotifySink␍ClassTopic.asp">System.Runtime.InteropServices.ITypeLibImporterNotifySink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesLCIDConversionAttribute␍ClassTopic.asp">System.Runtime.InteropServices.LCIDConversionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesLIBFLAGS␍ClassTopic.asp">System.Runtime.InteropServices.LIBFLAGS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesMarshalDirectiveException␍ClassTopic.asp">System.Runtime.InteropServices.MarshalDirectiveException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesObjectCreationDelegate␍ClassTopic.asp">System.Runtime.InteropServices.ObjectCreationDelegate␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesPARAMDESC␍ClassTopic.asp">System.Runtime.InteropServices.PARAMDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesPARAMFLAG␍ClassTopic.asp">System.Runtime.InteropServices.PARAMFLAG␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesPreserveSigAttribute␍ClassTopic.asp">System.Runtime.InteropServices.PreserveSigAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesPrimaryInteropAssemblyAttribute␍ClassTopic.asp">System.Runtime.InteropServices.PrimaryInteropAssemblyAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesProgIdAttribute␍ClassTopic.asp">System.Runtime.InteropServices.ProgIdAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesRegistrationServices␍ClassTopic.asp">System.Runtime.InteropServices.RegistrationServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesRuntimeEnvironment␍ClassTopic.asp">System.Runtime.InteropServices.RuntimeEnvironment␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesSafeArrayRankMismatchException␍ClassTopic.asp">System.Runtime.InteropServices.SafeArrayRankMismatchException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesSafeArrayTypeMismatchException␍ClassTopic.asp">System.Runtime.InteropServices.SafeArrayTypeMismatchException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesSEHException␍ClassTopic.asp">System.Runtime.InteropServices.SEHException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesSTATSTG␍ClassTopic.asp">System.Runtime.InteropServices.STATSTG␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesSYSKIND␍ClassTopic.asp">System.Runtime.InteropServices.SYSKIND␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTYPEATTR␍ClassTopic.asp">System.Runtime.InteropServices.TYPEATTR␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTYPEDESC␍ClassTopic.asp">System.Runtime.InteropServices.TYPEDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTYPEFLAGS␍ClassTopic.asp">System.Runtime.InteropServices.TYPEFLAGS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTYPEKIND␍ClassTopic.asp">System.Runtime.InteropServices.TYPEKIND␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTYPELIBATTR␍ClassTopic.asp">System.Runtime.InteropServices.TYPELIBATTR␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibConverter␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibConverter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibExporterFlags␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibExporterFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibFuncAttribute␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibFuncAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibFuncFlags␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibFuncFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibImporterFlags␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibImporterFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibTypeAttribute␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibTypeAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibTypeFlags␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibTypeFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibVarAttribute␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibVarAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesTypeLibVarFlags␍ClassTopic.asp">System.Runtime.InteropServices.TypeLibVarFlags␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIBindCtx␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIBindCtx␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIConnectionPoint␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIConnectionPoint␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIConnectionPointContainer␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIConnectionPointContainer␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIEnumConnectionPoints␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIEnumConnectionPoints␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIEnumConnections␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIEnumConnections␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIEnumMoniker␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIEnumMoniker␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIEnumString␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIEnumString␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIEnumVARIANT␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIEnumVARIANT␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIMoniker␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIMoniker␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIPersistFile␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIPersistFile␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIRunningObjectTable␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIRunningObjectTable␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMIStream␍ClassTopic.asp">System.Runtime.InteropServices.UCOMIStream␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMITypeComp␍ClassTopic.asp">System.Runtime.InteropServices.UCOMITypeComp␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMITypeInfo␍ClassTopic.asp">System.Runtime.InteropServices.UCOMITypeInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUCOMITypeLib␍ClassTopic.asp">System.Runtime.InteropServices.UCOMITypeLib␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesUnknownWrapper␍ClassTopic.asp">System.Runtime.InteropServices.UnknownWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesVARDESC␍ClassTopic.asp">System.Runtime.InteropServices.VARDESC␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeInteropServicesVARFLAGS␍ClassTopic.asp">System.Runtime.InteropServices.VARFLAGS␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingActivatedClientTypeEntry␍ClassTopic.asp">System.Runtime.Remoting.ActivatedClientTypeEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingActivatedServiceTypeEntry␍ClassTopic.asp">System.Runtime.Remoting.ActivatedServiceTypeEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingActivationUrlAttribute␍ClassTopic.asp">System.Runtime.Remoting.Activation.UrlAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsBaseChannelObjectWithProperties␍ClassTopic.asp">System.Runtime.Remoting.Channels.BaseChannelObjectWithProperties␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsBaseChannelSinkWithProperties␍ClassTopic.asp">System.Runtime.Remoting.Channels.BaseChannelSinkWithProperties␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsBaseChannelWithProperties␍ClassTopic.asp">System.Runtime.Remoting.Channels.BaseChannelWithProperties␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsChannelDataStore␍ClassTopic.asp">System.Runtime.Remoting.Channels.ChannelDataStore␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsChannelServices␍ClassTopic.asp">System.Runtime.Remoting.Channels.ChannelServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsClientChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.ClientChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannel␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannel␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannelDataStore␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannelDataStore␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannelReceiver␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannelReceiver␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannelReceiverHook␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannelReceiverHook␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannelSender␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannelSender␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIChannelSinkBase␍ClassTopic.asp">System.Runtime.Remoting.Channels.IChannelSinkBase␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientChannelSink␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientChannelSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientChannelSinkProvider␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientChannelSinkProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientFormatterSink␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientFormatterSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientFormatterSinkProvider␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientFormatterSinkProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIClientResponseChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.IClientResponseChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIServerChannelSink␍ClassTopic.asp">System.Runtime.Remoting.Channels.IServerChannelSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIServerChannelSinkProvider␍ClassTopic.asp">System.Runtime.Remoting.Channels.IServerChannelSinkProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIServerChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.IServerChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIServerFormatterSinkProvider␍ClassTopic.asp">System.Runtime.Remoting.Channels.IServerFormatterSinkProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsIServerResponseChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.IServerResponseChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsITransportHeaders␍ClassTopic.asp">System.Runtime.Remoting.Channels.ITransportHeaders␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsServerChannelSinkStack␍ClassTopic.asp">System.Runtime.Remoting.Channels.ServerChannelSinkStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsServerProcessing␍ClassTopic.asp">System.Runtime.Remoting.Channels.ServerProcessing␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsSinkProviderData␍ClassTopic.asp">System.Runtime.Remoting.Channels.SinkProviderData␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingChannelsTransportHeaders␍ClassTopic.asp">System.Runtime.Remoting.Channels.TransportHeaders␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsContextProperty␍ClassTopic.asp">System.Runtime.Remoting.Contexts.ContextProperty␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsCrossContextDelegate␍ClassTopic.asp">System.Runtime.Remoting.Contexts.CrossContextDelegate␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContextPropertyActivator␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContextPropertyActivator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContributeClientContextSink␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContributeClientContextSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContributeDynamicSink␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContributeDynamicSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContributeEnvoySink␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContributeEnvoySink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContributeObjectSink␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContributeObjectSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingContextsIContributeServerContextSink␍ClassTopic.asp">System.Runtime.Remoting.Contexts.IContributeServerContextSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingIChannelInfo␍ClassTopic.asp">System.Runtime.Remoting.IChannelInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingIEnvoyInfo␍ClassTopic.asp">System.Runtime.Remoting.IEnvoyInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingInternalRemotingServices␍ClassTopic.asp">System.Runtime.Remoting.InternalRemotingServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingIObjectHandle␍ClassTopic.asp">System.Runtime.Remoting.IObjectHandle␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingIRemotingTypeInfo␍ClassTopic.asp">System.Runtime.Remoting.IRemotingTypeInfo␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingLifetimeClientSponsor␍ClassTopic.asp">System.Runtime.Remoting.Lifetime.ClientSponsor␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingLifetimeILease␍ClassTopic.asp">System.Runtime.Remoting.Lifetime.ILease␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingLifetimeISponsor␍ClassTopic.asp">System.Runtime.Remoting.Lifetime.ISponsor␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingLifetimeLeaseState␍ClassTopic.asp">System.Runtime.Remoting.Lifetime.LeaseState␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingLifetimeLifetimeServices␍ClassTopic.asp">System.Runtime.Remoting.Lifetime.LifetimeServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingAsyncResult␍ClassTopic.asp">System.Runtime.Remoting.Messaging.AsyncResult␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingCallContext␍ClassTopic.asp">System.Runtime.Remoting.Messaging.CallContext␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingConstructionCall␍ClassTopic.asp">System.Runtime.Remoting.Messaging.ConstructionCall␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingConstructionResponse␍ClassTopic.asp">System.Runtime.Remoting.Messaging.ConstructionResponse␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingHeader␍ClassTopic.asp">System.Runtime.Remoting.Messaging.Header␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingHeaderHandler␍ClassTopic.asp">System.Runtime.Remoting.Messaging.HeaderHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingILogicalThreadAffinative␍ClassTopic.asp">System.Runtime.Remoting.Messaging.ILogicalThreadAffinative␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingIMessageCtrl␍ClassTopic.asp">System.Runtime.Remoting.Messaging.IMessageCtrl␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingIMessageSink␍ClassTopic.asp">System.Runtime.Remoting.Messaging.IMessageSink␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingIMethodCallMessage␍ClassTopic.asp">System.Runtime.Remoting.Messaging.IMethodCallMessage␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingInternalMessageWrapper␍ClassTopic.asp">System.Runtime.Remoting.Messaging.InternalMessageWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingIRemotingFormatter␍ClassTopic.asp">System.Runtime.Remoting.Messaging.IRemotingFormatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingLogicalCallContext␍ClassTopic.asp">System.Runtime.Remoting.Messaging.LogicalCallContext␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingMessageSurrogateFilter␍ClassTopic.asp">System.Runtime.Remoting.Messaging.MessageSurrogateFilter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingMethodCall␍ClassTopic.asp">System.Runtime.Remoting.Messaging.MethodCall␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingMethodCallMessageWrapper␍ClassTopic.asp">System.Runtime.Remoting.Messaging.MethodCallMessageWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingMethodResponse␍ClassTopic.asp">System.Runtime.Remoting.Messaging.MethodResponse␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingMethodReturnMessageWrapper␍ClassTopic.asp">System.Runtime.Remoting.Messaging.MethodReturnMessageWrapper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingOneWayAttribute␍ClassTopic.asp">System.Runtime.Remoting.Messaging.OneWayAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingRemotingSurrogateSelector␍ClassTopic.asp">System.Runtime.Remoting.Messaging.RemotingSurrogateSelector␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMessagingReturnMessage␍ClassTopic.asp">System.Runtime.Remoting.Messaging.ReturnMessage␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapAttribute␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapFieldAttribute␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapFieldAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapMethodAttribute␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapMethodAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapOption␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapOption␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapParameterAttribute␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapParameterAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataSoapTypeAttribute␍ClassTopic.asp">System.Runtime.Remoting.Metadata.SoapTypeAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001ISoapXsd␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.ISoapXsd␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapAnyUri␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapAnyUri␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapBase64Binary␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapBase64Binary␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapDate␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDate␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapDateTime␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDateTime␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapDay␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDay␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapDuration␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDuration␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapEntities␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapEntities␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapEntity␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapEntity␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapHexBinary␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapHexBinary␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapId␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapId␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapIdref␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapIdref␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapIdrefs␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapIdrefs␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapInteger␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapInteger␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapLanguage␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapLanguage␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapMonth␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapMonth␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapMonthDay␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapMonthDay␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapName␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapName␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNcName␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNcName␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNegativeInteger␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNegativeInteger␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNmtoken␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNmtoken␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNmtokens␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNmtokens␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNonNegativeInteger␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNonNegativeInteger␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNonPositiveInteger␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNonPositiveInteger␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNormalizedString␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNormalizedString␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapNotation␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNotation␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapPositiveInteger␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapPositiveInteger␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapQName␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapQName␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapTime␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapTime␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapToken␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapToken␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapYear␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapYear␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataW3cXsd2001SoapYearMonth␍ClassTopic.asp">System.Runtime.Remoting.Metadata.W3cXsd2001.SoapYearMonth␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingMetadataXmlFieldOrderOption␍ClassTopic.asp">System.Runtime.Remoting.Metadata.XmlFieldOrderOption␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingProxiesProxyAttribute␍ClassTopic.asp">System.Runtime.Remoting.Proxies.ProxyAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingProxiesRealProxy␍ClassTopic.asp">System.Runtime.Remoting.Proxies.RealProxy␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingRemotingConfiguration␍ClassTopic.asp">System.Runtime.Remoting.RemotingConfiguration␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingRemotingException␍ClassTopic.asp">System.Runtime.Remoting.RemotingException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingRemotingServices␍ClassTopic.asp">System.Runtime.Remoting.RemotingServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingRemotingTimeoutException␍ClassTopic.asp">System.Runtime.Remoting.RemotingTimeoutException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingServerException␍ClassTopic.asp">System.Runtime.Remoting.ServerException␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingServicesEnterpriseServicesHelper␍ClassTopic.asp">System.Runtime.Remoting.Services.EnterpriseServicesHelper␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingServicesITrackingHandler␍ClassTopic.asp">System.Runtime.Remoting.Services.ITrackingHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingServicesTrackingServices␍ClassTopic.asp">System.Runtime.Remoting.Services.TrackingServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingSoapServices␍ClassTopic.asp">System.Runtime.Remoting.SoapServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingTypeEntry␍ClassTopic.asp">System.Runtime.Remoting.TypeEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingWellKnownClientTypeEntry␍ClassTopic.asp">System.Runtime.Remoting.WellKnownClientTypeEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeRemotingWellKnownServiceTypeEntry␍ClassTopic.asp">System.Runtime.Remoting.WellKnownServiceTypeEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormatter␍ClassTopic.asp">System.Runtime.Serialization.Formatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormatterConverter␍ClassTopic.asp">System.Runtime.Serialization.FormatterConverter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersBinaryBinaryFormatter␍ClassTopic.asp">System.Runtime.Serialization.Formatters.Binary.BinaryFormatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersInternalRM␍ClassTopic.asp">System.Runtime.Serialization.Formatters.InternalRM␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersInternalST␍ClassTopic.asp">System.Runtime.Serialization.Formatters.InternalST␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersServerFault␍ClassTopic.asp">System.Runtime.Serialization.Formatters.ServerFault␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersSoapFault␍ClassTopic.asp">System.Runtime.Serialization.Formatters.SoapFault␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormattersSoapMessage␍ClassTopic.asp">System.Runtime.Serialization.Formatters.SoapMessage␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationFormatterServices␍ClassTopic.asp">System.Runtime.Serialization.FormatterServices␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationObjectIDGenerator␍ClassTopic.asp">System.Runtime.Serialization.ObjectIDGenerator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationObjectManager␍ClassTopic.asp">System.Runtime.Serialization.ObjectManager␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationSerializationBinder␍ClassTopic.asp">System.Runtime.Serialization.SerializationBinder␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationSerializationEntry␍ClassTopic.asp">System.Runtime.Serialization.SerializationEntry␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationSerializationInfoEnumerator␍ClassTopic.asp">System.Runtime.Serialization.SerializationInfoEnumerator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeSerializationSurrogateSelector␍ClassTopic.asp">System.Runtime.Serialization.SurrogateSelector␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemRuntimeArgumentHandle␍ClassTopic.asp">System.RuntimeArgumentHandle␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityAllowPartiallyTrustedCallersAttribute␍ClassTopic.asp">System.Security.AllowPartiallyTrustedCallersAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyCryptoConfig␍ClassTopic.asp">System.Security.Cryptography.CryptoConfig␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyDSACryptoServiceProvider␍ClassTopic.asp">System.Security.Cryptography.DSACryptoServiceProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyHMACSHA1␍ClassTopic.asp">System.Security.Cryptography.HMACSHA1␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyKeyedHashAlgorithm␍ClassTopic.asp">System.Security.Cryptography.KeyedHashAlgorithm␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyMACTripleDES␍ClassTopic.asp">System.Security.Cryptography.MACTripleDES␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyMaskGenerationMethod␍ClassTopic.asp">System.Security.Cryptography.MaskGenerationMethod␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyPasswordDeriveBytes␍ClassTopic.asp">System.Security.Cryptography.PasswordDeriveBytes␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyPKCS1MaskGenerationMethod␍ClassTopic.asp">System.Security.Cryptography.PKCS1MaskGenerationMethod␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRandomNumberGenerator␍ClassTopic.asp">System.Security.Cryptography.RandomNumberGenerator␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRC2␍ClassTopic.asp">System.Security.Cryptography.RC2␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRC2CryptoServiceProvider␍ClassTopic.asp">System.Security.Cryptography.RC2CryptoServiceProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRijndael␍ClassTopic.asp">System.Security.Cryptography.Rijndael␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRijndaelManaged␍ClassTopic.asp">System.Security.Cryptography.RijndaelManaged␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRNGCryptoServiceProvider␍ClassTopic.asp">System.Security.Cryptography.RNGCryptoServiceProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSA␍ClassTopic.asp">System.Security.Cryptography.RSA␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSACryptoServiceProvider␍ClassTopic.asp">System.Security.Cryptography.RSACryptoServiceProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAOAEPKeyExchangeDeformatter␍ClassTopic.asp">System.Security.Cryptography.RSAOAEPKeyExchangeDeformatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAOAEPKeyExchangeFormatter␍ClassTopic.asp">System.Security.Cryptography.RSAOAEPKeyExchangeFormatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAParameters␍ClassTopic.asp">System.Security.Cryptography.RSAParameters␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAPKCS1KeyExchangeDeformatter␍ClassTopic.asp">System.Security.Cryptography.RSAPKCS1KeyExchangeDeformatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAPKCS1KeyExchangeFormatter␍ClassTopic.asp">System.Security.Cryptography.RSAPKCS1KeyExchangeFormatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAPKCS1SignatureDeformatter␍ClassTopic.asp">System.Security.Cryptography.RSAPKCS1SignatureDeformatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyRSAPKCS1SignatureFormatter␍ClassTopic.asp">System.Security.Cryptography.RSAPKCS1SignatureFormatter␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographySHA1Managed␍ClassTopic.asp">System.Security.Cryptography.SHA1Managed␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographySHA384Managed␍ClassTopic.asp">System.Security.Cryptography.SHA384Managed␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographySHA512Managed␍ClassTopic.asp">System.Security.Cryptography.SHA512Managed␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyTripleDES␍ClassTopic.asp">System.Security.Cryptography.TripleDES␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyTripleDESCryptoServiceProvider␍ClassTopic.asp">System.Security.Cryptography.TripleDESCryptoServiceProvider␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityCryptographyX509CertificatesX509Certificate␍ClassTopic.asp">System.Security.Cryptography.X509Certificates.X509Certificate␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityNamedPermissionSet␍ClassTopic.asp">System.Security.NamedPermissionSet␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsEnvironmentPermission␍ClassTopic.asp">System.Security.Permissions.EnvironmentPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsFileDialogPermission␍ClassTopic.asp">System.Security.Permissions.FileDialogPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsFileDialogPermissionAccess␍ClassTopic.asp">System.Security.Permissions.FileDialogPermissionAccess␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsIsolatedStorageFilePermission␍ClassTopic.asp">System.Security.Permissions.IsolatedStorageFilePermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsIsolatedStoragePermissionAttribute␍ClassTopic.asp">System.Security.Permissions.IsolatedStoragePermissionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsPrincipalPermission␍ClassTopic.asp">System.Security.Permissions.PrincipalPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsPrincipalPermissionAttribute␍ClassTopic.asp">System.Security.Permissions.PrincipalPermissionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsPublisherIdentityPermission␍ClassTopic.asp">System.Security.Permissions.PublisherIdentityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsPublisherIdentityPermissionAttribute␍ClassTopic.asp">System.Security.Permissions.PublisherIdentityPermissionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsReflectionPermission␍ClassTopic.asp">System.Security.Permissions.ReflectionPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsRegistryPermission␍ClassTopic.asp">System.Security.Permissions.RegistryPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsSecurityPermission␍ClassTopic.asp">System.Security.Permissions.SecurityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsSecurityPermissionAttribute␍ClassTopic.asp">System.Security.Permissions.SecurityPermissionAttribute␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsSiteIdentityPermission␍ClassTopic.asp">System.Security.Permissions.SiteIdentityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsStrongNameIdentityPermission␍ClassTopic.asp">System.Security.Permissions.StrongNameIdentityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsStrongNamePublicKeyBlob␍ClassTopic.asp">System.Security.Permissions.StrongNamePublicKeyBlob␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsUIPermission␍ClassTopic.asp">System.Security.Permissions.UIPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsUrlIdentityPermission␍ClassTopic.asp">System.Security.Permissions.UrlIdentityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPermissionsZoneIdentityPermission␍ClassTopic.asp">System.Security.Permissions.ZoneIdentityPermission␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyAllMembershipCondition␍ClassTopic.asp">System.Security.Policy.AllMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyApplicationDirectory␍ClassTopic.asp">System.Security.Policy.ApplicationDirectory␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyApplicationDirectoryMembershipCondition␍ClassTopic.asp">System.Security.Policy.ApplicationDirectoryMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyFileCodeGroup␍ClassTopic.asp">System.Security.Policy.FileCodeGroup␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyFirstMatchCodeGroup␍ClassTopic.asp">System.Security.Policy.FirstMatchCodeGroup␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyHash␍ClassTopic.asp">System.Security.Policy.Hash␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyHashMembershipCondition␍ClassTopic.asp">System.Security.Policy.HashMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyNetCodeGroup␍ClassTopic.asp">System.Security.Policy.NetCodeGroup␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyPermissionRequestEvidence␍ClassTopic.asp">System.Security.Policy.PermissionRequestEvidence␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyPolicyStatement␍ClassTopic.asp">System.Security.Policy.PolicyStatement␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyPublisher␍ClassTopic.asp">System.Security.Policy.Publisher␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyPublisherMembershipCondition␍ClassTopic.asp">System.Security.Policy.PublisherMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicySite␍ClassTopic.asp">System.Security.Policy.Site␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicySiteMembershipCondition␍ClassTopic.asp">System.Security.Policy.SiteMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyStrongName␍ClassTopic.asp">System.Security.Policy.StrongName␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyStrongNameMembershipCondition␍ClassTopic.asp">System.Security.Policy.StrongNameMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyUnionCodeGroup␍ClassTopic.asp">System.Security.Policy.UnionCodeGroup␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyUrl␍ClassTopic.asp">System.Security.Policy.Url␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyUrlMembershipCondition␍ClassTopic.asp">System.Security.Policy.UrlMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyZone␍ClassTopic.asp">System.Security.Policy.Zone␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPolicyZoneMembershipCondition␍ClassTopic.asp">System.Security.Policy.ZoneMembershipCondition␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPrincipalWindowsIdentity␍ClassTopic.asp">System.Security.Principal.WindowsIdentity␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPrincipalWindowsImpersonationContext␍ClassTopic.asp">System.Security.Principal.WindowsImpersonationContext␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemSecurityPrincipalWindowsPrincipal␍ClassTopic.asp">System.Security.Principal.WindowsPrincipal␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingCompressedStack␍ClassTopic.asp">System.Threading.CompressedStack␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingIOCompletionCallback␍ClassTopic.asp">System.Threading.IOCompletionCallback␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingThreadStart␍ClassTopic.asp">System.Threading.ThreadStart␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingTimerCallback␍ClassTopic.asp">System.Threading.TimerCallback␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingWaitCallback␍ClassTopic.asp">System.Threading.WaitCallback␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemThreadingWaitOrTimerCallback␍ClassTopic.asp">System.Threading.WaitOrTimerCallback␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemTimeZone␍ClassTopic.asp">System.TimeZone␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemTypedReference␍ClassTopic.asp">System.TypedReference␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemUnhandledExceptionEventHandler␍ClassTopic.asp">System.UnhandledExceptionEventHandler␍</a><br>
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/cpref/html/frlrfSystemWeakReference␍ClassTopic.asp">System.WeakReference␍</a><br>
-22
View File
@@ -1,22 +0,0 @@
* Status of the various pieces of the class library
You can browse the status of the class library and see who has
registered to work on what parts of the system. These list
work-in-progress components currently.
Browse the current <a href="class-status/index.html">status</a>.
You can also download the XML <a
href="class-status/maintainers.xml">maintainers</a> file that
contains the actual maintainers list.
You can also download the master <a
href="class-status/class.xml">Class Status XML</a> file.
* List of not-implemented classes
These classes are only classes pending from the CORLIB assembly
other assemblies are not yet listed. We are going to be working
on an improved status display for pending classes.
Please bear with us in the meantime.
-105
View File
@@ -1,105 +0,0 @@
* Class Library Documentation
** Summary
<em>
[This documentation is in progress. Comments to
<a href="mailto:[email protected]">[email protected]</a>.]
</em>
While using inline XML tags to document classes is certainly
convenient, properly verbose documentation quickly overwhelms
the code, making the source quite difficult to read. Furthermore,
multilingual documentation makes for significant additional text.
With these considerations in mind, Mono will use external XML files
for type documentation, rather than documenting the source inline.
Several tools will be created for managing, verifying, generating,
and updating class library documentation, including:
<ul>
<li><a href="#docstub">docstub</a></li>
<li><a href="#docverify">docverify</a></li>
<li><a href="#docconv">docconv</a></li>
<li><a href="#docgen">docgen</a></li>
<li><a href="#monodoc">monodoc</a></li>
</ul>
** XML Documentation Files and Formats
*** Monodoc XML
This XML is similar to the XML documentation described in the
C# standard, with added tags for internationalization and a
slightly different structure. Documentation and a DTD/Schema
for Monodoc XML is forthcoming.
Monodoc XML does not contain any definitive type information,
and is only useful in conjunction with the Type definition for
whatever is being documented.
This XML could be generated by hand, by monostub, or by monodoc.
*** Assembly XML
This XML is generated by combining runtime type information for
an Assembly with a collection of Monodoc XML files. The resultant
XML (in the best of cases) contains both complete type and
documentary information for each type in the assembly.
This XML contains enough information to be transformed into
user-accessible documentation. It is likely that scripts
will be created to generate reference docs as HTML and other
formats.
Documentation and a DTD/Schema for Assembly XML is forthcoming.
** Documentation Tools
*** <a name="docstub">docstub</a>
Given a type name and an assembly, generates stub Monodoc XML
documentation for the type. Optionally, docstub can attempt to
populate initial documentation from an XML file in the format
published along with the ECMA standard.
*** <a name="docverify">docverify</a>
Given a Monodoc XML file and an assembly, verifies that the
documentation matches the compiled type. Checks signatures,
parameters, et cetera.
*** <a name="docconv">docconv</a>
Converts from Monodoc XML to the standard C# compiler-emitted
XML format and vice versa.
*** <a name="docgen">docgen</a>
Given an assembly and a collection of Monodoc XML files, creates
an Assembly XML file containing all documentation and type information
available. This output is suitable for transforming into user
documentation.
*** <a name="monodoc">monodoc</a>
A GUI tool for documentation and translation, this encapsulates
the capabilities of docstub, docverify, docconv, and docgen in
a friendly user interface. In addition, monodoc provides features
to ease translation, such as side-by-side editing and coverage
statistics.
** Status and Roadmap
*** 2 January 2002
As I write the initial version of this document, very little has been
implemented. Currently, there is a:
<ul>
<li>Preliminary version of the Monodoc XML format</li>
<li>Preliminary version of the docstub utility</li>
</ul>
Here's hoping for frequent, productive updates.
** Get Involved
Help us define Mono's documentation structure! Subscribe to
<a href="http://lists.ximian.com/mailman/listinfo/mono-docs-list">mono-docs-list</a>
and wade right in.
+3
View File
@@ -1,5 +1,8 @@
* Contacting the Mono team.
You can contact the Mono Team by sending e-mail to `<a
href="mailto:[email protected]">[email protected]</a>'.
You can contact the general forum of discussion by sending
e-mail to <a href="mailto:[email protected]">[email protected]</a>
+27 -13
View File
@@ -21,12 +21,16 @@
** To start contributing
To start developing classes or to contribute to the compiler,
you only need Windows and the .NET Framework 1.0 SDK. Please notice
you only need Windows and the .NET Beta 2 SDK. Please notice
that you do not need Visual Studio (although you can use it if
you want).
You can get it <a href="http://msdn.microsoft.com/downloads/default.asp?url=/downloads/sample.asp?url=/msdn-files/027/000/976/msdncompositedoc.xml&frame=true">here</a>
Alternatively you can use the <a
href="#compile-service">compilation service</a> that was setup by
Derek.
** Bug reporting
If you find bugs in Mono, please make sure you enter a bug
@@ -40,21 +44,32 @@
A few smaller tasks are here, dropped in no particular order:
<ul>
* <b>MCS compilation process:</b> Currently MCS does
not build with a single `make' command. This should
be fixed (this being part of the `Joel Test' that
software has to pass).
* <b>Mono/doc and web site:</b> They need to be
packaged up in the official `distribution'
* Adding serialization support to all the classes.
We have many classes, but we have not implemented in
many cases the serialization and re-incarnation support in
them (this is pretty straight forward code, and simple,
but there is a lot to be done here).
</ul>
* Emacs support
<a name="compile-service">
* C# Compilation Service
Brad Merryl's C# mode for Emacs is available: <a
href="http://www.cybercom.net/~zbrad/DotNet/Emacs/">http://www.cybercom.net/~zbrad/DotNet/Emacs/</a>
If you are working on a class for the Mono project, but do not
have a C# compiler available or a Windows machine to run the
.NET SDK, you can use the compilation service that Derek
Holden setup in the following URL: <a
href="http://toilet.2y.net:8080/">http://toilet.2y.net:8080/</a>
The service will let you compile a source file and get back a
list of errors in the class file. You will not be able to run
the code, but at least you can get some code written that will
help us further down the line.
You can contact (and thank) Derek for this service by sending
him nice email at <a
href="mailto:[email protected]">[email protected]</a>
* Books on C# and DotNet.
@@ -64,12 +79,11 @@
* Dotnet Resources (<a href="http://www.dotnetexperts.com/resources/">
http://www.dotnetexperts.com/resources</a>)
* O'Reilly C# Essentials (<a href="http://www.oreilly.com/catalog/csharpess/">
* O'Really C# Essentials (<a href="http://www.oreilly.com/catalog/csharpess/">
http://www.oreally.com/catalog/csharpess</a>)
* O'Really .NET Essentials (<a href="http://www.oreilly.com/catalog/dotnetfrmess/">
http://www.oreally.com/catalog/dotnetfrmess</a>)
</ul>
* Special note
+17 -2
View File
@@ -31,5 +31,20 @@
** Class Library documentation
We are moving to a new setup for documenting the class libraries,
you can read about it <a href="classlib-doc.html">here</a>
When contributing to the Class Library effort, please use the
inline XML documentation tags to document your classes so we
can automatically generate the documentation from the class
libraries.
If you provide examples, please do not embed them into the
source code, as that will make the source code harder to read
and maintain. Instead, put examples for your code into a
subdirectory of the class libraries. Make your sample
code a full standalone application that people can compile.
Ideally the Mono documentation browser will let you edit, modify
and run the sample programs.
+88 -175
View File
@@ -14,15 +14,13 @@
GNU make tools. These may be obtained for the Windows
environment from <a href="http://www.cygwin.com">cygwin.com</a>.
You will also need to get GLIB and <a
href="http://www.freedesktop.org/software/pkgconfig">pkgconfig</a>.
You can get Glib 1.3 from: <a
href="ftp://ftp.gtk.org/pub/gtk/v1.3/glib-1.3.12.tar.gz">ftp://ftp.gtk.org/pub/gtk/v1.3/glib-1.3.12.tar.gz</a>.
You will also need to get GLIB, from: <a
href="ftp://ftp.gtk.org/pub/gtk/v1.2/glib-1.2.10.tar.gz">ftp://ftp.gtk.org/pub/gtk/v1.2/glib-1.2.10.tar.gz</a>
<a name="sources">
** Sources
We provide both <a href="#packaged">packaged and tested</a>
We provide both <a href="packaged">packaged and tested</a>
tarballs (those are known to compile and pass `make
distcheck') as well as <a href="snapshots">daily snapshots</a> done
at 10pm Boston Time
@@ -38,188 +36,26 @@
should. They should give you a window to see what we are up to.
You might also want to track our development using the <a
href="http://lists.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list</a>
href="http://mail.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list</a>
mailing list.
<a name="packaged">
At this point, you should look into using one of the <a
href="snapshots">nightly snapshots</a> or getting the
sources from <a href="anoncvs.html">Anonymous CVS</a>
as significant work has gone into Mono since our last stable
release in September.
We will be making a packaged release sometime in January.
<a name="install">
** Installing the software
To install and work on the compiler and the class libraries,
follow these instructions:
*** Released and tested packages
<ul>
* Install <a href="http://www.cygwin.com">CygWin</a> first.
* Untar the MCS distribution (see below for information).
* Go into the MCS directory and type `make windows'
</ul>
To compile the mono runtime on windows:
<ul>
* Install <a href="http://www.cygwin.com">CygWin</a> first.
* Install in /usr/local in your Cygwin world the following
packages (compiled by <a href="mailto:[email protected]">Alex Graveley</a>):
<ul>
<li><a href="http://primates.ximian.com/~alex/glib-1.3.12-win32-native.zip">
http://primates.ximian.com/~alex/glib-1.3.12-win32-native.zip</a>
<li><a href="http://primates.ximian.com/~alex/pkg-config-0.8-cygwin.zip">
http://primates.ximian.com/~alex/pkg-config-0.8-cygwin.zip</a>
</ul>
These already contain libiconv linked statically.
</ul>
If you want to compile things yourself you will need to compile and install:
<ul>
* Install <a href="http://www.freedesktop.org/software/pkgconfig">pkgconfig</a>.
(I had to change line 674 of
<nobr><tt>pkgconfig-0.8.0/glib-1.2.8/gstrfuncs.c</tt></nobr> from
<nobr><tt>extern char *strsignal (int sig);</tt></nobr> to
<nobr><tt>extern const char *strsignal (int sig);</tt></nobr>.
<pre>
tar xzvf pkgconfig-0.8.0.tar.gz
cd pkgconfig-0.8.0
./configure --prefix=/usr
make
make install
</pre>
* Install <a href="ftp://ftp.gnu.org/pub/gnu/libiconv/libiconv-1.7.tar.gz">libiconv</a>.
If you dont have MS Visual C/C++ 4.0 or 5.0 or 6.0 you can also try
the binary package available at <a
href="http://www.gimp.org/win32/libiconv-dev-20001007.zip">here</a>.
* Install the newest glib (1.3 version). You can download it
from <a href="ftp://ftp.gtk.org/pub/gtk/v1.3/glib-1.3.12.tar.gz">here</a>.
* Unpack the mono distribution:
<pre>
tar xzvf mono-XXX.tar.gz
</pre>
* Configure, compile and install:
For official releases, use:
<pre>
./configure --prefix=//c/mono
make
make install
</pre>
For Snapshots, or CVS trees, use:
<pre>
./autogen.sh --prefix=//c/mono
make
make install
</pre>
* Some people observed problems with autoconf 2.52. Installing
autoconf 2.13 helped in those cases (don't forget to do a
`make maintainer-clean' after the update)
</ul>
To compile the mono runtime on Unix:
<ul>
* Install <a href="http://www.freedesktop.org/software/pkgconfig">pkgconfig</a>.
* Install newest glib (cvs head version). You can download it
from <a href="ftp://ftp.gtk.org/pub/gtk/v1.3/glib-1.3.12.tar.gz">here</a>. Maybe you need
to modify your ACLOCAL_FLAGS for pkgconfig, for example if
your install prefix is /usr/local:
<pre>
export ACLOCAL_FLAGS="-I /usr/local/share/aclocal"
</pre>
* Download the mono distribution
* Unpack the mono distribution:
<pre>
tar xzvf mono-XXX.tar.gz
</pre>
* Configure, compile and install:
For official releases:
<pre>
./configure
make
make install
</pre>
For snapshots and CVS trees:
<pre>
./autogen.sh
make
make install
</pre>
</ul>
<a name="oldreleases">
*** Old releases.
<ul>
<a name="sep-18">
<b>September 18, 2001</b>
<ul>
* <a href="archive/mono-0.7.tar.gz">mono-0.7.tar.gz</a>: Mono Runtime 0.7 release.
* <a href="archive/mcs-0.7.tar.gz">mcs-0.7.tar.gz</a>: Mono Compiler and Classes 0.7 release.
* <a href="archive/mono-0.7">Release Notes.</a>
</ul>
<a name="august-22">
<b>August 22, 2001</b>
<ul>
* <a href="archive/mono-0.6.tar.gz">mono-0.6.tar.gz</a>: Mono Runtime 0.6 release.
* <a href="archive/mono-compiler-classes-0.6.tar.gz">mono-compiler-classes-0.6.tar.gz</a>: Mono Compiler and Classes 0.6 release.
* <a href="archive/mono-0.6">Release Notes.</a>
</ul>
<a name="july-29">
<b>July 29, 2001</b>
<ul>
* <a href="archive/mono-0.5.tar.gz">mono-0.5.tar.gz</a>: Mono Runtime 0.5 release.
* <a href="archive/mono-0.5">Release Notes.</a>
* <a href="archive/mono-0.5.tar.gz">mono-0.5.tar.gz</a>: Mono Runtime 0.5 release
* <a href="archive/mono-0.5">Release Notes</a>
</ul>
<a name="july-22">
<b>July 22, 2001</b>
<ul>
* <a href="archive/mcs-22-Jul-2001.tar.gz">mcs-22-Jul-2001.tar.gz</a>: CVS snapshot.
* <a href="archive/mcs-Jul-22-Jul-19-2001.tar.gz">Differences since 19</a>: CVS snapshot.
* <a href="archive/mcs-22">Release Notes.</a>
* <a href="archive/mcs-22">Release Notes</a>
</ul>
<a name="july-19">
@@ -227,7 +63,7 @@ make install
<ul>
* <a href="archive/mcs-19-Jul-2001.tar.gz">mcs-19-Jul-2001.tar.gz</a>: CVS snapshot.
* <a href="archive/mcs-Jul-17-Jul-19-2001.tar.gz">Differences since 17</a>: CVS snapshot.
* <a href="archive/mcs-19">Release Notes.</a>
* <a href="archive/mcs-19">Release Notes</a>
</ul>
<a name="july-17">
@@ -235,7 +71,7 @@ make install
<ul>
* <a href="archive/mcs-17-Jul-2001.tar.gz">mcs-17-Jul-2001.tar.gz</a>: CVS snapshot.
* <a href="archive/mcs-Jul-15-Jul-17-2001.tar.gz">Differences since 15</a>: CVS snapshot.
* <a href="archive/mcs-17">Release Notes.</a>
* <a href="archive/mcs-17">Release Notes</a>
</ul>
<a name="july-15">
@@ -243,7 +79,7 @@ make install
<ul>
* <a href="archive/mcs-15-Jul-2001.tar.gz">mcs-15-Jul-2001.tar.gz</a>: CVS snapshot.
* <a href="archive/mono-0.4.tar.gz">mono-0.4.tar.gz</a>: Packaged Source Code.
* <a href="archive/mono-0.4">Release Notes.</a>
* <a href="archive/mono-0.4">Release Notes</a>
</ul>
<a name="july-14">
@@ -262,3 +98,80 @@ make install
</ul>
</ul>
<a name="install">
** Installing the software
To install and work on the compiler and the class libraries,
follow these instructions:
<ul>
* Install <a href="http://www.cygwin.com">CygWin</a> first.
* Untar the MCS distribution (see below for information).
* Go into the MCS directory and type `make windows'
</ul>
To compile the mono runtime on windows:
<ul>
* Install <a href="http://www.cygwin.com">CygWin</a> first.
* Once installed, in a terminal window or a cygwin
window (a shortcut should be on your desktop), untar the glib
distribution:
<pre>
tar xzvf glib-1.2.10.tar.gz
</pre>
* Configure, compile and install glib, like this:
<pre>
./configure --prefix=/usr
make
make install
</pre>
* Unpack the mono distribution:
<pre>
tar xzvf mono-XXX.tar.gz
</pre>
* Configure, compile and install:
<pre>
./configure --prefix=//c/mono
make
make install
</pre>
</ul>
To compile the mono runtime on Unix:
<ul>
* Download the mono distribution
* Unpack the mono distribution:
<pre>
tar xzvf mono-XXX.tar.gz
</pre>
* Configure, compile and install:
<pre>
./configure
make
make install
</pre>
</ul>
** CVS
We are trying to figure out where to put our CVS repository.
We are debating between the GNOME CVS or SourceForge. Watch
this spot.
+431 -372
View File
@@ -1,42 +1,18 @@
<a href="#basics">Basics</a><br>
<a href="#ximian">The Ximian Role in the Mono project</a><br>
<a href="#gnome">Mono and GNOME</a><br>
<a href="#gui">Building GUI applications with Mono</a><br>
<a href="#msft">Mono and Microsoft</a><br>
<a href="#pnpproject">Mono and the Portable.NET Project</a><br>
<a href="#webservices">Web Services</a><br>
<a href="#devel">Development Tools and Issues</a><br>
<a href="#java">Mono and Java</a><br>
<a href="#extending">Extending Mono</a><br>
<a href="#portability">Portability</a><br>
<a href="#reuse">Reusing Existing Code</a><br>
<a href="#gcc">Mono and GCC</a><br>
<a href="#performance">Performance</a><br>
<a href="#licensing">Licensing</a><br>
<a href="#etc">Miscellaneous Questions</a><br>
<a name="basics"></a>
** Basics
Q: What is Mono exactly?
Q: Is Mono the same as Microsoft's .NET initiative?
A: The Mono Project is an open development initiative sponsored by
Ximian that is working to develop an open source, Linux-based
version of the Microsoft .NET development platform. Its objective
is to enable Linux developers to build and
deploy cross-platform .NET Applications. The project will
implement various technologies developed by Microsoft that have now
been submitted to the ECMA for standardization.
A: It is not.
Q: What is the difference between Mono and the .NET Initiative?
.NET is a company-wide initiative at Microsoft that
encompasses many different areas. The .NET development framework,
Passport, Biztalk, new server products, and anything that is
remotely connected to .NET gets the ".NET-stamping" treatment.
Some components of Microsoft's .NET initiative have been announced
and some others are in the works.
A: The ".NET Initiative" is a somewhat nebulous company-wide effort by
Microsoft, one part of which is a cross-platform development
framework. Mono is an implementation of the development framework,
but not an implementation of anything else related to the .NET
Initiative, such as Passport, software-as-a-service, or
corporate rebranding.
Mono is a project to implement several technologies developed by
Microsoft that have now been submitted to the ECMA Standards Body.
Q: What technologies are included in Mono?
@@ -54,18 +30,20 @@ A: Mono contains a number of components useful for building new
* A compiler for the C# language. In the future we
might work on other compilers that target the Common
Language Runtime.
Windows has compilers that target the
virtual machine from
<a href="http://msdn.microsoft.com/net/thirdparty/default.asp#lang">a
number of languages:</a> Managed C++, Java Script,
Eiffel, Component Pascal, APL, Cobol, Oberon, Perl,
Python, Scheme, Smalltalk, Standard ML, Haskell,
Mercury and Oberon.
</ul>
Windows has compilers that target the virtual machine from <a
href="http://msdn.microsoft.com/net/thirdparty/default.asp#lang">a
number of languages:</a> Managed C++, Java Script, Eiffel,
Component Pascal, APL, Cobol, Oberon, Perl, Python, Scheme,
Smalltalk, Standard ML, Haskell, Mercury and Oberon.
The CLR and the Common Type System (CTS) enables applications and
libraries to be written in a collection of different languages that
target the byte code
libraries to be written in a number of languages. Classes and
methods created in one language can be used from a different
language.
This means for example that if you define a class to do algebraic
manipulation in C#, that class can be reused from any other
@@ -87,58 +65,48 @@ Q: Will you implement the .NET Framework SDK class libraries?
A: Yes, we will be implementing the APIs of the .NET Framework SDK
class libraries.
Q: Will you offer an ECMA-compliant set of class libraries?
Q: What does Mono stand for?
A: Eventually we will. Our current focus is on interoperating
with the Microsoft SDK, but we will also offer an ECMA compliant
set of libraries.
A: Mono is the word for `Monkey' in Spanish. We like monkeys.
Q: What does the name "Mono" mean?
A: Mono is the word for `monkey' in Spanish. We like monkeys.
Q: Is Mono usable?
A: The JIT engine is usable on Intel x86 machines. An interpreter
can be used on other non-Intel x86 machines.
The class libraries are not yet mature enough to run real applications,
but if you are interested in trying out Mono, you can definetly start
testing things out as many programs run.
The C# compiler has made significant progress, it can even compile
itself now, but it can not yet be ran with our class libraries as
they are missing some features.
It only means a number of other things: monochromatic (hence the
gray theme used in the Web site).
Q: When will you ship it?
A: It is premature to target a shipdate for the code, but we
anticipate that it will be available some time in the middle of
2002.
A: We do not know when the code will be shipped. The more
contributions we get to the project, the sooner it will ship.
A rough estimate is that we might be able to run our C# compiler on
Linux by the end of the year. That means running the Windows
Executable generated by a Microsoft .NET compiler on the Linux
platform.
We expect that doing GUI applications will require more work on the
class libraries. That could take another six months.
Q: How can I contribute?
A: Check the <a href="contributing.html">contributing</a> section.
Q: Aren't you just copying someone else's work?
Q: You guys should innovate instead of copying.
A: We are interested in providing the best tools for programmers to
develop applications for Free Operating Systems. We also want to help
provide the interoperability that will allow those systems to fit in
with other standards.
A: In this particular case, we see a clear advantage in the platform
and we are interested in using the features of the CLI on open source systems.
For more background, read the <a href="http://www.go-mono.com/rationale.html">Mono
Project white paper</a>.
the project.
We have decided that we should spend our limited resources towards
implementing an existing specification instead of designing and
implementing our own.
Designing and implementing our own would be possible, but it doesn't make
sense to do that just because the specification comes from a
proprietary vendor.
<a name="ximian"></a>
** The Ximian Role in the Mono Project
** Ximian
Q: Why is Ximian working on .NET?
A: Ximian is interested in providing the best tools for programmers to
A: We are interested in providing the best tools for programmers to
develop applications for Free Operating Systems.
For more information, read the project <a
@@ -146,150 +114,90 @@ A: Ximian is interested in providing the best tools for programmers to
Q: Will Ximian be able to take on a project of this size?
A: Of course not. Ximian a supporter of the Mono project, but the only way
to implement something of this size is for the entire free software
community to get involved. Visit the <a href="contributing.html">contributing</a>
page if you'd like to help out.
A: Ximian will not be able to taken on the whole project on its own.
Mono will be a free software/open source community project, that is
the only way we can hope to implement something of this size. You
can <a href="contributing.html">contribute</a> to this effort.
Q: What pieces will Ximian be working on?
A: We will devote most of our resources to work on the pieces which are
on the critical path to release a development and execution
environment. Once the project is at a stage where it is useful in
the real world, it will achieve a critical mass of developers to
improve it further.
A: We will focus on building a development and execution
environment.
Q: Will Ximian offer Mono commercially?
The idea is to get Mono to a state of that would allow
third parties to actually be able to use it real-world development.
A: When Mono is ready to be shipped Ximian will offer a commercial
support and services for Mono.
Q: Why does Ximian even care?
Q: Will you wait until Mono is finished?
A: We like the features that the CLI and its related technologies
bring to the table. An exciting addition to the developer toolkit.
The goal of Mono is to bring this technology to non-Windows
platforms (although we hope Mono will also run on Windows, for
debugging and comparative purposes).
A: Mono will ship on various stages as they mature. Some people
require only a subset of the technologies, those will ship first.
More advanced features will take more time to develop. A support
timeline will be available in June 2002.
** Licensing
<a name="gnome"></a>
** Mono and GNOME
Q: Will I be able to write proprietary applications that run with
Mono?
Q: How is Mono related to GNOME?
A: Yes. The licensing scheme is planned to allow proprietary
developers to write applications with Mono.
Q: What license is Mono on?
A: In a number of ways. This project was born out of the need of
providing improved tools for the GNOME community, and will use
existing components that have been developed for GNOME when they
are available. For example, we plan to use Gtk+ and Libart to
implement Winforms and the Drawing2D API and are considering
GObject support.
A: The C# Compiler is released under the terms of the GPL. The class
libraries will be under the LGPL or the GPL with a special
exception. The runtime libraries are under the LGPL.
Q: Has the GNOME Foundation or the GNOME team adopted Mono?
Since the LGPL is not suitable for embedded systems development, we
are also licensing the libraries under the GPL with the following exception:
A: Mono is too new to be adopted by those groups. We hope that the
tools that we will provide will be adopted by free software
programmers including the GNOME Foundation members and the GNOME
project generally.
If you link this library against your own program, then you do not
need to release the source code for that program. However, any
changes that you make to the library itself, or to any native
methods upon which the library relies, must be re-distributed in
accordance with the terms of the GPL.
Q: Should GNOME programmers switch over to Mono now?
This is similar in spirit to <a
href="http://www.gnu.org/software/classpath/classpath.html">GNU
Classpath.</a>
A: It is still far to early for discussions of "switching over." No
pieces of Mono will be ready within the next six months, and a
complete implementation is roughly one year away.
Q: But in Object Oriented Programming I need to subclass your library
functions, does that mean that I am making modifications to your
library and hence I would have to distribute my sources under the
LGPL?
We encourage GNOME developers to continue using the existing tools,
libraries and components. Improvements made to GNOME will have an
impact on Mono, as they would be the "backend" for various classes.
A: No. Object Oriented Programming in the class library is a well
understood interface barrier, so you can actually develop
proprietary applications with the Mono libraries.
Q: Will Mono include compatibility with Bonobo components? What is the
relationship between Mono and Bonobo?
Q: Will you accept code under the XXX License?
A: Yes, we will provide a set of classes for implementing and using
Bonobo components from within Mono. Mono should allow you to write
Bonobo components more easily, just like .NET on Windows allows you
to export .NET components to COM.
A: If the XXX License is compatible with the license we use in that
specific piece of code, then yes. If you want to use the BSD license, make
sure you use the BSD license without the advertisement clause (The
`Ousterhout License').
Q: Does Mono depend on GNOME?
** Mono and .NET
A: No, Mono does not depend on GNOME. We use a few packages produced by
the GNOME team like the `glib' library.
Q: But will I be able to build GNOME applications?
A: Yes, we will enable people to write GNOME applications using Mono.
<a name="gui"></a>
** GUI applications
Q: Will Mono enable GUI applications to be authored?
A: Yes, you will be able to build GUI applications. Indeed, that is our
main focus. We will provide both the Windows.Forms API and the Gtk# AP(I.
Q: What is the difference between Gtk# and System.Windows.Forms?
A: Gtk# is a set of bindings for the Gtk+ toolkit for C# (and other
CIL-enabled languages). System.Windows.Forms is an API defined
by Microsoft to build GUI applications.
Q: Will I be able to run my smart clients on systems powered by Mono?
A: As long as your applications are 100% .NET and do not make use
of P/Invoke to call Win32 functions, your smart client applications
will run on Mono platforms.
<a name="msft"></a>
** Mono and Microsoft
Q: Is Microsoft helping Ximian with this project?
A: No. Ximian CTO Miguel de Icaza had a friendly conversation with
Microsoft software architect David Stutz, but that's about the
extent of the contact. Microsoft is interested in other
implementations of .NET and are willing to help make the ECMA spec
more accurate for this purpose.
Ximian representatives have also spoken with Sam Ruby at the ECMA
TG3 committee to discuss the same issues.
Q: Is Microsoft or Corel paying Ximian to do this?
A: No.
Q: Do you fear that Microsoft will change the spec and render Mono
useless?
A: No. Microsoft proved with the CLI and the C# language that it was
possible to create a powerful foundation for many languages to
interoperate. We will always have that.
Even if changes happened in the platform which were undocumented,
the existing platform would a value on its own.
Q: Are you writing Mono from the ECMA specs?
A: Yes, we are writing them from the ECMA specs and the published
materials in print about .NET.
Q: If my applications use Mono, will I have to pay a service fee?
Q: If applications use Mono, does that mean that I have to pay a service fee?
A: No. Mono is not related to Microsoft's initiative of
software-as-a-service.
software-as-a-service.
Q: Is the Mono Project is related to the Microsoft Hailstorm effort? Is
Ximian endorsing Hailstorm?
Q: If you implement .NET, will I depend on Microsoft Passport to run my software?
A: No. The Mono Project is focused on providing a compatible set of
tools for the Microsoft .NET development platform. It does not
address, require, or otherwise endorse the MS Passport-based
Hailstorm single sign-on system that is part of Windows XP and
other services.
A: No. The .NET Framework is a runtime infrastructure and collection
of class libraries. Passport may be required to access certain web
services written for that framework, but only if the programmer
chooses Passport as the authentication mechanism.
Q: Will Mono or .NET applications depend on Microsoft Passport?
Q: Is .NET just a bunch of marketing slogans?
A: No. MS Passport is unrelated to running .NET compatible applications
produced with the Mono tools. The only thing you will need is a
just-in-time compiler (JIT).
A: Although the `.NET initiative' is still quite nebulous, The .NET Framework
has been available for some time. Mono is not an implementation of the .NET
initiative, just the development framework.
Q: What is a 100% .NET application?
@@ -298,83 +206,149 @@ A: A `100% .NET application' is one that only uses the APIs defined
applications would in theory run unmodified on Windows, Linux,
HP-UX, Solaris, MacOS X and others.
Q: If Microsoft will release a port of their .NET platform under the
`Shared Source' license, why should I bother with anything else?
Q: But Microsoft will release a port of the real thing under the
`Shared Source' license, why bother with anything else?
A: The Shared Source implementation will be expensive and its uses
will be tightly restricted, especially for commercial use. We are
working towards an implementation that will grant a number of
important rights to recipients: use for any purpose,
A: The Shared Source implementation will not be usable for commercial
purposes. We are working towards an implementation that will grant
a number of rights to recipients: use for any purpose,
redistribution, modification, and redistribution of modifications.
This is what we call <a
href="http://www.gnu.org/philosophy/free-sw.html">Free Software</a>
Q: Is Mono a free implementation of Passport?
** Passport
A: No. Mono is just a runtime, a compiler and a set of class
libraries.
Q: Is this a free implementation of Passport?
Q: Will the System.Web.Security.PassportIdentity class mean
A: No. Passport is part of Microsoft's Hailstorm initiative. Mono
is just a runtime, a compiler and a set of class libraries.
Q: Will the System.Web.Security.PassportIdentity class, mean
that my software will depend on Passport?
A: No. Applications may use that API to contact a Passport site, but
are not required to do so.
A: No. That just means that applications might use that API to
contact a Passport site.
As long as your application does not use Passport, you will not
need Passport.
It might even be possible to implement that class with
a set of dummy functions, or use an alternate Passport implementation.
We do not know at this time whether the Passport protocol is
documented and whether we will be able to talk to
passport.com
Q: But that must mean that you are tied to Passport!
A: All the contrary. The implementation could keep the interface (for
the sake of simplicity, lets say it implements the method `Login'
and `GetUserName').
We could implement `Login' and `GetUserName' by talking to XNS or
any other decentralized systems. Or any other system that the
industry standarizes on.
Q: What is your opinion?
A: You can read my personal <a href="passport.html">opinion on
passport</a>.
Q: Will Mono running on Linux make Passport available for Linux?
A: No. However, the Passport toolkit for Linux-based web servers is
available from Microsoft.
A: The Passport toolkit for Linux-based web servers is available from
Microsoft.
Again, Mono has nothing to do with Passport.
** Mono and Windows
Q: Will Mono allow me to run Microsoft Office on Linux?
A: No, it will not. Microsoft Office is a Windows application. To
learn more about running Windows applications on Intel Unix systems
refer to <a href="http://www.winehq.com">the Wine Project</a>.
run Windows applications on Intel Unix systems refer to <a
href="http://www.winehq.com">the Wine Project</a>
<a name="pnpproject"></a>
** Mono and Portable.NET
** GNOME
Q: What are the differences between Mono and Portable.NET?
Q: How is this related to GNOME?
A: Most of Mono is being written using C#, with only
a few parts written in C (The JIT engine, the runtime, the
interfaces to the garbage collection system). The Portable.NET
Project is using C for development.
A: In a number of ways:
* Mono will use existing
components that have been developed for GNOME when it makes
sense. For example on X systems, we will use Gtk+ and
Libart to implement Winforms and the Drawing2D API.
For database access, we will use LibGDA (not really
depending on GNOME, but related to).
* This project was born out of the need of providing improved
tools for the GNOME community.
* We would like to add support to our CLR implementation to
deal with GObjects (in GNOME 1.x, they are called
GtkObjects), and allow Mono developers to provide GObjects
or use and extend existing GObjects.
Q: Has the GNOME Foundation or the GNOME team adopted Mono?
A: Mono is too new to be adopted by those groups. We hope that the
tools that we will provide will be adopted by free software
programmers including the GNOME Foundation members and the GNOME
project generally.
Q: Should GNOME programmers switch over to Mono?
A: Mono will not be ready even within the next six months, and a
complete implementation is probably one year away.
We encourage GNOME developers to continue using the existing tools,
libraries and components. Improvements made to GNOME will have an
impact on Mono, as they will provide the "backend" for various
classes.
Q: Will Mono include compatibility with Bonobo components?
A: Yes, we will provide a set of classes for implementing and using
Bonobo components from within Mono.
Q: Does Mono replace Bonobo?
A: Bonobo is very focused on cross-application component reuse. Mono
will provide a Bonobo framework to allow you to develop Bonobo
components and use Bonobo components on Unix.
Mono should allow you to write Bonobo components more easily, just
like .NET on Windows allows you to export .NET components to COM.
** Mono and the Web
Q: Is Mono a way of running Java applets?
A: No.
<a name="webservices"></a>
** Web Services
Q: How is Mono related to Web Services?
Q: Is Mono just a new way of writing Web Services?
A: Mono is only related to Web Services in that it will implement the
same set of classes that have been authored in the .NET Framework
to simplify and streamline the process of building Web Services.
A: No.
But most importantly, Mono is an Open Source implementation of the
.NET Framework.
Q: Can I author Web Services with Mono?
A: You will be able to write Web Services on .NET that run on Mono and
viceversa.
Q: If Mono implements the SDK classes, will I be able to write and
execute .NET Web Services with it?
Q: If this implements the SDK classes, will I be able to write and
execute .NET Web Services with this?
A: Yes. When the project is finished, you will be able to use the
same technologies that are available through the .NET Framework SDK
on Windows to write Web Services.
A: Yes, you will.
Q: What about Soup? Can I use Soup without Mono?
When the project is finished, you will be able to use the same
technologies that are available through the .NET Framework SDK on
Windows to write Web Services.
A: Soup is a library for GNOME applications to create SOAP servers and
SOAP clients, and can be used without Mono. You can browse the
source code for soup using <a
href="http://cvs.gnome.org/bonsai">GNOME's Bonsai</a>.
Q: What about Soup?
A: Soup is a library for GNOME applications to create SOAP server and
SOAP clients. You can browse the source code for soup using <a
href="http://cvs.gnome.org/bonsai">GNOME's Bonsai</a>
Q: Can I use CORBA?
@@ -392,8 +366,7 @@ Q: Can I serialize my objects to other things other than XML?
A: Yes, although the serializing tools have not yet been planned, and
you would probably have to implement them yourself.
<a name="devel"></a>
** Development Tools and Issues
** Development Tools
Q: Will it be possible to use the CLI features without using bytecodes
or the JIT?
@@ -406,34 +379,14 @@ A: Yes. The CLI engine will be made available as a shared library.
Q: Will you have new development tools?
A: With any luck, Free Software enthusiasts will contribute tools to
A: Hopefully Free Software enthusiasts will contribute tools to
improve the developer environment. These tools could be developed
initially using the Microsoft implementation of the CLI and then
initially using Microsoft implementation of the CLI and then
executed later with Mono.
Q: What kind of rules make the Common Intermediate Language useful for
JITers?
A: The main rule is that the stack in the CLI is not a general purpose
stack. You are not allowed to use it for other purposes than
computing values and passing arguments to functions or return
values.
At any given call or return instruction, the types on the stack
have to be the same independently of the flow of execution of your
code.
Q: Is it true that the CIL is ideal for JITing and not efficient for
interpreters?
A: The CIL is better suited to be JITed than JVM byte codes, but you
can interpret them as trivially as you can interpret JVM byte
codes.
<a name="java">
** Mono and Java
Q: Why don't you use Java? After all, there are many languages that
Q: What about using Java? After all there are many languages that
target the Java VM.
A: You can get very good tools for doing Java development on free
@@ -452,36 +405,32 @@ A: You can get very good tools for doing Java development on free
wide variety of programming languages, and has a set of rules
designed to be optimal for JITers.
Q: What kind of rules make the Common Intermediate Language useful for
JITers?
A: The main rule is that the stack in the CLI is not a general purpose
stack. You are not allowed to use it for other purposes than
computing values and passing arguments to functions or return
values.
At any given call or return instruction, the types on the stack
have to be the same independently of the flow of execution of your
code.
Q: I heard that the CIL is ideal for JITing and not efficient for
interpreters, is this the case?
A: The CIL is better suited to be JITed than JVM byte codes, but you
can interpret them as trivially as you can interpret JVM byte
codes.
Q: Could Java target the CLI?
A: Yes, Java could target the CLI. We have details on a <a
href="ideas.html#guavac">project</a> that someone could take on to
make this happen.
make this happen.
Microsoft has an implemenation of the Java language called J# that
can target the CIL execution engine.
Q: Is it possible to write a JVM byte code to CIL convertor?
A: Yes, it is possible. Here are a few starting point:
<ul>
* A byte code representation is really a flattened forest of
trees. Look at the Mono JIT engine to see how we compute
the basic blocks (this is used to figure out the "trees").
The forest is just an array of trees.
Indeed, run the JIT engine with -d (mono -d prog.exe) and
you will see how these trees look like.
You will have to do something similar for Java.
* Each "forest of trees" has a meaning. This meaning can now
be translated into the equivalent "meaning" in CLR-land.
</ul>
<a name="extending"></a>
** Extending Mono
Q: Would you allow other classes other than those in the
@@ -500,18 +449,7 @@ A: Yes. The Microsoft class collection is very big, but it is by no
For more information on extending Mono, see our <a
href="ideas.html">ideas</a> page.
Q: Do you plan to Embrace and Extend .NET?
A: Embracing a good technology is good. Extending technologies in
incompatible ways is bad for the users, so we do not plan on
extending the technologies.
If you have innovative ideas, and want to create new classes, we
encourage you to make those classes operate correctly well in both
Mono and .NET
<a name="portability"></a>
** Portability
** Mono and portability
Q: Will Mono only work on Linux?
@@ -525,23 +463,9 @@ A: Our main intention at Ximian is to be able to develop GNOME
applications with Mono, but if you are interested in providing a
port of the Winform classes to other platforms (frame buffer or
MacOS X for example), we would gladly integrate them, as long
they are under an open source license.
they are under a Free Software License.
Q: Will Mono run on Windows?
A: We hope so. Currently some parts of Mono only run on Windows
(the C# compiler is a .NET executable) and other parts have only
been compiled on Linux, but work on Windows with Cygwin.
Q: Will Mono depend on GNOME?
A: It will depend only if you are using a particular assembly (for
example, for doing GUI applications). If you are just interested
in Mono for implementing a `Hello World Enterprise P2P Web
Service', you will not need any GNOME components.
<a name="reuse"></a>
** Reusing Existing Code
** Reusing existing Code
Q: What projects will you reuse or build upon?
@@ -551,9 +475,9 @@ A: We want to get Mono in the hands of programmers soon. We are
Q: What about Intel's research JIT framework, ORP?
A: At this time, we are investigating whether we can use elements of
ORP for Mono. ORP is a research JIT engine that has a clearly
ORP for Mono. ORP is a research JIT engine that has a clear
defined API that splits the JIT from the GC system and the actual
byte code implementation.
byte code implementation. It is a research product.
Q: What about using GNU Lightning?
@@ -561,17 +485,154 @@ A: We are also researching <a
href="http://www.gnu.org/software/lightning/lightning.html">GNU
Lightning</a>.
<a name="gcc"></a>
<a name="ximian-and-microsoft">
** Ximian and Microsoft
Q: I read that Microsoft is helping out Ximian, is this true?
A: Initial contact between David Stutz and Miguel de Icaza happened.
It was a friendly conversation. Microsoft is interested in other
implementing .NET and are willing to help make the ECMA spec more
accurate for this purpose.
We were initially contacted by Sam Ruby at the ECMA TG3 committee
to discuss the same issue. And we are glad to have good contacts
to ask questions about the specs.
Q: Is Microsoft paying Ximian to do this?
A: No, we are doing this for purely selfish reasons. We are upgrading
our development platform to build better applications on Unix and other
systems.
Q: Do you fear that Microsoft will change the spec and render Mono useless?
A: No. Microsoft proved with the CLI and the C# language that it was
possible to create a powerful foundation for many languages to
interoperate. We will always have that.
Even if changes happened in the platform which were undocumented
(which is very unlikely), the existing platform has a value on its
own.
Miguel once explained its motivation for working on Mono to Dave
Winer, and his mail got posted <a
href="http://scriptingnews.userland.com/stories/storyReader$1275">here</a>
Q: Didn't Miguel de Icaza say that `Unix Sucks'?
A: Yes, he did, as a catch phrase in his opening remark on the Ottawa
Linux Symposium. His talk focused on various ways to improve Unix.
There is a paper describing some ways to improve Unix at:
<a href="http://primates.ximian.com/~miguel/bongo-bong.html">
http://primates.ximian.com/~miguel/bongo-bong.html</a>
Q: Didn't Ximian's Miguel work for Microsoft?
A: Actually, Nat Friedman (Ximian's co-founder) did work as an
intern for Microsoft for one summer but Miguel did not.
Q: Did Nat and Miguel meet at Microsoft?
A: They met online on the Linux IRC network; They met in person for
the first time in 1997.
** Mono and Microsoft
Q: How can you expect Mono to compete with Microsoft, wont this
require an effort too large?
A: You are right. Mono will never become a reality without the help
of other contributors. Ximian is a small company that can not
finish Mono alone. We will be working with members of the
community to deliver the product.
Q: Is Microsoft and Corel involved in the Mono implementation?
A: No, they are not.
Q: Are you writing Mono from the ECMA specs?
A: yes, we are writing them from the ECMA specs and the published
materials in print about .NET
Q: What happens if Microsoft `Embraces and Extends' the CLI standard
and keeps Mono out of the play?
A: There are various explanations to this question. The first one is
that the benefits that you can get from the CLI are going to be
there with or without `embracing and extending'. We might not be
able to run every .NET Windows application on Mono. But remember:
it was already easy for someone to just use PInvoke to tie their
application to Windows.
The bottom line is that the advantages of having a CLI runtime will
be with us, no matter if Microsoft forks their version to be
incompatible.
Q: What if Microsoft changes the interface, and all of a sudden
applications break?
A: If they change their released API, every application that was
developed against it will break.
That being said, Microsoft have a pretty good record of keeping
backwards binary compatibility.
** Acronyms
Q: What is the difference between CLR (Common Language Runtime) and
CLI (Common Language Infrastructure)?
A: CLI is the specification of an execution system. The Microsoft
implementation of this specification is named CLR.
Unless we come up with our own acronym, we could just call ours
also CLR, just because it would do exactly the same thing the
Microsoft implementation does.
<a name="gcc">
** Mono and GCC
Q: Are you working on a GCC front-end to C#? A GCC backend that will
generate CIL images? What about making a front-end to GCC that
takes CIL images and generates native code?
Q: Should someone work on a GCC front-end to C#?
A: We are currently seeking volunteers for those projects.
Visit the <a href="contributing.html">contributing</a> section if
you are interested.
A: I would love if someone does, and we would love to help anyone that
takes on that task, but we do not have the time or expertise to
build a C# compiler with the GCC engine. I find it a lot more fun
personally to work on C# on a C# compiler, which has an intrinsic
beauty.
We can provide help and assistance to anyone who would like to work
on this task.
Q: Should someone make a GCC backend that will generate CIL images?
A: I would love to see a backend to GCC that generates CIL images. It
would provide a ton of free compilers that would generate CIL
code. This is something that people would want to look into
anyways for Windows interoperation in the future.
Again, we would love to provide help and assistance to anyone
interested in working in such a project.
Q: What about making a front-end to GCC that takes CIL images and
generates native code?
A: I would love to see this, specially since GCC supports this same
feature for Java Byte Codes. You could use the metadata library
from Mono to read the byte codes (ie, this would be your
"front-end") and generate the trees that get passed to the
optimizer.
Ideally our implementation of the CLI will be available as a shared
library that could be linked with your application as its runtime
support.
Again, we would love to provide help and assistance to anyone
interested in working in such a project.
Q: But would this work around the GPL in the GCC compiler and allow
people to work on non-free front-ends?
@@ -580,65 +641,68 @@ A: People can already do this by targeting the JVM byte codes (there
Q: Why are you writing a JIT engine instead of a front-end to GCC?
A: We want the JIT engine and runtime engine because they will be able
to execute CIL executables generated on Windows, and so no recompilation
will be necessary.
A: The JIT engine and runtime engine will be able to execute CIL
executables generated on Windows.
** Mono and Portability
Q: Will Mono work on other variants of Unix?
A: Yes. We do not expect to add any gratuitous incompatibilities.
Q: Will Mono run on Windows?
A: Hopefully yes. Currently some parts of Mono only run on Windows
(the C# compiler is a .NET executable) and other parts have only
been compiled on Linux, but work on Windows with Cygwin.
Q: Will Mono depend on GNOME?
A: It will depend only if you are using a particular assembly (for
example, for doing GUI applications). If you are just interested
in Mono for implementing a `Hello World Enterprise P2P Web
Service', you will not need any GNOME component.
<a name="performance"></a>
** Performance
Q: How fast will Mono be?
Q: How fast will be Mono?
A: We can not predict the future, but a conservative estimate is that
it would be at least `as fast as other JIT engines'.
We would like to ship various JIT engines with Mono, just like
Microsoft has done with their .NET development platform. We could
provide a faster, JIT for fast load times but lower performance,
and an and an optimizing JIT that would be slower at generating
code but produce more optimal output.
Now, wishfully thinking I hope that we will ship various JITs with
Mono just like Microsoft has done. A fast JITer when maximum
performance is not needed, but fast load times are important; And
an optimizing JITer that would be slower at generating code but
produce more optimal output.
The CIL has some advantages over the Java byte code: it is really
an intermediate representation and there are a number of
restrictions on how you can emit CIL code that simplify creating
better JIT engines.
For example, on the CIL, the stack is not really an abstraction
available for the code generator to use at will. Rather, it is a
way of creating a postfix representation of the parsed tree. At
any given call point or return point, the contents of the stack are
For example, on the CIL the stack is not really an abstraction
available for the code generator to use at will: it is just a way
of creating a postfix representation of the parsed tree. At any
given call point or return point, the contents of the stack are
expected to contain the same object types independently of how the
instruction was reached.
instructions was reached.
<a name="licensing"></a>
** Licensing
Q: Will I be able to write proprietary applications that run with
Mono?
** Mono and Portable.NET
A: Yes. The licensing scheme is planned to allow proprietary
developers to write applications with Mono.
Q: What license or licenses are you using for the Mono Project?
Q: What are the differences between Mono and Portable.NET?
A: The C# Compiler is released under the terms of the <a
href="http://www.opensource.org/licenses/gpl-license.html">GNU GPL</a>. The runtime
libraries are under the <a
href="http://www.opensource.org/licenses/lgpl-license.html">GNU
Library GPL</a>. And the class libraries are released
under the terms of the <a
href="http://www.opensource.org/licenses/mit-license.html">MIT X11</a>
license.
A: Most of Mono is being written using C#, the only pieces written in
C are those who have to absolutely be built using C (The JIT
engine, the runtime, the interfaces to the garbage collection
system).
Q: I would like to contribute code to Mono under a particular
license. What licenses will you accept?
The C# compiler and the tools will become reusable C# components.
A: We will have to evaluate the licenses for compatibility first,
but as a general rule, we will accept the code under the same
terms of the "container" module.
Portable.NET is building its components out of C pieces.
<a name="etc"></a>
** Miscellaneous Questions
** Assorted questions
Q: You say that the CLI allows multiple languages to execute on the
same environment. Isn't this the purpose of CORBA?
@@ -656,8 +720,3 @@ A: The key difference between CORBA (and COM) and the CLI is that the
space. This means much less copying and no need for reference
counting.
Q: Will you support COM?
A: The runtime will support XPCOM on Unix systems and COM on Windows.
Most of the code for dynamic trampolines exists already.
+16 -9
View File
@@ -1,14 +1,5 @@
* Ideas
There are many pending classes that need to be implemented. Those have
the highest impact in the Mono project: the sooner they are done,
the sooner we can start using this platform to create new and
exciting applications.
That being said, if you are not very excited about working on class
libraries, here is a list of other related projects or tools that might
be useful to the Mono project.
Here are a few ideas of tools, classes and projects that you
could start. More are forthcoming.
@@ -51,6 +42,22 @@ TODO=multimedia
href="mailto:[email protected]">Vladimir</a> has
researched the problem before
TODO=gtk,Gtk+ wrappers for Mono and .NET
* Wrap the Gtk+ API. This is simple and can be done
on Windows as Gtk+ 2.0 works on Windows.
This work can also be used on Windows and will
enable developers on Windows to use some of Gtk+'s
advanced features.
The idea is to wrap the Gtk+ API and allow us to
build GUI applications using Gtk+ and in the future
other Gtk+-based libraries from Mono (Gal, GtkHTML).
There is extensive knowledge on wrapping Gtk+ in
other languages (this has been done this for Perl,
Python, Java, Scheme, Haskel and other languages in
the past).
</ul>
<a name="projects">
+2 -603
View File
@@ -1,6 +1,3 @@
<table>
<tr>
<td>
<a href="http://www.ximian.com">Ximian</a> announced the
launch of the Mono project, an effort to create an Open Source
implementation of the .NET Development Framework.
@@ -8,7 +5,7 @@
Mono includes: <a href="c-sharp.html">a compiler</a> for the
C# language, a <a href="runtime.html">runtime</a> for the
Common Language Infrastructure and a set of <a
href="class-library.html">class libraries</a>.
href="class-librayr.html">class libraries</a>
You can read our <a href="rationale.html">rationale</a> for
this project. If you have questions about the project, please
@@ -25,634 +22,47 @@
You can contact the team at: <a
href="mailto:[email protected]">[email protected]</a>
</td>
<td>
<table border=1 bgcolor>
<tr>
</tr>
<tr>
<td>
<b><a href="c-sharp.html">Compiler</a></b>
</td>
<td>
Self hosting on .NET.<br>
Classes need work to run on Mono.
</td>
</tr>
<tr>
<td>
<b>JIT</b>
</td>
<td>
Linux/x86 working.
</td>
</tr>
<tr>
<td>
<b>Interpreter</b>
</td>
<td>
Linux/PPC working<br>
In progress: StrongARM, SPARC.
</td>
</tr>
<tr>
<td>
<b><a href="class-status.html">Classes</a></b>
</td>
<td>
Corlib for self hosting mostly done
</td>
</tr>
</table>
</td>
</tr>
</table>
** Feb 11, 2002
We would like to welcome all the new developers that have
joined the project in the last couple of days. The classes
are rapidly moving.
An explanation of the relationship between <a
href="http://mail.gnome.org/archives/gnome-hackers/2002-February/msg00031.html">GNOME
and Mono</a>.
Nick is still leading our test suite platform. I can not
stress how important it is to have a good regression test suite
for our platform, as buggy class libraries are what are
stopping the compiler from running completely on Linux.
We are of course psyched to see Mono run on
non-Linux systems. Work is moving on native code generation
for StrongARM, PowerPC, and SPARC as well as porting Mono to
other systems.
There are a couple of debates on the Mono list on implementing
a set of web server classes for <a
href="http://lists.ximian.com/archives/public/mono-list/2002-February/002911.html">enabling
ASP.NET</a> on Mono.
Paolo also <a
href="http://lists.ximian.com/archives/public/mono-list/2002-February/002944.html">
posted a list of pending tasks</a> to enable the compiler to run on Linux
** Feb 10, 2002
Mike Kestner has posted an <a
href="http://lists.ximian.com/archives/public/gtk-sharp-list/2002-February/000024.html">Update
on his Gtk#</a> activities.
** Feb 4, 2002
Adam has done <a
href="http://lists.ximian.com/archives/public/mono-list/2002-February/002808.html">Qt
bindings</a> for .NET. Adam is cool.
** Jan 29, 2002
Dan Lewis has contributed a major missing set of classes to
Mono: <a
href="http://lists.ximian.com/archives/public/mono-list/2002-January/002745.html">
System.Text.RegularExpressions</a>.
This is a fully .NET compatible implementation of the .NET regular expressions,
fully Unicode aware. This contribution is very appreciated, as implementing this
was not entirely trivial (supporting Unicode, plus a regex engine which is a super
set of the Perl regex engine).
** Jan 28, 2002
The Mono contributors have relicensed the Class Libraries under
the terms of the
<a href="http://www.opensource.org/licenses/mit-license.html">MIT X11</a> license.
This license is an Open Source license, and is used by other projects
(most notably, the XFree86 project).
The runtime (JIT, metadata library, interpreter) remains under
the LGPL and the C# compiler remains under the GPL.
Our <a
href="http://www.ximian.com/about_us/press_center/press_releases/mono_partners.html">Press
Release</a>
Press coverage: <a
href="http://news.com.com/2100-1001-823734.html">CNet</a>, <a
href="http://www.wired.com/news/technology/0,1282,50037-2,00.html">Wired</a>,
<a href="http://www.infoworld.com/articles/hn/xml/02/01/28/020128hnopennet.xml">InfoWorld</a>,
<a href="http://www.newsforge.com/article.pl?sid=02/01/27/2232231">NewsForge</a>.
** Jan 23, 2002
New mailing list: <a href="mailto:[email protected]">[email protected]</a>.
This mailing list will receive automatically the patches that are submitted
to the Mono CVS to any of its modules.
This allows anyone who wants to participate in the peer-review of the
code submitted to CVS to receive patches on e-mail. It should also
expose to everyone the changes that are being done by the team every day.
** Jan 21, 2002
Dick has got a simple web server running with Mono (`MIS: Mono
Internet Server') that is mostly used to test our IO layer, a
<a href="http://primates.ximian.com/~miguel/dick-mis-server.png">screenshot</a>
Paolo and Dietmar are busy making our runtime self sufficient on
non-Windows platforms.
C# compiler front: A lot of focus in the past weeks after
the C# became self hosting has been in making the compiler a useful
tool for development: improve error handling, provide better error
reports, fixing all known bugs, and finally profiling of the compiler
has begun.
** Jan 8, 2002
Our compiler has been self-supporting since January 3rd. In
the meantime, we have been busy working on making it run on
Linux. Today Paolo got more work done on Reflection.Emit and
the compiler compiled `console.cs' (a sample Mono program) on
Linux.
** Jan 4, 2002
Dietmar landed the Unicode support patch. Class libraries and
runtimes are now fully Unicode aware. The details are <a href=
"http://lists.ximian.com/archives/public/mono-list/2002-January/002409.html">
here</a>
Last minute breaking news: Paolo got our compiler in Linux to
compile fib.cs, patches are comming tomorrow once we have
ChangeLog entries.
** Jan 4, 2002
Mike Kestner posted an update on Gtk# <a
href="http://lists.ximian.com/archives/public/gtk-sharp-list/2002-January/000021.html"><i>New
year, new direction</i></a>.
Gtk# will be our foundation on which we will be implementing
System.Windows.Forms.
** Jan 3, 2002
Mono C# compiler becomes self-sufficient. We can now continue
development of the compiler with itself.
Work on the class libraries is still underway for having a full
self hosting system. We hope to achieve our goal of self-hosting
on Linux before the end of the month.
Join the fun by downloading either tonight's <a
href="snapshots">snapshot</a> or getting your sources from our
<a href="anoncvs.html">Anonymous CVS server</a>.
** Dec 28, 2001
After a lot of work, the C# compiler can compile itself.
There are still errors in the generated image, but they are
being fixed quickly.
We will soon have the first non-Microsoft C# implementation!
** Dec 18, 2001
JIT: More work on our IO abstraction layer (Dick).
JIT: exception handling for unmanaged code (Dietmar)
System.Reflection: Support for PropertyInfo and
PropertyBuilder as well as the various queries for MethodBase.
C#: Pre-processor; Rewrite of MemberLookup which fixed many of
the outstanding issues. More bug fixing allows it to compile
more programs.
** Dec 14, 2001
Dietmar has improved the register allocation and now Mono performs
two to three times as fast as it did yesterday. Amazing.
The compiler keeps moving along, explicit interface
implementation is there.
** Dec 11, 2001
The JIT engine can now run all the compiler regression tests as
well as assorted other programs, many more opcodes added
recently. Currently the JIT engine uses a very simplistic register
allocator (just enough to allow us to focus on feature completeness)
and that will be the next major task to improve performance and
reduce spills and reloads.
On the C# compiler front: language features are now pretty
much complete. The big missing tasks are unsafe code support,
visibility, explicit interface implementation plus static flow
analysis. There are many small bugs that need to be addressed.
You can get your copy of the <a href="snapshots">latest Mono</a>
More work is also required on fixing the foundation class
libraries, it is easy to find spots now since Nick got the
`make test' going.
** Dec 1, 2001
AnonCVS access to Mono is here (updated every hour). Thanks
to <a href="http://www.hispalinux.es">HispaLinux</a> and Jesus
Climent for helping to set this up.
** Nov 30, 2001
All tests from the mono runtime work with the JIT engine now
(Dietmar).
Recursive enumeration definition in the C# compiler are
working now (Ravi).
More work on the Web classes (Gaurav).
** Nov 28, 2001
JIT land: Paolo got GDB support into the JIT engine while
Dietmar added exceptions support to it.
The C# compiler supports all array initializations now, and the
switch statement as well as fixing many existing bugs. Many
new more tests.
Nick keeps working on improving our class library test suite.
Dick has almost completed the Mono IO layer.
** Nov 16, 2001
<blockquote>
Mike Kestner has posted an <a
href="http://lists.ximian.com/archives/public/gtk-sharp-list/2001-November/000015.html">update</a>
on Gtk# development.
</blockquote>
** Nov 14, 2001
<blockquote>
Paolo today got the Mono C# compiler running <a
href="http://lists.ximian.com/archives/public/mono-list/2001-November/001941.html">on
Linux</a>. It compiles a sample program and then the sample
program is executed.
Mutator unary operators (++ and --) in the compiler are fully
functional, they used to only work on variables, and now they
are complete.
To sum things up: The Mono C# compiler is written in C# and
uses the .NET classes to get its work done. To make this work
on Linux work has to happen in various fronts:
<ul>
* The C# compiler is being worked on and can compile
many programs now (our test suite at this point is
made up of 40 tests).
* The class libraries need to be mature enough to support
the compiler, particularly System.Reflection.Emit (which is
what Paolo has been working on lately).
The compiler currently requires 103 classes from the
.NET runtime (you can get the list by running: <b>monodis --typeref mcs.exe</b>
* The interpreter should be mature enough to run the actual
compiler byte codes and the corlib bytecodes.
</ul>
At the same time, Dietmar is working on the JIT engine which will
replace our interpreter in production.
</blockquote>
** Nov 12, 2001
<blockquote>
Dietmar got value types working on the JIT engine. Sean has
got assembly loading in the runtime (required for NUnit).
More progress on enumerations and attributes from Ravi.
Nick keeps working on improving our class libraries.
</blockquote>
** Nov 8, 2001
<blockquote>
Enumerations, array access and attributes for the C# compiler are into the CVS now.
Full array support is not complete, but moving along.
</blockquote>
** Nov 5, 2001
<blockquote>
Dietmar's new set of patches to the JIT have 20 out of 33
tests running now.
</blockquote>
** Nov 4, 2001
<blockquote>
Mike Kestner, main Gtk# contributor has posted a very interesting <a
href="http://lists.ximian.com/archives/public/gtk-sharp-list/2001-November/000013.html">
update</a> on his work on Gtk#.
Ravi commited the initial support for Attributes in the
compiler.
Many HTML Controls from Leen checked into CVS.
Paolo checked in his new System.Reflection and
System.Reflection.Emit implementations. He has been working
steadily on this huge task for a few weeks now. This is the
foundation for the Mono C# compiler, and hence a very
important piece of the puzzle.
</blockquote>
** Nov 3, 2001
<blockquote>
Many clean ups have been going into the class library by Nick Drochak.
Mega patch from Dietmar: he commited the flow analysis code
for the JITer.
A lot of work has been going into the WebControls by Gaurav (4
new controls plus improved and bug fixed base classes).
</blockquote>
** Nov 1, 2001
<blockquote>
Ravi commited the caller-side method selection of methods with
variable length arguments. Now he depends on Miguel finishing
the array handling support.
</blockquote>
** Oct 27, 2001
<blockquote>
Lots of classes for System.Web from Gaurav were commited this
morning.
Some large recent developments:
The Decimal implementation from Martin Weindel has been
partially integrated (we need to put the internalcalls in
place now and compile and link the decimal code).
Derek Holden commited recently the IntegerFormatter code into
the CVS, so we got a pretty comprehensive integer formatting
engine that we can finally use all over the place.
Compiler got support for lock as well as assorted bug fixes.
Ravi is still working on array support (and then we can
optimize foreach for the array case).
Dietmar is busy working on flow analysis on the JITer, the
previous mechanism of generating the forest was wrong. Paolo
has been a busy bee reworking the System.Reflection.Emit
support code, and we should have some pretty nice stuff next
week. Dick on the other hand is still working on the
WaitOne/WaitAll emulation code. WaitAll is like select on
steroids: it can wait for different kinds of objects: files,
mutexes, events and a couple of others.
Mike Kestner is busy working on Gtk# which is now using the
.defs files to quickly wrap the API.
</blockquote>
** Oct 18, 2001
<blockquote>
Reworking expressions to support cleanly indexers and
properties. <a href="http://www.nat.org/evolution.php3">11
days</a> until Evolution 1.0 ships.
Ximian users around the world <!--a
href="http://www.bez.it/IMAGES/nora.jpg"-->rejoice<!--/a--> with
recent C# compiler progress.
</blockquote>
** Oct 17, 2001
<blockquote>
Delegate support has been checked into the compiler
(definition and invocation); break/continue implemented.
</blockquote>
** Oct 15, 2001
<blockquote>
JIT engine supports many of the object constructs now (object
creation, vtable setup, interface table setup).
The C# compiler now has almost full property support (only
missing bit are pre-post increment/decrement operations),
delegates are now created (still missing delegate invocation).
try/catch/finally is also supported in the compiler now.
System.Decimal implementation is in, as well as many crypto
classes.
</blockquote>
** Oct 5, 2001
<blockquote>
Sergey has released his first version of the <b>ilasm</b>
assembler written in C#. You can get it from his web page:
<a
href="http://mono.eurosoft.od.ua">http://mono.eurosoft.od.ua</a>.
The plan is to integrate ildasm into the Mono CVS soon. This
component should in theory also be reusable for SharpDevelop
eventually.
</blockquote>
** Oct 4, 2001
<blockquote>
Our System.Reflection.Emit implementation created its first
executable today. This means that a very simple .NET program
that was compiled on Windows was able to generate a .NET program
while running on Linux using the Mono runtime.
The various piece of the puzzle are starting to get together:
the compiler can compile simple programs now and we are
basically focusing on completeness now.
</blockquote>
** Sep 28, 2001
<blockquote>
<a
href="http://www.icsharpcode.net/OpenSource/SD/default.asp">Sharp
Develop 0.80</a> was released today.
</blockquote>
** Sep 26, 2001
<blockquote>
More progress: more opcodes are working (Paolo); The compiler
runs up to a point in Mint (Paolo); operator overloading works
(both unary and binary) all over the place (Miguel); Completed decimal
type conversions (Miguel); New build system in place based on
Ant (Sean and Sergey); Refactored and documented the
internals of the JIT engine (Dietmar); StatementExpressions
handled correctly (Miguel).
</blockquote>
** Sep 21, 2001
<blockquote>
A couple of news-worthy items: Dick got the initial thread
support into mint; Paolo implemented many new opcodes; Dietmar
got long operations and mul/div working on the JITer; Ravi rewrote
the Method selector for expressions to be conformant; Miguel
got i++ working. All in tonight's snapshot
</blockquote>
** Sep 19, 2001
<blockquote>
Paolo has written a section on <a href="porting.html">Porting
Mono</a> to othre architectures.
</blockquote>
** Sep 18, 2001
<blockquote>
<A a href="download.html#sep-18">Mono 0.7</a> has been
released (runtime engine, class libraries
and C# compiler). Check the <a href="archive/mono-0.7">Mono
0.7 announcement</a> for details
</blockquote>
** Sep 17, 2001
<blockquote>
Mike Kestner's Gtk# (Gtk-sharp) was checked into the CVS
repository. Gtk# can run a simple hello world application.
The binding is nice, as it maps Gtk+ signals to delegates in
C#. You can see the Gtk# Hello World program <a href="src/HelloWorld.cs">here</a>
Gtk-sharp should be available on the next snapshot set.
</blockquote>
** Sep 10, 2001
<blockquote>
Dietmar checked in his CIL tree/forest regeneration and most
importantly, the x86 instruction selector burg grammar.
</blockquote>
** Sep 5, 2001
<blockquote>
The MCS compiler <b>can compile the sample Hello World</b>
application and generate a Windows/CIL executable that runs!
This executable runs with the Mono Interpreter of course (see
August 28)
</blockquote>
** Sep 4, 2001
<blockquote>
Dietmar checked into CVS the `monoburg' architecture
independent instruction selector for the JIT engine.
</blockquote>
** Aug 28, 2001
<blockquote>
<b>.NET Hello World is working under Mono!</b> The latest snapshots
will let you run it.
Hello World consits of 1821 CIL instructions,
performs 66 subroutine calls and loads 12 classes from the corlib.dll
Good work Mono team!
</blockquote>
** Aug 23, 2001
<blockquote>
Lloyd Dupont has announced his OpenGL bindings for C#, they
are available here: <a
href="http://csgl.sourceforge.net">http://csgl.sourceforge.net</a>
</blockquote>
** Aug 22, 2001
<blockquote>
New version of the Mono Runtime, Compiler and Classes has been
<a
href="download.html#august-22">released.</a> Check the <a
href="archive/mono-0.6">0.6 announcement</a>.
</blockquote>
** Aug 20, 2001
<blockquote>
A new <a href="contributing.html#compile-service">Compilation
service</a> has been made available by Derek to allow people
without access to the <a
href="http://msdn.microsoft.com/downloads/default.asp?url=/downloads/sample.asp?url=/msdn-files/027/000/976/msdncompositedoc.xml&frame=true">.NET SDK</a>
</blockquote>
** Aug 3, 2001
<blockquote>
Daily snapshots of mcs and mono are now available, they will
run every night at 10pm Boston time.
</blockquote>
** Jul 29, 2001
<blockquote>
Mono Runtime 0.5 has been <a
href="download.html#july-29">released.</a> Check the <a
href="archive/mono-0.5">release notes</a>
</blockquote>
** Jul 25, 2001
<blockquote>
The slides for <A href="Presentations/O-Reilly">my
presentation</a> at <a href="http://www.oreilly.com">O'Reilly
Open Source Software Convention</a>
</blockquote>
** Jul 22, 2001
<blockquote>
Another release of the class libraries is out, check the <a
href="archive/mcs-22">MCS 22-July Release Notes</a>. You can
get the new class libraries from <a
href="download.html#july-22">here</a>
</blockquote>
** Jul 19, 2001
<blockquote>
Another release of the class libraries is out, check the <a
href="archive/mcs-19">MCS 19-July Release Notes</a>. You can
get the new class libraries from <a
href="download.html#july-19">here</a>
</blockquote>
** Jul 17, 2001
<blockquote>
Another release of the class libraries is out, check the <a
href="archive/mcs-17">MCS 17-July Release Notes</a>. You can
get the new class libraries from <a
@@ -663,42 +73,31 @@
Got Sean's new <a href="class-status/index.html">Class
Status</a> web pages up. These are a lot better than mine, and
we are now keeping better track of contributors.
</blockquote>
** Jul 15, 2001
<blockquote>
Another release of Mono is out, check the <a
href="archive/mono-0.4">Mono 0.4 Release Notes</a>. Get it <a
href="download.html#july-15">here</a>.
</blockquote>
** Jul 14, 2001
<blockquote>
A <a
href="http://lists.ximian.com/archives/public/mono-list/2001-July/000399.html">new
href="http://mail.ximian.com/archives/public/mono-list/2001-July/000399.html">new
release</a> of the
runtime, compiler and classes has been made. Get it <a href="download.html#july-14">here</a>
</blockquote>
** Jul 12, 2001
<blockquote>
I keep getting questions about my opinion on Passport, even when
Mono has <b>nothing</b> to do with it. I finally <a
href="passport.html">wrote something.</a>
</blockquote>
** Jul 9, 2001
<blockquote>
Project launched.
</blockquote>
** O'Reilly
<blockquote>
Brian posted a story on <a
href="http://www.oreillynet.com/dotnet">O'Reilly Network .NET</a>
</blockquote>
+1 -18
View File
@@ -1,17 +1,5 @@
* Updates
<b>Sep 20, 2001:</b> Microsoft has just announced some changes
to passport that are rather interesting. This document
reflects the Passport system without taking into account the
new changes.
Read about it <a href="http://www.microsoft.com/presspass/features/2001/sep01/09-20passport.asp">here</a>.
For an analysis of security problems with passport, check <a
href="http://avirubin.com/passport.html">http://avirubin.com/passport.html</a>.
The bottom line is that you should not put any sensitive
information on passport.
I have received many comments from people, and I have updated
the page accordingly. From removing incorrect statements, to
fixing typos, to include mentions to other software pieces.
@@ -256,7 +244,7 @@
A few people have said: `Mono will allow Passport to be
available for Linux and that is bad'. This is plain
misinformation.
missinformation.
Currently, you can obtain Passport for Linux from Microsoft
itself and deploy it today on your Web server. Mono does not
@@ -278,11 +266,6 @@
competing system to Passport could be designed, but I will let
<a href="http://www.nat.org/">him</a> post his own story.
** Other Passport Comments
An interesting study on the security of passport is available at: <a
href="http://avirubin.com/passport.html">http://avirubin.com/passport.html</a>
** Other Alternatives
Some people have pointed out <a
-384
View File
@@ -1,384 +0,0 @@
Microsoft.Win32.Registry
Microsoft.Win32.RegistryHive
Microsoft.Win32.RegistryKey
System.__ComObject
System.Activator
System.ArgIterator
System.AssemblyLoadEventHandler
System.AsyncCallback
System.CrossAppDomainDelegate
System.DBNull
System.Diagnostics.SymbolStore.SymDocumentType
System.Diagnostics.SymbolStore.SymLanguageType
System.Diagnostics.SymbolStore.SymLanguageVendor
System.EventHandler
System.Globalization.CompareInfo
System.Globalization.HebrewCalendar
System.Globalization.HijriCalendar
System.Globalization.JapaneseCalendar
System.Globalization.KoreanCalendar
System.Globalization.RegionInfo
System.Globalization.SortKey
System.Globalization.StringInfo
System.Globalization.TaiwanCalendar
System.Globalization.TextElementEnumerator
System.Globalization.TextInfo
System.Globalization.ThaiBuddhistCalendar
System.IO.EndOfStreamException
System.IO.IsolatedStorage.IsolatedStorage
System.IO.IsolatedStorage.IsolatedStorageException
System.IO.IsolatedStorage.IsolatedStorageFile
System.IO.PathTooLongException
System.Reflection.Emit.MethodRental
System.Reflection.MemberFilter
System.Reflection.Missing
System.Reflection.ModuleResolveEventHandler
System.Reflection.Pointer
System.Reflection.StrongNameKeyPair
System.Reflection.TargetException
System.Reflection.TypeFilter
System.ResolveEventHandler
System.Runtime.CompilerServices.AccessedThroughPropertyAttribute
System.Runtime.CompilerServices.CallConvCdecl
System.Runtime.CompilerServices.CallConvFastcall
System.Runtime.CompilerServices.CallConvStdcall
System.Runtime.CompilerServices.CallConvThiscall
System.Runtime.CompilerServices.CompilationRelaxationsAttribute
System.Runtime.CompilerServices.CompilerGlobalScopeAttribute
System.Runtime.CompilerServices.CustomConstantAttribute
System.Runtime.CompilerServices.DateTimeConstantAttribute
System.Runtime.CompilerServices.DecimalConstantAttribute
System.Runtime.CompilerServices.DiscardableAttribute
System.Runtime.CompilerServices.IDispatchConstantAttribute
System.Runtime.CompilerServices.IsVolatile
System.Runtime.CompilerServices.IUnknownConstantAttribute
System.Runtime.CompilerServices.RequiredAttributeAttribute
System.Runtime.InteropServices.ArrayWithOffset
System.Runtime.InteropServices.AssemblyRegistrationFlags
System.Runtime.InteropServices.AutomationProxyAttribute
System.Runtime.InteropServices.BIND_OPTS
System.Runtime.InteropServices.BINDPTR
System.Runtime.InteropServices.CALLCONV
System.Runtime.InteropServices.ClassInterfaceAttribute
System.Runtime.InteropServices.ClassInterfaceType
System.Runtime.InteropServices.CoClassAttribute
System.Runtime.InteropServices.ComAliasNameAttribute
System.Runtime.InteropServices.ComConversionLossAttribute
System.Runtime.InteropServices.ComEventInterfaceAttribute
System.Runtime.InteropServices.COMException
System.Runtime.InteropServices.ComImportAttribute
System.Runtime.InteropServices.ComInterfaceType
System.Runtime.InteropServices.ComMemberType
System.Runtime.InteropServices.ComRegisterFunctionAttribute
System.Runtime.InteropServices.ComSourceInterfacesAttribute
System.Runtime.InteropServices.ComUnregisterFunctionAttribute
System.Runtime.InteropServices.ComVisibleAttribute
System.Runtime.InteropServices.CONNECTDATA
System.Runtime.InteropServices.CurrencyWrapper
System.Runtime.InteropServices.DESCKIND
System.Runtime.InteropServices.DispatchWrapper
System.Runtime.InteropServices.DispIdAttribute
System.Runtime.InteropServices.DISPPARAMS
System.Runtime.InteropServices.ELEMDESC
System.Runtime.InteropServices.ErrorWrapper
System.Runtime.InteropServices.EXCEPINFO
System.Runtime.InteropServices.Expando.IExpando
System.Runtime.InteropServices.ExporterEventKind
System.Runtime.InteropServices.ExtensibleClassFactory
System.Runtime.InteropServices.FILETIME
System.Runtime.InteropServices.FUNCDESC
System.Runtime.InteropServices.FUNCFLAGS
System.Runtime.InteropServices.FUNCKIND
System.Runtime.InteropServices.GCHandle
System.Runtime.InteropServices.GuidAttribute
System.Runtime.InteropServices.HandleRef
System.Runtime.InteropServices.ICustomAdapter
System.Runtime.InteropServices.ICustomFactory
System.Runtime.InteropServices.ICustomMarshaler
System.Runtime.InteropServices.IDispatchImplAttribute
System.Runtime.InteropServices.IDispatchImplType
System.Runtime.InteropServices.IDLDESC
System.Runtime.InteropServices.IDLFLAG
System.Runtime.InteropServices.IMPLTYPEFLAGS
System.Runtime.InteropServices.ImportedFromTypeLibAttribute
System.Runtime.InteropServices.ImporterEventKind
System.Runtime.InteropServices.InterfaceTypeAttribute
System.Runtime.InteropServices.InvalidComObjectException
System.Runtime.InteropServices.InvalidOleVariantTypeException
System.Runtime.InteropServices.INVOKEKIND
System.Runtime.InteropServices.IRegistrationServices
System.Runtime.InteropServices.ITypeLibConverter
System.Runtime.InteropServices.ITypeLibExporterNameProvider
System.Runtime.InteropServices.ITypeLibExporterNotifySink
System.Runtime.InteropServices.ITypeLibImporterNotifySink
System.Runtime.InteropServices.LCIDConversionAttribute
System.Runtime.InteropServices.LIBFLAGS
System.Runtime.InteropServices.MarshalDirectiveException
System.Runtime.InteropServices.ObjectCreationDelegate
System.Runtime.InteropServices.PARAMDESC
System.Runtime.InteropServices.PARAMFLAG
System.Runtime.InteropServices.PreserveSigAttribute
System.Runtime.InteropServices.PrimaryInteropAssemblyAttribute
System.Runtime.InteropServices.ProgIdAttribute
System.Runtime.InteropServices.RegistrationServices
System.Runtime.InteropServices.RuntimeEnvironment
System.Runtime.InteropServices.SafeArrayRankMismatchException
System.Runtime.InteropServices.SafeArrayTypeMismatchException
System.Runtime.InteropServices.SEHException
System.Runtime.InteropServices.STATSTG
System.Runtime.InteropServices.SYSKIND
System.Runtime.InteropServices.TYPEATTR
System.Runtime.InteropServices.TYPEDESC
System.Runtime.InteropServices.TYPEFLAGS
System.Runtime.InteropServices.TYPEKIND
System.Runtime.InteropServices.TYPELIBATTR
System.Runtime.InteropServices.TypeLibConverter
System.Runtime.InteropServices.TypeLibExporterFlags
System.Runtime.InteropServices.TypeLibFuncAttribute
System.Runtime.InteropServices.TypeLibFuncFlags
System.Runtime.InteropServices.TypeLibImporterFlags
System.Runtime.InteropServices.TypeLibTypeAttribute
System.Runtime.InteropServices.TypeLibTypeFlags
System.Runtime.InteropServices.TypeLibVarAttribute
System.Runtime.InteropServices.TypeLibVarFlags
System.Runtime.InteropServices.UCOMIBindCtx
System.Runtime.InteropServices.UCOMIConnectionPoint
System.Runtime.InteropServices.UCOMIConnectionPointContainer
System.Runtime.InteropServices.UCOMIEnumConnectionPoints
System.Runtime.InteropServices.UCOMIEnumConnections
System.Runtime.InteropServices.UCOMIEnumMoniker
System.Runtime.InteropServices.UCOMIEnumString
System.Runtime.InteropServices.UCOMIEnumVARIANT
System.Runtime.InteropServices.UCOMIMoniker
System.Runtime.InteropServices.UCOMIPersistFile
System.Runtime.InteropServices.UCOMIRunningObjectTable
System.Runtime.InteropServices.UCOMIStream
System.Runtime.InteropServices.UCOMITypeComp
System.Runtime.InteropServices.UCOMITypeInfo
System.Runtime.InteropServices.UCOMITypeLib
System.Runtime.InteropServices.UnknownWrapper
System.Runtime.InteropServices.VARDESC
System.Runtime.InteropServices.VARFLAGS
System.Runtime.Remoting.ActivatedClientTypeEntry
System.Runtime.Remoting.ActivatedServiceTypeEntry
System.Runtime.Remoting.Activation.UrlAttribute
System.Runtime.Remoting.Channels.BaseChannelObjectWithProperties
System.Runtime.Remoting.Channels.BaseChannelSinkWithProperties
System.Runtime.Remoting.Channels.BaseChannelWithProperties
System.Runtime.Remoting.Channels.ChannelDataStore
System.Runtime.Remoting.Channels.ChannelServices
System.Runtime.Remoting.Channels.ClientChannelSinkStack
System.Runtime.Remoting.Channels.IChannel
System.Runtime.Remoting.Channels.IChannelDataStore
System.Runtime.Remoting.Channels.IChannelReceiver
System.Runtime.Remoting.Channels.IChannelReceiverHook
System.Runtime.Remoting.Channels.IChannelSender
System.Runtime.Remoting.Channels.IChannelSinkBase
System.Runtime.Remoting.Channels.IClientChannelSink
System.Runtime.Remoting.Channels.IClientChannelSinkProvider
System.Runtime.Remoting.Channels.IClientChannelSinkStack
System.Runtime.Remoting.Channels.IClientFormatterSink
System.Runtime.Remoting.Channels.IClientFormatterSinkProvider
System.Runtime.Remoting.Channels.IClientResponseChannelSinkStack
System.Runtime.Remoting.Channels.IServerChannelSink
System.Runtime.Remoting.Channels.IServerChannelSinkProvider
System.Runtime.Remoting.Channels.IServerChannelSinkStack
System.Runtime.Remoting.Channels.IServerFormatterSinkProvider
System.Runtime.Remoting.Channels.IServerResponseChannelSinkStack
System.Runtime.Remoting.Channels.ITransportHeaders
System.Runtime.Remoting.Channels.ServerChannelSinkStack
System.Runtime.Remoting.Channels.ServerProcessing
System.Runtime.Remoting.Channels.SinkProviderData
System.Runtime.Remoting.Channels.TransportHeaders
System.Runtime.Remoting.Contexts.ContextProperty
System.Runtime.Remoting.Contexts.CrossContextDelegate
System.Runtime.Remoting.Contexts.IContextPropertyActivator
System.Runtime.Remoting.Contexts.IContributeClientContextSink
System.Runtime.Remoting.Contexts.IContributeDynamicSink
System.Runtime.Remoting.Contexts.IContributeEnvoySink
System.Runtime.Remoting.Contexts.IContributeObjectSink
System.Runtime.Remoting.Contexts.IContributeServerContextSink
System.Runtime.Remoting.IChannelInfo
System.Runtime.Remoting.IEnvoyInfo
System.Runtime.Remoting.InternalRemotingServices
System.Runtime.Remoting.IObjectHandle
System.Runtime.Remoting.IRemotingTypeInfo
System.Runtime.Remoting.Lifetime.ClientSponsor
System.Runtime.Remoting.Lifetime.ILease
System.Runtime.Remoting.Lifetime.ISponsor
System.Runtime.Remoting.Lifetime.LeaseState
System.Runtime.Remoting.Lifetime.LifetimeServices
System.Runtime.Remoting.Messaging.AsyncResult
System.Runtime.Remoting.Messaging.CallContext
System.Runtime.Remoting.Messaging.ConstructionCall
System.Runtime.Remoting.Messaging.ConstructionResponse
System.Runtime.Remoting.Messaging.Header
System.Runtime.Remoting.Messaging.HeaderHandler
System.Runtime.Remoting.Messaging.ILogicalThreadAffinative
System.Runtime.Remoting.Messaging.IMessageCtrl
System.Runtime.Remoting.Messaging.IMessageSink
System.Runtime.Remoting.Messaging.IMethodCallMessage
System.Runtime.Remoting.Messaging.InternalMessageWrapper
System.Runtime.Remoting.Messaging.IRemotingFormatter
System.Runtime.Remoting.Messaging.LogicalCallContext
System.Runtime.Remoting.Messaging.MessageSurrogateFilter
System.Runtime.Remoting.Messaging.MethodCall
System.Runtime.Remoting.Messaging.MethodCallMessageWrapper
System.Runtime.Remoting.Messaging.MethodResponse
System.Runtime.Remoting.Messaging.MethodReturnMessageWrapper
System.Runtime.Remoting.Messaging.OneWayAttribute
System.Runtime.Remoting.Messaging.RemotingSurrogateSelector
System.Runtime.Remoting.Messaging.ReturnMessage
System.Runtime.Remoting.Metadata.SoapAttribute
System.Runtime.Remoting.Metadata.SoapFieldAttribute
System.Runtime.Remoting.Metadata.SoapMethodAttribute
System.Runtime.Remoting.Metadata.SoapOption
System.Runtime.Remoting.Metadata.SoapParameterAttribute
System.Runtime.Remoting.Metadata.SoapTypeAttribute
System.Runtime.Remoting.Metadata.W3cXsd2001.ISoapXsd
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapAnyUri
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapBase64Binary
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDate
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDateTime
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDay
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapDuration
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapEntities
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapEntity
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapHexBinary
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapId
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapIdref
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapIdrefs
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapInteger
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapLanguage
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapMonth
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapMonthDay
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapName
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNcName
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNegativeInteger
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNmtoken
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNmtokens
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNonNegativeInteger
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNonPositiveInteger
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNormalizedString
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapNotation
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapPositiveInteger
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapQName
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapTime
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapToken
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapYear
System.Runtime.Remoting.Metadata.W3cXsd2001.SoapYearMonth
System.Runtime.Remoting.Metadata.XmlFieldOrderOption
System.Runtime.Remoting.Proxies.ProxyAttribute
System.Runtime.Remoting.Proxies.RealProxy
System.Runtime.Remoting.RemotingConfiguration
System.Runtime.Remoting.RemotingException
System.Runtime.Remoting.RemotingServices
System.Runtime.Remoting.RemotingTimeoutException
System.Runtime.Remoting.ServerException
System.Runtime.Remoting.Services.EnterpriseServicesHelper
System.Runtime.Remoting.Services.ITrackingHandler
System.Runtime.Remoting.Services.TrackingServices
System.Runtime.Remoting.SoapServices
System.Runtime.Remoting.TypeEntry
System.Runtime.Remoting.WellKnownClientTypeEntry
System.Runtime.Remoting.WellKnownServiceTypeEntry
System.Runtime.Serialization.Formatter
System.Runtime.Serialization.FormatterConverter
System.Runtime.Serialization.Formatters.Binary.BinaryFormatter
System.Runtime.Serialization.Formatters.InternalRM
System.Runtime.Serialization.Formatters.InternalST
System.Runtime.Serialization.Formatters.ServerFault
System.Runtime.Serialization.Formatters.SoapFault
System.Runtime.Serialization.Formatters.SoapMessage
System.Runtime.Serialization.FormatterServices
System.Runtime.Serialization.ObjectIDGenerator
System.Runtime.Serialization.ObjectManager
System.Runtime.Serialization.SerializationBinder
System.Runtime.Serialization.SerializationEntry
System.Runtime.Serialization.SerializationInfoEnumerator
System.Runtime.Serialization.SurrogateSelector
System.RuntimeArgumentHandle
System.Security.AllowPartiallyTrustedCallersAttribute
System.Security.Cryptography.CryptoConfig
System.Security.Cryptography.DSACryptoServiceProvider
System.Security.Cryptography.HMACSHA1
System.Security.Cryptography.KeyedHashAlgorithm
System.Security.Cryptography.MACTripleDES
System.Security.Cryptography.MaskGenerationMethod
System.Security.Cryptography.PasswordDeriveBytes
System.Security.Cryptography.PKCS1MaskGenerationMethod
System.Security.Cryptography.RandomNumberGenerator
System.Security.Cryptography.RC2
System.Security.Cryptography.RC2CryptoServiceProvider
System.Security.Cryptography.Rijndael
System.Security.Cryptography.RijndaelManaged
System.Security.Cryptography.RNGCryptoServiceProvider
System.Security.Cryptography.RSA
System.Security.Cryptography.RSACryptoServiceProvider
System.Security.Cryptography.RSAOAEPKeyExchangeDeformatter
System.Security.Cryptography.RSAOAEPKeyExchangeFormatter
System.Security.Cryptography.RSAParameters
System.Security.Cryptography.RSAPKCS1KeyExchangeDeformatter
System.Security.Cryptography.RSAPKCS1KeyExchangeFormatter
System.Security.Cryptography.RSAPKCS1SignatureDeformatter
System.Security.Cryptography.RSAPKCS1SignatureFormatter
System.Security.Cryptography.SHA1Managed
System.Security.Cryptography.SHA384Managed
System.Security.Cryptography.SHA512Managed
System.Security.Cryptography.TripleDES
System.Security.Cryptography.TripleDESCryptoServiceProvider
System.Security.Cryptography.X509Certificates.X509Certificate
System.Security.NamedPermissionSet
System.Security.Permissions.EnvironmentPermission
System.Security.Permissions.FileDialogPermission
System.Security.Permissions.FileDialogPermissionAccess
System.Security.Permissions.IsolatedStorageFilePermission
System.Security.Permissions.IsolatedStoragePermissionAttribute
System.Security.Permissions.PrincipalPermission
System.Security.Permissions.PrincipalPermissionAttribute
System.Security.Permissions.PublisherIdentityPermission
System.Security.Permissions.PublisherIdentityPermissionAttribute
System.Security.Permissions.ReflectionPermission
System.Security.Permissions.RegistryPermission
System.Security.Permissions.SecurityPermission
System.Security.Permissions.SecurityPermissionAttribute
System.Security.Permissions.SiteIdentityPermission
System.Security.Permissions.StrongNameIdentityPermission
System.Security.Permissions.StrongNamePublicKeyBlob
System.Security.Permissions.UIPermission
System.Security.Permissions.UrlIdentityPermission
System.Security.Permissions.ZoneIdentityPermission
System.Security.Policy.AllMembershipCondition
System.Security.Policy.ApplicationDirectory
System.Security.Policy.ApplicationDirectoryMembershipCondition
System.Security.Policy.FileCodeGroup
System.Security.Policy.FirstMatchCodeGroup
System.Security.Policy.Hash
System.Security.Policy.HashMembershipCondition
System.Security.Policy.NetCodeGroup
System.Security.Policy.PermissionRequestEvidence
System.Security.Policy.PolicyStatement
System.Security.Policy.Publisher
System.Security.Policy.PublisherMembershipCondition
System.Security.Policy.Site
System.Security.Policy.SiteMembershipCondition
System.Security.Policy.StrongName
System.Security.Policy.StrongNameMembershipCondition
System.Security.Policy.UnionCodeGroup
System.Security.Policy.Url
System.Security.Policy.UrlMembershipCondition
System.Security.Policy.Zone
System.Security.Policy.ZoneMembershipCondition
System.Security.Principal.WindowsIdentity
System.Security.Principal.WindowsImpersonationContext
System.Security.Principal.WindowsPrincipal
System.Threading.CompressedStack
System.Threading.IOCompletionCallback
System.Threading.ThreadStart
System.Threading.TimerCallback
System.Threading.WaitCallback
System.Threading.WaitOrTimerCallback
System.TimeZone
System.TypedReference
System.UnhandledExceptionEventHandler
System.WeakReference
-63
View File
@@ -1,63 +0,0 @@
* How to port Mono to your preferred architecture
** Endian, 64 bits and unaligned access issues
The first thing to do is to check that the metadata handling
library works on your target processor. You may use the disassembler
on simple programs and check that you get sensible results
(assuming it compiles at all on your system:-).
The main issue is to write macros that read unaligned
little endian shorts/ints/longs/float/doubles: look into
mono/metadata/endian.h. There may be other spots in the code that are
unsafe at reading/writing to some datatypes that require special
alignment, but there should be few such issues and they need to be fixed.
Once this stuff is sorted out, you should be able to run the interpreter
on simple programs that don't require delegates, P/Invoke functions etc..
** Generating assembly bytecodes for the target processor
Next, you need to provide the support code for generating assembly bytecode
for your target platform (in mono/arch/{ppc,sparc,alpha,*}).
The code should be more or less like the code in x86-codegen.h:
macros that produce fast in-line code. You don't need to provide
code to create every possible code, at first, just the code to
create trampolines and execute them is fine (you'll need to research
how the call convention works on your platform): that would be, for
example, the prolog and epilog code in a function, code to pass function
parameters and deal with the return value and so on.
libffi in gcc or the xptcall sources in mozilla may be helpful to
understand how the calling convention works, if you can't find a specification.
You'd need a processor manual to know how to create the assembly binary data.
This requires a lot of reading if you're not familiar with the assembly for your
target platform. Manuals for many processors are available as PDF files on the
web site of the respective vendors. Note that some processors require you to
flush the I-cache before executing the code: have a look at how the same thing is
done in GNU lightning.
** Getting the interpreter to work
Once you can generate binary code, you can start working on a
mono_create_trampoline() function for your platform: this function will receive
a MonoMethod that describes the arguments and the return type of a C function
and will create the code to call such function. When this function is complete
you'll be able to run more sample programs, that use System.IO, P/Invoke
functions etc.
To support delegates you'll need to write a mono_create_method_pointer()
function that creates a native function: this can be used to call the
method using the runtime's calling convention (it's basically the reverse
of mono_create_trampoline()).
** The final step: porting the JIT
At this point you'd need to have a more complete code generation header file
and you can start writing the machine description file for the monoburg
system. This code (jit/tesjit.c) will require some machine specific tweaks,
but hopefully all you have to do is create the grammar that emit assembly
code from the IR tree. Work is at the early stages also for x86 on this stuff
as we are still testing various solutions: you'd want to read about burg-like
code-generator generators (the LCC book is a good starting point).
+1 -1
View File
@@ -18,7 +18,7 @@
standard GNOME interfaces for easily creating reusable,
language independent components, controls and compound
documents. This technology is known as <a
href="http://developer.ximian.com/tech/bonobo.html">Bonobo</a>.
href="http://www.ximian.com/tech/bonobo.php3">Bonobo</a>.
Interfaces to Bonobo exist for C, Perl, Python, and
Java.
+8 -75
View File
@@ -7,64 +7,9 @@
If you want to send suggestions for links, address them to <a
mailto="[email protected]">[email protected]</a>.
** Microsoft .NET
The <a href="http://msdn.microsoft.com/net">Microsoft.NET site.
The Microsoft .NET Framework 1.0 can be downloaded <a href="http://msdn.microsoft.com/downloads/default.asp?url=/downloads/sample.asp?url=/msdn-files/027/000/976/msdncompositedoc.xml&frame=true">here</a>
New methods that are not documented in 1.0, are documented <a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dndotnet/html/framewrkaddend.asp">here</a>
** IRC
Various Mono contributors get together on channel #mono on
irc.gnome.org
** Mono related sites.
Sergey's web page on Mono resources:
<a href="http://mono.eurosoft.od.ua">http://mono.eurosoft.od.ua</a>
Got Dot Net:
<a href="http://www.gotdotnet.com">Got Dot Net</a>
MonoMail:
<a href="http://sourceforge.net/projects/monomail/">http://sourceforge.net/projects/monomail/</a>
Zip classes:
Mike's port to .NET: <a
href="http://www.icsharpcode.net/OpenSource/NZipLib/default.asp">NZipLib</a>
.NET Security:
A nice overview paper on the <a
href="http://www.foundstone.com/pdf/dotnet-security-framework.pdf">.NET
security architecture</a>.
** Presentations
Currently the <a href="Presentations/O-Reilly">O'Reilly</a>
presentation on Mono is available. Arturo Espinosa has given
a talk on Mono in Mexico and made <a
href="http://construct.ximian.com/~arturo/Presentations/Mono.OLS/html">spanish
slides</a>
** ECMA Documentation.
You can get the documentation for the ECMA specs from:
<ul>
* <a href="http://www.ecma.ch/ecma1/STAND/ecma-334.htm">C# Language Specification</a>
* <a href="http://www.ecma.ch/ecma1/STAND/ecma-335.htm">Common Language Infrastructure</a>
</ul>
Older copies of the standard are available here:
You can get the documentation for the ECMA specs from a number of sites:
<ul>
* <a href="http://msdn.microsoft.com/net/ecma">At MSDN</a>
* <a href="http://www.dotnetexperts.com">Dot Net Experts</a>
@@ -94,25 +39,21 @@
<ul>
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-list">mono-list:</a></b>
href="http://mail.ximian.com/mailman/listinfo/mono-list">mono-list:</a></b>
The general Mono discussion list.
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-announce-list">mono-announce-list:</a></b>
href="http://mail.ximian.com/mailman/listinfo/mono-announce-list">mono-announce-list:</a></b>
Announcements of Mono developments.
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-gc-list">mono-gc-list:</a></b>
href="http://mail.ximian.com/mailman/listinfo/mono-gc-list">mono-gc-list:</a></b>
Discussion on the GC system of Mono.
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list:</a></b>
href="http://mail.ximian.com/mailman/listinfo/mono-cvs-list">mono-cvs-list:</a></b>
Track the CVS activity of Mono on this mailing list. You can get <a href="http://www.go-mono.com/snapshots">daily snapshots</a> as well.
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-docs-list">mono-docs-list:</a></b>
href="http://mail.ximian.com/mailman/listinfo/mono-docs-list">mono-docs-list:</a></b>
Discussion on the documentation of Mono.
* <b><a
href="http://lists.ximian.com/mailman/listinfo/mono-patches">mono-patches:</a></b>
Track the development of Mono in realtime: receive patches to the Mono repository
by email.
* <b><a
href="http://discuss.develop.com/dotnet.html">Dotnet mailing
list at Develop.com:</a></b> The guys at Develop Mentor run
this general purpose mailing list.
@@ -146,8 +87,8 @@
version of JavaScript) in C#
* <a
href="http://xmarks.sourceforge.net/doc.html">XMarks DOC.NET:</a> For
creating online documentation you can browse for your own classes.
href="http://www.southern-storm.com.au/portable_net.html">Portable.NET:</a>
Another implementation of the CLI and C# compiler.
* <a href="http://nunit.sourceforge.net">NUnit:</a> A
testing framework for .NET classes.
@@ -157,14 +98,6 @@
* <a href="http://www.intel.com/research/mrl/orp">ORP:</a> A research
JIT/VM/GC system from Intel.
* C# bindings for OpenGL and SDL are available here: <a
href="http://csgl.sourceforge.net">http://csgl.sourceforge.net</a>
* <a
href="http://www.southern-storm.com.au/portable_net.html">Portable.NET:</a>
Another implementation of the CLI and C# compiler.
</ul>
** GNOME Documentation
+51 -60
View File
@@ -1,98 +1,96 @@
* The Mono runtime
The Mono runtime implements a JIT engine for the CIL virtual
machine (as well as a byte code interpreter, this is to
quickly port it to new systems), the class loader, the garbage
collector, threading system and metadata access libraries.
The Mono runtime will implement the JIT engine (and a byte
code interpreter for quickly porting to new systems), the
class loader, the garbage collector, threading system and
metadata access libraries.
We currently have two runtimes:
<ul>
* <b>mono:</b> The Just In Time compiler implemented
using a BURS instruction selector. We only support
x86 machines in the JIT engine at this point.
* <b>mint:</b> The Mono interpreter. This is an
easy-to-port runtime engine.
</ul>
Currently both runtimes are missing garbage collection. We
are planning on using the ORP GC engine and deploy it by
middle January.
* COM and XPCOM
We plan on adding support for XPCOM on Unix and COM on Microsoft
Windows later in our development process.
Currently the runtime has an image loader and metadata access
entry points. The runtime comes with a simple interpreter
that can execute very simple programs.
** Executing MSIL/CIL images
The code will load an executable and map the references to
external assemblies to our own version of the assemblies on
Linux.
GNU/Linux.
Our roadmap looks like this, this has been updated as of
<b>Dec 18, 2001</b>:
<b>Jul 15, 2001</b>:
<ul>
* Milestone 1: <b>Done</b> Fully read and parse all CIL byte-codes
* Milestone 1: <b> Done</b> Fully read and parse all CIL byte-codes
and metadata tokens (ie, a disassembler).
* Milestone 2: <b>Done</b> Complete an interpreter for CIL byte
* Milestone 2: Complete an interpreter for CIL byte
codes. This interpreter can be used temporarly to
run CIL byte code on a system where no JIT is
available.
* Milestone 3: <b>Done</b>Define an <i>lburg</i>-like
instruction selector for the JITer for Intel.
* Milestone 3: Define an <i>lburg</i>-like instruction
selector for the JITer for Intel. Although slower
at JITing than a streaming JITer, it generates
better code. The same grammar can later be used for
the stream jitter.
* Milestone 4: <b>Done</b> Implement JITer. This is where our
current efforts are focused on, the JITer currently runs
all of the code we have tested on it. The major limitation
is that our class libraries are not complete, and hence not
every application can be ran.
* Milestone 4: Implement JITer.
* Milestone 5: Port of the JITer to non IA32 systems.
</ul>
A setup similar to the Kaffe JIT engine will be used to
A setup similar to the Kaffe JIT engine can be used to
layout the code to support non-IA32 architectures. Our work
will be focused on getting a IA32 version running first.
The JIT engine should work on Linux and Win32, although you
will need to install the CygWin32 development tools to get a
Unix-like compilation environment (mostly we use GNU make in
a few of the makefiles).
Unix-like compilation environment.
** JIT Engine (<b>updated, Dec 18th, 2001</b>)
** JIT Engine (<b>updated, Jul 14th, 2001</b>)
The JIT engine uses a code-generator generator approach for
compilation. Given the properties of CIL byte codes, we can
take full advantage of a real instruction selector for our
code generator.
We will be using a code-generator generator approach for our
JITer. Given the properties of CIL byte codes, we can take
full advantage of a real instruction selector for our code
generator.
There are a couple of books that deal with this technique: "A
Retargetable C Compiler" and "Advanced Compiler Design and
Implementation" are good references. You can also get a
technical description of <a
href="http://research.microsoft.com/copyright/accept.asp?path=http://www.research.microsoft.com/~drh/pubs/iburg.pdf&pub=ACM">lbrug</a>.
href="http://research.microsoft.com/copyright/accept.asp?path=http://www.research.microsoft.com/~drh/pubs/iburg.pdf&pub=ACM">lbrug</a>
A few papers that describe the instruction selector:
Previously we had looked at a number of JIT engines and tools,
but they would not take full advantage of the CIL properties:
<ul>
* <a href="http://research.microsoft.com/copyright/accept.asp?path=http://www.research.microsoft.com/~drh/pubs/interface.pdf&pub=wiley">A code generation interface for ANSI C</a>
* <a
href="http://www.intel.com/research/mrl/orp/">ORP</a>
* <a
href="http://www.gnu.org/software/lightning/">GNU
Lightning</a>
* <a href="http://research.microsoft.com/copyright/accept.asp?path=http://www.research.microsoft.com/~drh/pubs/iburg.pdf&pub=ACM">Engineering efficient code generators using tree matching and dynamic programming.</a>
* <a href="http://www.eecs.harvard.edu/~nr/toolkit/">NJ Machine
Toolkit</a>.).
* VCODE.
</ul>
** Garbage Collection
We will be using the Intel ORP GC engine as it provides a precise
garbage collector engine, similar to what is available on the
.NET environment.
We have decided to implement a generational tracing garbage
collector, which is very similar to the one being used by
.NET. For an introduction to the garbage collection system
used by Microsoft's CLR implementation, you can read this book
on <a
href="http://www.amazon.com/exec/obidos/ASIN/0471941484/o/qid=992556433/sr=2-1/ref=aps_sr_b_1_1/103-5866388-0492603">Garbage
Collection.</a>
Another consideration is to use the same interface that ORP
uses to its Garbage Collection system and reuse that GC system
instead of rolling our own, as the ORP system is pretty advanced
and is independent of the rest of ORP.
Although using a conservative garbage collector like Bohem's
would work, all the type information is available at runtime,
@@ -116,15 +114,6 @@
<a href="http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnmag00/html/GCI2.asp">http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dnmag00/html/GCI2.asp</a>
</ul>
** IO and threading
The ECMA runtime and the .NET runtime assume an IO model and a
threading model that is very similar to the Win32 API.
Dick Porter has been working on the Mono abstraction layer
that allows our runtime to execute code that depend on this
behaviour.
** Useful links
Paolo Molaro found a few interesting links:
@@ -149,5 +138,7 @@
PInvoke is the mechanism we are using to wrap Unix API calls
as well as talking to system libraries.
Initially we used libffi, but it was fairly slow, so we have
reused parts of the JIT work to create efficient PInvoke trampolines.
We hvae implemented PInvoke through libffi, but we are likely
going to roll our own system as the runtime matures, specially
as the interpreter is approaching completion, and we move into
the JITer.
+22 -24
View File
@@ -4,37 +4,35 @@
Pieces of Mono that have been implemented:
<ul>
* C# compiler: The C# compiler is mostly complete, a
few features are still missing (for a list, check the
<a href="c-sharp.html">C# compiler</a> web page.
* C# compiler: The C# parser can now generate parse
trees. Next up: semantic analysis, compiler lever optimizations
and code generation.
* The Mono JIT engine (<b>mono</b>. The JIT engine is
the real
virtual machine that we support. It is
functionaly currently, and we are improving its speed,
and making it feature complete.
* Metadata library: Can currently parse and load
information from .NET modules (executables and DLL
files).
* A CIL bytecode interpreter (<b>mint</b>). This
currently can be used to run most .NET programs
(modulo the fact that our class libraries are not
complete enough).
<b>mint</b> can currently run the Mono C# compiler and
the compiler generates valid code.
We use <b>mint</b> as a reference implementation of
the runtime environment and to quickly support
non-Intel platforms.
* Metadata library: Can currently parse
and load information from .NET modules (executables
and DLL files).
Parsing of Exception tables is missing.
* Disassembler: Can disassemble .NET modules. Still
lacking exception handling as well as useful debugging
tools (hex dumping, token dumping).
* Class Libraries: You can check the current status in the
<a href="class-status.html">Class Status page</a>.
<a href="class-status.html">Class Status page</a>
</ul>
Tasks on the critical path:
<ul>
* Bytecode interpreter: (Paolo is working on this).
* Simple JIT: Not implemented yet (Miguel is
researching the code-generator generator)
* Garbage collection engine (Dick is working on this, his contact is <a
href="mailto:[email protected]">[email protected]</a>
</ul>
If you want to work on any task here, please mail <a
href="mailto:[email protected]">[email protected]</a>
+2 -27
View File
@@ -1,12 +1,4 @@
* Testing
Testing is an important part of the Mono project: every one of its
three major components has a test suite tailored for its needs. This
is very helpful, because in the course of developing the software it
is very common to introduce bugs in existing code. A test suite
helps us fix the bugs as soon as they are introduced.
** Class Library Tests
* Testing
All classes in Mono libraries should have comprehensive unit test
suites to go with them. Unit testing is a software engineering
@@ -25,22 +17,5 @@
be written to demonstrate the bug and its fix. Then, if
you ever reintroduce the bug, you will know immediately. For
more info, read <a
href="http://junit.sourceforge.net/doc/testinfected/testing.htm">
href="http://nunit.sourceforge.net/doc/testinfected/testing.html">
JUnit Test Infected: Programmers Love Writing Tests</a>.
** Getting Started
We welcome all contributions to the Class Libary Test Suite.
There is information to help you get started in CVS at
mcs/class/doc/NUnitGuidelines. Once you have written your test, please
post it to <a href="resources.html#mailing">mono-list</a>.
Someone will make sure to add the file or apply the patch as
appropriate. If you plan to be an on-going contributor and
would like to get cvs account, email <a href="mailto:[email protected]">miguel</a>.
Normally, after you send a couple of well-written new files
and/or patches to the list, you will be given cvs access.
+1 -6
View File
@@ -3,19 +3,14 @@
0,Mono,rationale.html,rationale.src
1,Runtime,runtime.html,runtime.src
1,Classes,class-library.html,class-library.src
1,Gtk#,http://gtk-sharp.sourceforge.net
2,Class Status,class-status.html,class-status.src
1,C# Compiler,c-sharp.html,c-sharp.src
1,Status,status.html,status.src
0,Download,download.html,download.src
1,CVS access,ccvs.html,ccvs.src
1,AnonCVS access,anoncvs.html,anoncvs.src
0,Contributing,contributing.html,contributing.src
1,Documentation,documentation.html,documentation.src
1,Class Docs,classlib-doc.html,classlib-doc.src
1,Test Suite,testing.html,testing.src
1,Tools,tools.html,tools.src
1,Porting,porting.html,porting.src
0,Download,download.html,download.src
0,Resources,resources.html,resources.src
1,Ideas,ideas.html,ideas.src
1,Passport,passport.html,passport.src
-76
View File
@@ -1,76 +0,0 @@
#!/usr/bin/perl
$q = 1;
# Modified by Edwin Lima ([email protected]; [email protected])
# Date: 08/21/01, The Netherlands
# $f: Variable used as a flag to create the list of questions on top of the question-answers set. This
# is the only way that I found to scan the questions which have a <CR><LF>, in such a way that I will not
# scan the answer together at same time.
# @aname: Buffer used to store the question-answers set to print them out just after the print of the
# questions.
# @href: Buffer used to store the anchors (only questions) to the questions-answers set on the bottom
# of the page.
# I opened explicitly the file for input (input.txt) but U can change this as it was originally.
#
#
#comment this line if you are not open the file explicitly
#open(IN, "input.txt") || die "cannot open file input.txt" ;
print("<A name=TOP>") ;
#Uncomment line bellow to make it work as it was originally.
while (<>){
#comment line bellow
#while (<IN>){
chop;
if (/^\* (.*)$/){
push(@aname,"<h1>$1</h1>\n");
# print $body;
} elsif (/^\*\* (.*)$/) {
push(@aname, "<h2>$1</h2>\n");
push(@href, "<h2>$1</h2>\n");
} elsif (/^\*\*\* (.*)$/) {
push(@aname, "<h3>$1</h3>\n");
} elsif (/^$/) {
push(@aname, "<p>\n");
# push(@href, "<p>\n"); NOT NEEDED
} elsif (/^\t\t\* (.*)$/) {
push(@aname, "<li>$1\n");
} elsif (/^Q: (.*)$/){
push(@aname, "<p><a name=\"q$q\"></a><b>Question $q:</b> $1\n");
push(@href,"<p><a href=\"#q$q\"><b>Question $q:</b></a> $1\n");
$f=1;
$q++;
} elsif (/^A: (.*)$/){
push(@aname,"<P>\n<A HREF=#TOP>Top</A>\n<P>");
push(@aname,"$1\n");
$f=0;
} elsif (/^TODO=(.*),$/){
push(@aname, "<a name=\"$1\">\n");
#push(@href, "<a name=\"$1\">\n");
} else {
push(@aname,"$_\n");
if ($f==1) {
push(@href,"$_\n");
}
}
}
foreach $line (@href) #"\n\n";
{
print $line;
}
foreach $line (@aname) #"\n\n";
{
print $line;
}
#comment this line if you are not open the file explicitly
# close(IN) || die "cannot close file" ;
+2 -10
View File
@@ -1,15 +1,9 @@
# Until we are ready:
# ../class-library \
SOURCES= \
../anoncvs \
../contributing \
../class-status \
../class-library \
../classlib-doc \
../class-status \
../contact \
../c-sharp \
../ccvs \
../documentation \
../download \
../faq \
@@ -23,7 +17,6 @@ SOURCES= \
../runtime \
../status \
../testing \
../porting \
../tools
all:
@@ -40,8 +33,7 @@ clean:
done;
push:
scp -P 2200 *.html [email protected]:/web/cvsmodules/mono
echo scp *.html [email protected]:/home/groups/m/mo/mononet/htdocs
scp *.html [email protected]:/web/cvsmodules/mono
push2:
scp *.html primates:public_html/xxx
+25 -27
View File
@@ -44,32 +44,30 @@ while (<COMMANDS>) {
chop;
my @command = split /,/;
if ($command[2] =~ /^http:/){
} else {
$n = $ARGV[2] . "/" . $command[2];
open OUTPUT, ">" . $n || die "Can not create $n";
my $content = "";
open INPUT, $command[3] || die "Can not open $command[3]";
while (<INPUT>) {
$content .= $_;
}
close INPUT;
my $line;
my $temp;
my $tit;
my $title;
$tit = $command[1];
foreach $line (@template) {
$temp = $line;
$title = "$tit / Mono";
$temp =~ s/#TITLE#/$title/;
$temp =~ s/#CONTENT#/$content/;
$temp =~ s/#MENU#/$menu/;
print OUTPUT $temp;
}
}
$n = $ARGV[2] . "/" . $command[2];
open OUTPUT, ">" . $n || die "Can not create $n";
my $content = "";
open INPUT, $command[3] || die "Can not open $command[3]";
while (<INPUT>) {
$content .= $_;
}
close INPUT;
my $line;
my $temp;
my $tit;
my $title;
$tit = $command[1];
foreach $line (@template) {
$temp = $line;
$title = "$tit / Mono";
$temp =~ s/#TITLE#/$title/;
$temp =~ s/#CONTENT#/$content/;
$temp =~ s/#MENU#/$menu/;
print OUTPUT $temp;
}
close OUTPUT;
}
+1 -1
View File
@@ -30,7 +30,7 @@
<tr>
<td><img src="images/pixel.gif"></td><!-- left border -->
<td colspan="4">
<a href="http://www.go-mono.com"><img src="images/mono.gif" border="0"></a></td>
<a href="/"><img src="images/mono.gif" border="0"></a></td>
<td><img src="images/pixel.gif"></td><!-- right border -->
</tr>
<tr>
-54
View File
@@ -1,54 +0,0 @@
Exception implementation (jit):
===============================
Stack unwinding:
================
We record the code address (start_address, size) of all methods. That way it is
possible to map an instruction pointer (IP) to the method information needed
for unwinding the stack:
void handle_exception ((struct sigcontext *ctx, gpointer obj)
{
info = mono_jit_info_table_find (mono_jit_info_table, ctx->ip);
if (info) { // we are inside managed code
if (ch = find_catch_handler ())
execute_catch_handler (ch, ctx, obj);
execute_all_finally_handler ();
// restore register, including IP and Frame pointer
restore_caller_saved_registers (ji, ctx);
// continue unwinding
handle_exception (ctx, obj);
} else {
// not implemented
}
}
Code generation:
================
leave: is simply translated into a branch to the target. If the leave
instruction is inside a finally block (but not inside another handler)
we call the finally handler before we branch to the target.
finally/endfinally: is translated into subroutine ending with a "return"
statement. The subroutine does not save EBP/ESP, because we need access to the
local variables of the enclosing method. We have to use a "call"
instruction to execute such finally handlers. This makes it possible to
execute them inside the stack unwinding code.
throw: we first save all regs into a sigcontext struct (we pass the
exception object in register ECX), and then call the stack unwinding
code.
catch handler: receives the exception object in ECX. They store that
object into a local variable, so that rethrow can access the object.
-47
View File
@@ -1,47 +0,0 @@
How to use the debug support in the jit.
----------------------------------------
You need to run mono inside gdb. The following command line swicthes
are available:
--stabs
--debug methodName
--stabs will create an assemblyname-stabs.s file for each assembly the
CLR program uses. Note that to properly display source code lines,
you need to disassemble the CLR executables with monodis before running
mono. Each IL assembly file needs to have the name <assemblyname>.il.
--debug methodName will insert a breakpoin at the beginning of
methodName's code, so that control is trasnfered to the debugger as soon
as it is entered. You may use this switch multiple times.
So, suppose you use the --debug switch, or hit a segfault inside a
jitted method. In a shell you need to compile the stab information
created with the --stabs option with the assembler:
as assemblyname-stabs.s -o assemblyname-stabs.o
Now, inside gdb, you can load the debug information with:
add-symbol-file assemblyname-stabs.o 0
And at this point the debugger should be able to print a correct
backtrace, you should be able to inspect method parameters and local
variables, disassemble methods and step through the code.
Note that apparently you can't unload a symbol file from gdb, so you
need to restart it if you restart the program to debug.
Name mangling.
-------------
Currently CLR methods are exposed as C functions, so, if they are not
static, their first argument will be called 'this'.
Method names are mangled in the following way: a long name is created
concatting namespace, class name and method name. '.' chars are changed
to underscaore '_'. To allow for overloading, the address of the
MonoMethod from wich the method was created is appended to the end of
the name (i's also handy since you can use the address in gdb to inspect
the MonoMethod).
-182
View File
@@ -1,182 +0,0 @@
Just some thoughts for the JITer:
General issues:
===============
We are designing a JIT compiler, so we have to consider two things:
- the quality of the generated code
- the time needed to generate that code
The current approach is to keep the JITer as simple as possible, and thus as
fast as possible. The generated code quality will suffer from that.
Register allocation is first done inside the trees of the forest, and each
tree can use the full set of registers. We simply split a tree if we get out of
registers, for example the following tree:
add(R0)
/ \
/ \
a(R0) add(R1)
/ \
/ \
b(R1) add(R2)
/ \
/ \
c(R2) b(R3)
can be transformed to:
stloc(t1) add(R0)
| / \
| / \
add(R0) a(R0) add(R1)
/ \ / \
/ \ / \
c(R0) b(R1) b(R1) t1(R2)
Please notice that the split trees use less registers than the original
tree.
Triggering JIT compilation:
===========================
The current approach is to call functions indirectly. The address to call is
stored in the MonoMethod structure. For each method we create a trampoline
function. When called, this function does the JIT compilation and replaces the
trampoline with the compiled method address.
We should consider using the CACAO approach, they do not use a trampoline at
all.
Register Allocation:
====================
With lcc you can assign a fixed register to a tree before register
allocation. For example this is needed by call, which return the value always
in EAX on x86. The current implementation works without such system, due to
special forest generation.
X86 Register Allocation:
========================
We can use 8bit or 16bit registers on the x86. If we use that feature we have
more registers to allocate, which maybe prevents some register spills. We
currently ignore that ability and always allocate 32 bit registers, because I
think we would gain very little from that optimisation and it would complicate
the code.
Different Register Sets:
========================
Most processors have more that one register set, at least one for floating
point values, and one for integers. Should we support architectures with more
that two sets? Does someone knows such an architecture?
64bit Integer Values:
=====================
I can imagine two different implementation. On possibility would be to treat
long (64bit) values simply like any other value type. This implies that we
call class methods for ALU operations like add or sub. Sure, this method will
be be a bit inefficient.
The more performant solution is to allocate two 32bit registers for each 64bit
value. We add a new non terminal to the monoburg grammar called long_reg. The
register allocation routines takes care of this non terminal and allocates two
32 bit registers for them.
Forest generation:
==================
Consider the following code:
OPCODE: STACK LOCALS
LDLOC.0 (5) [5,0]
LDC.1 (5,1) [5,0]
STLOC.0 (5) [1,0]
STLOC.1 () [1,5]
The current forest generation generates:
STLOC.0(LDC.1)
STLOC.1(LDLOC.0)
Which is wrong, since it stores the wrong value (1 instead of 5). Instead we
must generate something like:
STLOC.TMP(LDLOC.0)
STLOC.0(LDC.1)
STLOC.1(LDLOC.TMP)
Where STLOC.TMP saves the value into a new temporary variable.
We also need a similar solution for basic block boundaries when the stack depth
is not zero. We can simply save those values to new temporary values. Consider
the following basic block with one instruction:
LDLOC.1
This should generate a tree like:
STLOC.TMP(LDLOC.1) Please notice that an intelligent register allocator can
still allocate registers for those new variables.
DAG handling:
=============
Monoburg can't handle DAGs, instead we need real trees as input for
the code generator. So we have two problems:
1.) DUP instruction: This one is obvious - we need to store the value
into a temporary variable to solve the problem.
2.) function calls: Chapter 12.8, page 343 of "A retargetable C compiler"
explains that: "because listing a call node will give it a hidden reference
from the code list". I don't understand that (can someone explain that?), but
there is another reason to save return values to temporaries: Consider the
following code:
x = f(y) + g(z); // all functions return integers
We could generate such a tree for this expression: STLOC(ADD(CALL,CALL))
The problem is that both calls returns the value in the same register,
so it is non trivial to generate code for that tree. We must copy one
register into another one, which make register allocation more complex.
The easier solution is store the result of function calls to
temporaries. This leads to the following forest:
STLOC(CALL)
STLOC(CALL)
STLOC(ADD (LDLOC, LDLOC))
This is what lcc is doing, if I understood 12.8, page 342, 343?
Value Types:
============
The only CLI instructions which can handle value types are loads and stores,
either to local variable, to the stack or to array elements. Value types with a
size smaller than sizeof(int) are handled like any other basic type. For other
value types we load the base address and emit block copies to store them.
Possible Optimisations:
=======================
Miguel said ORP does some optimisation on IL level, for example moving array
bounds checking out of loops:
for (i = 0; i < N; i++) { check_range (a, i); a [i] = X; }
id transformed to:
if (in_range (a, 0, N)) { for (i = 0; i < N; i++) a[i] = X; }
else for (i = 0; i < N; i++) { check_range (a, i); a [i] = X; }
The "else" is only to keep original semantics (exception handling).
-58
View File
@@ -1,58 +0,0 @@
Object and VTable layout
========================
The first pointer inside an Object points to a MonoClass structure. Objects
also contains a MonoThreadsSync structure which is used by the mono Thread
implementation.
typedef struct {
MonoClass *class;
MonoThreadsSync synchronisation;
/* object specific data goes here */
} MonoObject;
The MonoClass contains all Class infos, the VTable and a pointer to static
class data.
typedef struct {
/* various class infos */
MonoClass *parent;
const char *name;
const char *name_space;
...
/* interface offset table */
gint *interface_offsets;
gpointer data; /* a pointer to static data */
/* the variable sized vtable is included at the end */
gpointer vtable [vtable_size];
} MonoClass;
Calling virtual functions:
==========================
Each MonoMethod (if virtual) has an associated slot, which is an index into the
VTable. So we can use the following code to compute the address of a virtual
function:
method_addr = object->class->vtable [method->slot];
Calling interface methods:
==========================
Each interface class is associated with an unique ID. The following code
computes the address of an interface function:
offset_into_vtable = object->class->interface_offsets [interface_id];
method_addr = object->class->vtable [offset_into_vtable + method->slot];
+16
View File
@@ -0,0 +1,16 @@
.deps
aclocal.m4
config.cache
config.guess
config.h
config.h.in
config.log
config.status
config.sub
configure
fficonfig.h
fficonfig.h.in
Makefile
Makefile.in
stamp-h
stamp-h.in
+440
View File
@@ -0,0 +1,440 @@
2001-08-16 Alex Graveley <[email protected]>
* libffi/include/fficonfig.h: Remove as this is auto generated.
* Makefile.am: Only generate libffi.a.
* configure.in: replace A[CM]_PROG_LIBTOOL with AC_PROG_RANLIB.
Sun Apr 22 15:50:08 2001 Anthony Green <[email protected]>
* include/ffi_common.h: Delete, after moving contents to...
* include/ffi_private.h: Subsume contents of ffi_common.h.
* include/Makefile.am (noinst_HEADERS): Remove ffi_common.h.
* include/Makefile.in: Rebuilt.
* arm/ffi.c, m68k/ffi.c, mips/ffi.c, powerpc/ffi.c, s390/ffi.c,
ia64/ffi.c: Include ffi_private.h, not ffi_common.h.
* alpha/ffi.c, sparc/ffi.c, x86/ffi.c: Don't include ffi_common.h.
* types.c, raw_api.c, java_raw_api.c, prep_cif.c: Don't include
ffi_common.h.
* debug.c: Include ffi_private.h instead of ffi_common.h.
* mips/ffi.c (calc_n32_struct_flags): Make static.
(FIX_ARGP): Remove call to debugging routine ffi_stop_here.
* mips/n32.S: Include ffi_private.h.
* mips/o32.S: Include ffi_private.h.
Sun Apr 22 12:37:44 2001 Anthony Green <[email protected]>
* README: Update some comments.
* Makefile.am (SUBDIRS): Add include so ffi.h gets installed.
* Makefile.in: Rebuilt.
* include/ffi.h: Change ALPHA to __alpha__ and SPARC to __sparc__.
* types.c: Ditto.
* prep_cif.c (ffi_prep_cif): Ditto.
* alpha/ffi.c, alpha/osf.S, sparc/ffi.c, sparc/v8.S, sparc/v9.S:
Include ffi_private.h.
* include/ffi_private.h (FFI_TYPE_LAST): Define.
Sun Apr 22 11:41:55 2001 Anthony Green <[email protected]>
* Revolutionary changes. Target independent ffi.h. dejagnu
testing infrastructure. Moved source files around. Configure no
longer depends on run tests.
Sat Apr 14 20:49:46 2001 Anthony Green <[email protected]>
* mips/ffi.c: Include mips/mips.h.
* mips/o32.S: Ditto.
* mips/n32.S: Ditto.
* mips/o32.s: Deleted.
* mips/n32.s: Deleted.
Sat Apr 14 20:47:13 2001 Anthony Green <[email protected]>
* mips/mips.h: Moved from include/ffi_mips.h.
Sat Apr 14 20:09:05 2001 Anthony Green <[email protected]>
* powerpc/asm.h: Remove rcs id.
Sat Apr 14 15:04:08 2001 Anthony Green <[email protected]>
* configure.in: Use AC_COMPILE_CHECK_SIZEOF and
AC_C_BIGENDIAN_CROSS.
Sat Apr 14 15:02:53 2001 Anthony Green <[email protected]>
* aclocal.m4: Rebuilt.
* acinclude.m4: New file. Add AC_COMPILE_CHECK_SIZEOF and
AC_C_BIGENDIAN_CROSS.
2001-04-8 Neale Ferguson <[email protected]>
* src/s390/ffi.c: New file.
* src/s390/sysv.S: New file.
* include/ffi.h: Add s390 bits.
* configure.in: Ditto.
* configure: Rebuild.
2001-03-29 Bryce McKinlay <[email protected]>
* configure.in: Use different syntax for subdirectory creation.
* configure: Rebuilt.
2001-03-27 Jon Beniston <[email protected]>
* configure.in: Added X86_WIN32 target (Win32, CygWin, MingW).
* configure: Rebuilt.
* Makefile.am: Added X86_WIN32 target support.
* Makefile.in: Rebuilt.
* include/ffi.h.in: Added X86_WIN32 target support.
* src/ffitest.c: Doesn't run structure tests for X86_WIN32 targets.
* src/types.c: Added X86_WIN32 target support.
* src/x86/win32.S: New file. Based on sysv.S, but with EH
stuff removed and made to work with CygWin's gas.
2001-03-26 Bryce McKinlay <[email protected]>
* configure.in: Make target subdirectory in build dir.
* Makefile.am: Override suffix based rules to specify correct output
subdirectory.
* Makefile.in: Rebuilt.
* configure: Rebuilt.
2001-03-23 Kevin B Hendricks <[email protected]>
* src/powerpc/ppc_closure.S: New file.
* src/powerpc/ffi.c (ffi_prep_args): Fixed ABI compatibility bug
involving long long and register pairs.
(ffi_prep_closure): New function.
(flush_icache): Likewise.
(ffi_closure_helper_SYSV): Likewise.
* include/ffi.h.in (FFI_CLOSURES): Define on PPC.
(FFI_TRAMPOLINE_SIZE): Likewise.
(FFI_NATIVE_RAW_API): Likewise.
* Makefile.in: Rebuilt.
* Makefile.am (EXTRA_DIST): Added src/powerpc/ppc_closure.S.
(TARGET_SRC_POWERPC): Likewise.
2001-03-19 Tom Tromey <[email protected]>
* Makefile.in: Rebuilt.
* Makefile.am (ffitest_LDFLAGS): New macro.
2001-03-02 Nick Clifton <[email protected]>
* include/ffi.h.in: Remove RCS ident string.
* include/ffi_mips.h: Remove RCS ident string.
* src/debug.c: Remove RCS ident string.
* src/ffitest.c: Remove RCS ident string.
* src/prep_cif.c: Remove RCS ident string.
* src/types.c: Remove RCS ident string.
* src/alpha/ffi.c: Remove RCS ident string.
* src/alpha/osf.S: Remove RCS ident string.
* src/arm/ffi.c: Remove RCS ident string.
* src/arm/sysv.S: Remove RCS ident string.
* src/mips/ffi.c: Remove RCS ident string.
* src/mips/n32.S: Remove RCS ident string.
* src/mips/o32.S: Remove RCS ident string.
* src/sparc/ffi.c: Remove RCS ident string.
* src/sparc/v8.S: Remove RCS ident string.
* src/sparc/v9.S: Remove RCS ident string.
* src/x86/ffi.c: Remove RCS ident string.
* src/x86/sysv.S: Remove RCS ident string.
2001-02-08 Joseph S. Myers <[email protected]>
* include/ffi.h.in: Change sourceware.cygnus.com references to
gcc.gnu.org.
2000-12-09 Richard Henderson <[email protected]>
* src/alpha/ffi.c (ffi_call): Simplify struct return test.
(ffi_closure_osf_inner): Index rather than increment avalue
and arg_types. Give ffi_closure_osf the raw return value type.
* src/alpha/osf.S (ffi_closure_osf): Handle return value type
promotion.
2000-12-07 Richard Henderson <[email protected]>
* src/raw_api.c (ffi_translate_args): Fix typo.
(ffi_prep_closure): Likewise.
* include/ffi.h.in [ALPHA]: Define FFI_CLOSURES and
FFI_TRAMPOLINE_SIZE.
* src/alpha/ffi.c (ffi_prep_cif_machdep): Adjust minimal
cif->bytes for new ffi_call_osf implementation.
(ffi_prep_args): Absorb into ...
(ffi_call): ... here. Do all stack allocation here and
avoid a callback function.
(ffi_prep_closure, ffi_closure_osf_inner): New.
* src/alpha/osf.S (ffi_call_osf): Reimplement with no callback.
(ffi_closure_osf): New.
2000-09-10 Alexandre Oliva <[email protected]>
* config.guess, config.sub, install-sh: Removed.
* ltconfig, ltmain.sh, missing, mkinstalldirs: Likewise.
* Makefile.in: Rebuilt.
* acinclude.m4: Include libtool macros from the top level.
* aclocal.m4, configure: Rebuilt.
2000-08-22 Alexandre Oliva <[email protected]>
* configure.in [i*86-*-freebsd*] (TARGET, TARGETDIR): Set.
* configure: Rebuilt.
2000-05-11 Scott Bambrough <[email protected]>
* libffi/src/arm/sysv.S (ffi_call_SYSV): Doubles are not saved to
memory correctly. Use conditional instructions, not branches where
possible.
2000-05-04 Tom Tromey <[email protected]>
* configure: Rebuilt.
* configure.in: Match `arm*-*-linux-*'.
From Chris Dornan <[email protected]>.
2000-04-28 Jakub Jelinek <[email protected]>
* Makefile.am (SUBDIRS): Define.
(AM_MAKEFLAGS): Likewise.
(Multilib support.): Add section.
* Makefile.in: Rebuilt.
* ltconfig (extra_compiler_flags, extra_compiler_flags_value):
New variables. Set for gcc using -print-multi-lib. Export them
to libtool.
(sparc64-*-linux-gnu*): Use libsuff 64 for search paths.
* ltmain.sh (B|b|V): Don't throw away gcc's -B, -b and -V options
for -shared links.
(extra_compiler_flags_value, extra_compiler_flags): Check these
for extra compiler options which need to be passed down in
compiler_flags.
2000-04-16 Anthony Green <[email protected]>
* README: Credit updates. Update supported platorms.
* include/ffi.h.in: Important size fixes from William Knottenblet
([email protected]).
2000-04-16 Anthony Green <[email protected]>
* configure: Rebuilt.
* configure.in: Change i*86-pc-linux* to i*86-*-linux*.
2000-04-14 Jakub Jelinek <[email protected]>
* include/ffi.h.in (SPARC64): Define for 64bit SPARC builds.
Set SPARC FFI_DEFAULT_ABI based on SPARC64 define.
* src/sparc/ffi.c (ffi_prep_args_v8): Renamed from ffi_prep_args.
Replace all void * sizeofs with sizeof(int).
Only compare type with FFI_TYPE_LONGDOUBLE if LONGDOUBLE is
different than DOUBLE.
Remove FFI_TYPE_SINT32 and FFI_TYPE_UINT32 cases (handled elsewhere).
(ffi_prep_args_v9): New function.
(ffi_prep_cif_machdep): Handle V9 ABI and long long on V8.
(ffi_V9_return_struct): New function.
(ffi_call): Handle FFI_V9 ABI from 64bit code and FFI_V8 ABI from
32bit code (not yet cross-arch calls).
* src/sparc/v8.S: Add struct return delay nop.
Handle long long.
* src/sparc/v9.S: New file.
* src/prep_cif.c (ffi_prep_cif): Return structure pointer
is used on sparc64 only for structures larger than 32 bytes.
Pass by reference for structures is done for structure arguments
larger than 16 bytes.
* src/ffitest.c (main): Use 64bit rint on sparc64.
Run long long tests on sparc.
* src/types.c (FFI_TYPE_POINTER): Pointer is 64bit on alpha and
sparc64.
(FFI_TYPE_LONGDOUBLE): long double is 128 bit aligned to 128 bits
on sparc64.
* configure.in (sparc-*-linux*): New supported target.
(sparc64-*-linux*): Likewise.
* configure: Rebuilt.
* Makefile.am: Add v9.S to SPARC files.
* Makefile.in: Likewise.
(LINK): Surround $(CCLD) into double quotes, so that multilib
compiles work correctly.
2000-04-04 Alexandre Petit-Bianco <[email protected]>
* configure: Rebuilt.
* configure.in: (i*86-*-solaris*): New libffi target. Patch
proposed by Bryce McKinlay.
2000-03-20 Tom Tromey <[email protected]>
* Makefile.in: Hand edit for java_raw_api.lo.
2000-03-08 Bryce McKinlay <[email protected]>
* config.guess, config.sub: Update from the gcc tree.
Fix for PR libgcj/168.
2000-03-03 Tom Tromey <[email protected]>
* Makefile.in: Fixed ia64 by hand.
* configure: Rebuilt.
* configure.in (--enable-multilib): New option.
(libffi_basedir): New subst.
(AC_OUTPUT): Added multilib code.
2000-03-02 Tom Tromey <[email protected]>
* Makefile.in: Rebuilt.
* Makefile.am (TARGET_SRC_IA64): Use `ia64', not `alpha', as
directory name.
2000-02-25 Hans Boehm <[email protected]>
* src/ia64/ffi.c, src/ia64/ia64_flags.h, src/ia64/unix.S: New
files.
* src/raw_api.c (ffi_translate_args): Fixed typo in argument
list.
(ffi_prep_raw_closure): Use ffi_translate_args, not
ffi_closure_translate.
* src/java_raw_api.c: New file.
* src/ffitest.c (closure_test_fn): New function.
(main): Define `rint' as long long on IA64. Added new test when
FFI_CLOSURES is defined.
* include/ffi.h.in (ALIGN): Use size_t, not unsigned.
(ffi_abi): Recognize IA64.
(ffi_raw): Added `flt' field.
Added "Java raw API" code.
* configure.in: Recognize ia64.
* Makefile.am (TARGET_SRC_IA64): New macro.
(libffi_la_common_SOURCES): Added java_raw_api.c.
(libffi_la_SOURCES): Define in IA64 case.
2000-01-04 Tom Tromey <[email protected]>
* Makefile.in: Rebuilt with newer automake.
1999-12-31 Tom Tromey <[email protected]>
* Makefile.am (INCLUDES): Added -I$(top_srcdir)/src.
1999-09-01 Tom Tromey <[email protected]>
* include/ffi.h.in: Removed PACKAGE and VERSION defines and
undefs.
* fficonfig.h.in: Rebuilt.
* configure: Rebuilt.
* configure.in: Pass 3rd argument to AM_INIT_AUTOMAKE.
Use AM_PROG_LIBTOOL (automake 1.4 compatibility).
* acconfig.h: Don't #undef PACKAGE or VERSION.
1999-08-09 Anthony Green <[email protected]>
* include/ffi.h.in: Try to work around messy header problem
with PACKAGE and VERSION.
* configure: Rebuilt.
* configure.in: Change version to 2.00-beta.
* fficonfig.h.in: Rebuilt.
* acconfig.h (FFI_NO_STRUCTS, FFI_NO_RAW_API): Define.
* src/x86/ffi.c (ffi_raw_call): Rename.
Mon Aug 9 18:33:38 1999 Rainer Orth <[email protected]>
* include/Makefile.in: Rebuilt.
* Makefile.in: Rebuilt
* Makefile.am (toolexeclibdir): Add $(MULTISUBDIR) even for native
builds.
Use USE_LIBDIR.
* configure: Rebuilt.
* configure.in (USE_LIBDIR): Define for native builds.
Use lowercase in configure --help explanations.
1999-08-08 Anthony Green <[email protected]>
* include/ffi.h.in (FFI_FN): Remove `...'.
1999-08-08 Anthony Green <[email protected]>
* Makefile.in: Rebuilt.
* Makefile.am (AM_CFLAGS): Compile with -fexceptions.
* src/x86/sysv.S: Add exception handling metadata.
1999-08-02 Kresten Krab Thorup <[email protected]>
* src/x86/ffi.c (ffi_closure_SYSV): New function.
(ffi_prep_incoming_args_SYSV): Ditto.
(ffi_prep_closure): Ditto.
(ffi_closure_raw_SYSV): Ditto.
(ffi_prep_raw_closure): More ditto.
(ffi_call_raw): Final ditto.
* include/ffi.h.in: Add definitions for closure and raw API.
* src/x86/ffi.c (ffi_prep_cif_machdep): Added case for
FFI_TYPE_UINT64.
* Makefile.am (libffi_la_common_SOURCES): Added raw_api.c
* src/raw_api.c: New file.
* include/ffi.h.in (ffi_raw): New type.
(UINT_ARG, SINT_ARG): New defines.
(ffi_closure, ffi_raw_closure): New types.
(ffi_prep_closure, ffi_prep_raw_closure): New declarations.
* configure.in: Add check for endianness and sizeof void*.
* src/x86/sysv.S (ffi_call_SYSV): Call fixup routine via argument,
instead of directly.
* configure: Rebuilt.
Thu Jul 8 14:28:42 1999 Anthony Green <[email protected]>
* configure.in: Add x86 and powerpc BeOS configurations.
From Makoto Kato <[email protected]>.
1999-05-09 Anthony Green <[email protected]>
* configure.in: Add warning about this being beta code.
Remove src/Makefile.am from the picture.
* configure: Rebuilt.
* Makefile.am: Move logic from src/Makefile.am. Add changes
to support libffi as a target library.
* Makefile.in: Rebuilt.
* aclocal.m4, config.guess, config.sub, ltconfig, ltmain.sh:
Upgraded to new autoconf, automake, libtool.
* README: Tweaks.
* LICENSE: Update copyright date.
* src/Makefile.am, src/Makefile.in: Removed.
1998-11-29 Anthony Green <[email protected]>
* include/ChangeLog: Removed.
* src/ChangeLog: Removed.
* src/mips/ChangeLog: Removed.
* src/sparc/ChangeLog: Remboved.
* src/x86/ChangeLog: Removed.
* ChangeLog.v1: Created.
+764
View File
@@ -0,0 +1,764 @@
The libffi version 1 ChangeLog archive.
Version 1 of libffi had per-directory ChangeLogs. Current and future
versions have a single ChangeLog file in the root directory. The
version 1 ChangeLogs have all been concatonated into this file for
future reference only.
--- libffi ----------------------------------------------------------------
Mon Oct 5 02:17:50 1998 Anthony Green <[email protected]>
* configure.in: Boosted rev.
* configure, Makefile.in, aclocal.m4: Rebuilt.
* README: Boosted rev and updated release notes.
Mon Oct 5 01:03:03 1998 Anthony Green <[email protected]>
* configure.in: Boosted rev.
* configure, Makefile.in, aclocal.m4: Rebuilt.
* README: Boosted rev and updated release notes.
1998-07-25 Andreas Schwab <[email protected]>
* m68k/ffi.c (ffi_prep_cif_machdep): Use bitmask for cif->flags.
Correctly handle small structures.
(ffi_prep_args): Also handle small structures.
(ffi_call): Pass size of return type to ffi_call_SYSV.
* m68k/sysv.S: Adjust for above changes. Correctly align small
structures in the return value.
* types.c (uint64, sint64) [M68K]: Change alignment to 4.
Fri Apr 17 17:26:58 1998 Anthony Green <[email protected]>
* configure.in: Boosted rev.
* configure,Makefile.in,aclocal.m4: Rebuilt.
* README: Boosted rev and added release notes.
Sun Feb 22 00:50:41 1998 Geoff Keating <[email protected]>
* configure.in: Add PowerPC config bits.
1998-02-14 Andreas Schwab <[email protected]>
* configure.in: Add m68k config bits. Change AC_CANONICAL_SYSTEM
to AC_CANONICAL_HOST, this is not a compiler. Use $host instead
of $target. Remove AC_CHECK_SIZEOF(char), we already know the
result. Fix argument of AC_ARG_ENABLE.
* configure, fficonfig.h.in: Rebuilt.
Tue Feb 10 20:53:40 1998 Richard Henderson <[email protected]>
* configure.in: Add Alpha config bits.
Tue May 13 13:39:20 1997 Anthony Green <[email protected]>
* README: Updated dates and reworded Irix comments.
* configure.in: Removed AC_PROG_RANLIB.
* Makefile.in, aclocal.m4, config.guess, config.sub, configure,
ltmain.sh, */Makefile.in: libtoolized again and rebuilt with
automake and autoconf.
Sat May 10 18:44:50 1997 Tom Tromey <[email protected]>
* configure, aclocal.m4: Rebuilt.
* configure.in: Don't compute EXTRADIST; now handled in
src/Makefile.in. Removed macros implied by AM_INIT_AUTOMAKE.
Don't run AM_MAINTAINER_MODE.
Thu May 8 14:34:05 1997 Anthony Green <[email protected]>
* missing, ltmain.sh, ltconfig.sh: Created. These are new files
required by automake and libtool.
* README: Boosted rev to 1.14. Added notes.
* acconfig.h: Moved PACKAGE and VERSION for new automake.
* configure.in: Changes for libtool.
* Makefile.am (check): make test now make check. Uses libtool now.
* Makefile.in, configure.in, aclocal.h, fficonfig.h.in: Rebuilt.
Thu May 1 16:27:07 1997 Anthony Green <[email protected]>
* missing: Added file required by new automake.
Tue Nov 26 14:10:42 1996 Anthony Green <[email protected]>
* acconfig.h: Added USING_PURIFY flag. This is defined when
--enable-purify-safety was used at configure time.
* configure.in (allsources): Added --enable-purify-safety switch.
(VERSION): Boosted rev to 1.13.
* configure: Rebuilt.
Fri Nov 22 06:46:12 1996 Anthony Green <[email protected]>
* configure.in (VERSION): Boosted rev to 1.12.
Removed special CFLAGS hack for gcc.
* configure: Rebuilt.
* README: Boosted rev to 1.12. Added notes.
* Many files: Cygnus Support changed to Cygnus Solutions.
Wed Oct 30 11:15:25 1996 Anthony Green <[email protected]>
* configure.in (VERSION): Boosted rev to 1.11.
* configure: Rebuilt.
* README: Boosted rev to 1.11. Added notes about GNU make.
Tue Oct 29 12:25:12 1996 Anthony Green <[email protected]>
* configure.in: Fixed -Wall trick.
(VERSION): Boosted rev.
* configure: Rebuilt
* acconfig.h: Needed for --enable-debug configure switch.
* README: Boosted rev to 1.09. Added more notes on building
libffi, and LCLint.
* configure.in: Added --enable-debug switch. Boosted rev to
1.09.
* configure: Rebuilt
Tue Oct 15 13:11:28 1996 Anthony Green <[email protected]>
* configure.in (VERSION): Boosted rev to 1.08
* configure: Rebuilt.
* README: Added n32 bug fix notes.
* Makefile.am: Added "make lint" production.
* Makefile.in: Rebuilt.
Mon Oct 14 10:54:46 1996 Anthony Green <[email protected]>
* README: Added web page reference.
* configure.in, README: Boosted rev to 1.05
* configure: Rebuilt.
* README: Fixed n32 sample code.
Fri Oct 11 17:09:28 1996 Anthony Green <[email protected]>
* README: Added sparc notes.
* configure.in, README: Boosted rev to 1.04.
* configure: Rebuilt.
Thu Oct 10 10:31:03 1996 Anthony Green <[email protected]>
* configure.in, README: Boosted rev to 1.03.
* configure: Rebuilt.
* README: Added struct notes.
* Makefile.am (EXTRA_DIST): Added LICENSE to distribution.
* Makefile.in: Rebuilt.
* README: Removed Linux section. No special notes now
because aggregates arg/return types work.
Wed Oct 9 16:16:42 1996 Anthony Green <[email protected]>
* README, configure.in (VERSION): Boosted rev to 1.02
* configure: Rebuilt.
Tue Oct 8 11:56:33 1996 Anthony Green <[email protected]>
* README (NOTE): Added n32 notes.
* Makefile.am: Added test production.
* Makefile: Rebuilt
* README: spell checked!
* configure.in (VERSION): Boosted rev to 1.01
* configure: Rebuilt.
Mon Oct 7 15:50:22 1996 Anthony Green <[email protected]>
* configure.in: Added nasty bit to support SGI tools.
* configure: Rebuilt.
* README: Added SGI notes. Added note about automake bug.
Mon Oct 7 11:00:28 1996 Anthony Green <[email protected]>
* README: Rewrote intro, and fixed examples.
Fri Oct 4 10:19:55 1996 Anthony Green <[email protected]>
* configure.in: -D$TARGET is no longer used as a compiler switch.
It is now inserted into ffi.h at configure time.
* configure: Rebuilt.
* FFI_ABI and FFI_STATUS are now ffi_abi and ffi_status.
Thu Oct 3 13:47:34 1996 Anthony Green <[email protected]>
* README, LICENSE: Created. Wrote some docs.
* configure.in: Don't barf on i586-unknown-linuxaout.
Added EXTRADIST code for "make dist".
* configure: Rebuilt.
* */Makefile.in: Rebuilt with patched automake.
Tue Oct 1 17:12:25 1996 Anthony Green <[email protected]>
* Makefile.am, aclocal.m4, config.guess, config.sub,
configure.in, fficonfig.h.in, install-sh, mkinstalldirs,
stamp-h.in: Created
* Makefile.in, configure: Generated
--- libffi/include --------------------------------------------------------
Tue Feb 24 13:09:36 1998 Anthony Green <[email protected]>
* ffi_mips.h: Updated FFI_TYPE_STRUCT_* values based on
ffi.h.in changes. This is a work-around for SGI's "simple"
assembler.
Sun Feb 22 00:51:55 1998 Geoff Keating <[email protected]>
* ffi.h.in: PowerPC support.
1998-02-14 Andreas Schwab <[email protected]>
* ffi.h.in: Add m68k support.
(FFI_TYPE_LONGDOUBLE): Make it a separate value.
Tue Feb 10 20:55:16 1998 Richard Henderson <[email protected]>
* ffi.h.in (SIZEOF_ARG): Use a pointer type by default.
* ffi.h.in: Alpha support.
Fri Nov 22 06:48:45 1996 Anthony Green <[email protected]>
* ffi.h.in, ffi_common.h: Cygnus Support -> Cygnus Solutions.
Wed Nov 20 22:31:01 1996 Anthony Green <[email protected]>
* ffi.h.in: Added ffi_type_void definition.
Tue Oct 29 12:22:40 1996 Anthony Green <[email protected]>
* Makefile.am (hack_DATA): Always install ffi_mips.h.
* ffi.h.in: Removed FFI_DEBUG. It's now in the correct
place (acconfig.h).
Added #include <stddef.h> for size_t definition.
Tue Oct 15 17:23:35 1996 Anthony Green <[email protected]>
* ffi.h.in, ffi_common.h, ffi_mips.h: More clean up.
Commented out #define of FFI_DEBUG.
Tue Oct 15 13:01:06 1996 Anthony Green <[email protected]>
* ffi_common.h: Added bool definition.
* ffi.h.in, ffi_common.h: Clean up based on LCLint output.
Added funny /*@...@*/ comments to annotate source.
Mon Oct 14 12:29:23 1996 Anthony Green <[email protected]>
* ffi.h.in: Interface changes based on feedback from Jim
Blandy.
Fri Oct 11 16:49:35 1996 Anthony Green <[email protected]>
* ffi.h.in: Small change for sparc support.
Thu Oct 10 14:53:37 1996 Anthony Green <[email protected]>
* ffi_mips.h: Added FFI_TYPE_STRUCT_* definitions for
special structure return types.
Wed Oct 9 13:55:57 1996 Anthony Green <[email protected]>
* ffi.h.in: Added SIZEOF_ARG definition for X86
Tue Oct 8 11:40:36 1996 Anthony Green <[email protected]>
* ffi.h.in (FFI_FN): Added macro for eliminating compiler warnings.
Use it to case your function pointers to the proper type.
* ffi_mips.h (SIZEOF_ARG): Added magic to fix type promotion bug.
* Makefile.am (EXTRA_DIST): Added ffi_mips.h to EXTRA_DIST.
* Makefile: Rebuilt.
* ffi_mips.h: Created. Moved all common mips definitions here.
Mon Oct 7 10:58:12 1996 Anthony Green <[email protected]>
* ffi.h.in: The SGI assember is very picky about parens. Redefined
some macros to avoid problems.
* ffi.h.in: Added FFI_DEFAULT_ABI definitions. Also added
externs for pointer, and 64bit integral ffi_types.
Fri Oct 4 09:51:37 1996 Anthony Green <[email protected]>
* ffi.h.in: Added FFI_ABI member to ffi_cif and changed
function prototypes accordingly.
Added #define @TARGET@. Now programs including ffi.h don't
have to specify this themselves.
Thu Oct 3 15:36:44 1996 Anthony Green <[email protected]>
* ffi.h.in: Changed ffi_prep_cif's values from void* to void**
* Makefile.am (EXTRA_DIST): Added EXTRA_DIST for "make dist"
to work.
* Makefile.in: Regenerated.
Wed Oct 2 10:16:59 1996 Anthony Green <[email protected]>
* Makefile.am: Created
* Makefile.in: Generated
* ffi_common.h: Added rcsid comment
Tue Oct 1 17:13:51 1996 Anthony Green <[email protected]>
* ffi.h.in, ffi_common.h: Created
--- libffi/src ------------------------------------------------------------
Mon Oct 5 02:17:50 1998 Anthony Green <[email protected]>
* arm/ffi.c, arm/sysv.S: Created.
* Makefile.am: Added arm files.
* Makefile.in: Rebuilt.
Mon Oct 5 01:41:38 1998 Anthony Green <[email protected]>
* Makefile.am (libffi_la_LDFLAGS): Incremented revision.
Sun Oct 4 16:27:17 1998 Anthony Green <[email protected]>
* alpha/osf.S (ffi_call_osf): Patch for DU assembler.
* ffitest.c (main): long long and long double return values work
for x86.
Fri Apr 17 11:50:58 1998 Anthony Green <[email protected]>
* Makefile.in: Rebuilt.
* ffitest.c (main): Floating point tests not executed for systems
with broken lond double (SunOS 4 w/ GCC).
* types.c: Fixed x86 alignment info for long long types.
Thu Apr 16 07:15:28 1998 Anthony Green <[email protected]>
* ffitest.c: Added more notes about GCC bugs under Irix 6.
Wed Apr 15 08:42:22 1998 Anthony Green <[email protected]>
* ffitest.c (struct5): New test function.
(main): New test with struct5.
Thu Mar 5 10:48:11 1998 Anthony Green <[email protected]>
* prep_cif.c (initialize_aggregate): Fix assertion for
nested structures.
Tue Feb 24 16:33:41 1998 Anthony Green <[email protected]>
* prep_cif.c (ffi_prep_cif): Added long double support for sparc.
Sun Feb 22 00:52:18 1998 Geoff Keating <[email protected]>
* powerpc/asm.h: New file.
* powerpc/ffi.c: New file.
* powerpc/sysv.S: New file.
* Makefile.am: PowerPC port.
* ffitest.c (main): Allow all tests to run even in presence of gcc
bug on PowerPC.
1998-02-17 Anthony Green <[email protected]>
* mips/ffi.c: Fixed comment typo.
* x86/ffi.c (ffi_prep_cif_machdep), x86/sysv.S (retfloat):
Fixed x86 long double return handling.
* types.c: Fixed x86 long double alignment info.
1998-02-14 Andreas Schwab <[email protected]>
* types.c: Add m68k support.
* ffitest.c (floating): Add long double parameter.
(return_ll, ldblit): New functions to test long long and long
double return value.
(main): Fix type error in assignment of ts[1-4]_type.elements.
Add tests for long long and long double arguments and return
values.
* prep_cif.c (ffi_prep_cif) [M68K]: Don't allocate argument for
struct value pointer.
* m68k/ffi.c, m68k/sysv.S: New files.
* Makefile.am: Add bits for m68k port. Add kludge to work around
automake deficiency.
(test): Don't require "." in $PATH.
* Makefile.in: Rebuilt.
Wed Feb 11 07:36:50 1998 Anthony Green <[email protected]>
* Makefile.in: Rebuilt.
Tue Feb 10 20:56:00 1998 Richard Henderson <[email protected]>
* alpha/ffi.c, alpha/osf.S: New files.
* Makefile.am: Alpha port.
Tue Nov 18 14:12:07 1997 Anthony Green <[email protected]>
* mips/ffi.c (ffi_prep_cif_machdep): Initialize rstruct_flag
for n32.
Tue Jun 3 17:18:20 1997 Anthony Green <[email protected]>
* ffitest.c (main): Added hack to get structure tests working
correctly.
Sat May 10 19:06:42 1997 Tom Tromey <[email protected]>
* Makefile.in: Rebuilt.
* Makefile.am (EXTRA_DIST): Explicitly list all distributable
files in subdirs.
(VERSION, CC): Removed.
Thu May 8 17:19:01 1997 Anthony Green <[email protected]>
* Makefile.am: Many changes for new automake and libtool.
* Makefile.in: Rebuilt.
Fri Nov 22 06:57:56 1996 Anthony Green <[email protected]>
* ffitest.c (main): Fixed test case for non mips machines.
Wed Nov 20 22:31:59 1996 Anthony Green <[email protected]>
* types.c: Added ffi_type_void declaration.
Tue Oct 29 13:07:19 1996 Anthony Green <[email protected]>
* ffitest.c (main): Fixed character constants.
(main): Emit warning for structure test 3 failure on Sun.
* Makefile.am (VPATH): Fixed VPATH def'n so automake won't
strip it out.
Moved distdir hack from libffi to automake.
(ffitest): Added missing -c for $(COMPILE) (change in automake).
* Makefile.in: Rebuilt.
Tue Oct 15 13:08:20 1996 Anthony Green <[email protected]>
* Makefile.am: Added "make lint" production.
* Makefile.in: Rebuilt.
* prep_cif.c (STACK_ARG_SIZE): Improved STACK_ARG_SIZE macro.
Clean up based on LCLint output. Added funny /*@...@*/ comments to
annotate source.
* ffitest.c, debug.c: Cleaned up code.
Mon Oct 14 12:26:56 1996 Anthony Green <[email protected]>
* ffitest.c: Changes based on interface changes.
* prep_cif.c (ffi_prep_cif): Cleaned up interface based on
feedback from Jim Blandy.
Fri Oct 11 15:53:18 1996 Anthony Green <[email protected]>
* ffitest.c: Reordered tests while porting to sparc.
Made changes to handle lame structure passing for sparc.
Removed calls to fflush().
* prep_cif.c (ffi_prep_cif): Added special case for sparc
aggregate type arguments.
Thu Oct 10 09:56:51 1996 Anthony Green <[email protected]>
* ffitest.c (main): Added structure passing/returning tests.
* prep_cif.c (ffi_prep_cif): Perform proper initialization
of structure return types if needed.
(initialize_aggregate): Bug fix
Wed Oct 9 16:04:20 1996 Anthony Green <[email protected]>
* types.c: Added special definitions for x86 (double doesn't
need double word alignment).
* ffitest.c: Added many tests
Tue Oct 8 09:19:22 1996 Anthony Green <[email protected]>
* prep_cif.c (ffi_prep_cif): Fixed assertion.
* debug.c (ffi_assert): Must return a non void now.
* Makefile.am: Added test production.
* Makefile: Rebuilt.
* ffitest.c (main): Created.
* types.c: Created. Stripped common code out of */ffi.c.
* prep_cif.c: Added missing stdlib.h include.
* debug.c (ffi_type_test): Used "a" to eliminate compiler
warnings in non-debug builds. Included ffi_common.h.
Mon Oct 7 15:36:42 1996 Anthony Green <[email protected]>
* Makefile.am: Added a rule for .s -> .o
This is required by the SGI compiler.
* Makefile: Rebuilt.
Fri Oct 4 09:51:08 1996 Anthony Green <[email protected]>
* prep_cif.c (initialize_aggregate): Moved abi specification
to ffi_prep_cif().
Thu Oct 3 15:37:37 1996 Anthony Green <[email protected]>
* prep_cif.c (ffi_prep_cif): Changed values from void* to void**.
(initialize_aggregate): Fixed aggregate type initialization.
* Makefile.am (EXTRA_DIST): Added support code for "make dist".
* Makefile.in: Regenerated.
Wed Oct 2 11:41:57 1996 Anthony Green <[email protected]>
* debug.c, prep_cif: Created.
* Makefile.am: Added debug.o and prep_cif.o to OBJ.
* Makefile.in: Regenerated.
* Makefile.am (INCLUDES): Added missing -I../include
* Makefile.in: Regenerated.
Tue Oct 1 17:11:51 1996 Anthony Green <[email protected]>
* error.c, Makefile.am: Created.
* Makefile.in: Generated.
--- libffi/src/x86 --------------------------------------------------------
Sun Oct 4 16:27:17 1998 Anthony Green <[email protected]>
* sysv.S (retlongdouble): Fixed long long return value support.
* ffi.c (ffi_prep_cif_machdep): Ditto.
Wed May 13 04:30:33 1998 Anthony Green <[email protected]>
* ffi.c (ffi_prep_cif_machdep): Fixed long double return value
support.
Wed Apr 15 08:43:20 1998 Anthony Green <[email protected]>
* ffi.c (ffi_prep_args): small struct support was missing.
Thu May 8 16:53:58 1997 Anthony Green <[email protected]>
* objects.mak: Removed.
Mon Dec 2 15:12:58 1996 Tom Tromey <[email protected]>
* sysv.S: Use .balign, for a.out Linux boxes.
Tue Oct 15 13:06:50 1996 Anthony Green <[email protected]>
* ffi.c: Clean up based on LCLint output.
Added funny /*@...@*/ comments to annotate source.
Fri Oct 11 16:43:38 1996 Anthony Green <[email protected]>
* ffi.c (ffi_call): Added assertion for bad ABIs.
Wed Oct 9 13:57:27 1996 Anthony Green <[email protected]>
* sysv.S (retdouble): Fixed double return problems.
* ffi.c (ffi_call): Corrected fn arg definition.
(ffi_prep_cif_machdep): Fixed double return problems
Tue Oct 8 12:12:49 1996 Anthony Green <[email protected]>
* ffi.c: Moved ffi_type definitions to types.c.
(ffi_prep_args): Fixed type promotion bug.
Mon Oct 7 15:53:06 1996 Anthony Green <[email protected]>
* ffi.c (FFI_*_TYPEDEF): Removed redundant ';'
Fri Oct 4 09:54:53 1996 Anthony Green <[email protected]>
* ffi.c (ffi_call): Removed FFI_ABI arg, and swapped
remaining args.
Wed Oct 2 10:07:05 1996 Anthony Green <[email protected]>
* ffi.c, sysv.S, objects.mak: Created.
(ffi_prep_cif): cif->rvalue no longer initialized to NULL.
(ffi_prep_cif_machdep): Moved machine independent cif processing
to src/prep_cif.c. Introduced ffi_prep_cif_machdep().
--- libffi/src/mips -------------------------------------------------------
Tue Feb 17 17:18:07 1998 Anthony Green <[email protected]>
* o32.S: Fixed typo in comment.
* ffi.c (ffi_prep_cif_machdep): Fixed argument processing.
Thu May 8 16:53:58 1997 Anthony Green <[email protected]>
* o32.s, n32.s: Wrappers for SGI tool support.
* objects.mak: Removed.
Tue Oct 29 14:37:45 1996 Anthony Green <[email protected]>
* ffi.c (ffi_prep_args): Changed int z to size_t z.
Tue Oct 15 13:17:25 1996 Anthony Green <[email protected]>
* n32.S: Fixed bad stack munging.
* ffi.c: Moved prototypes for ffi_call_?32() to here from
ffi_mips.h because extended_cif is not defined in ffi_mips.h.
Mon Oct 14 12:42:02 1996 Anthony Green <[email protected]>
* ffi.c: Interface changes based on feedback from Jim Blandy.
Thu Oct 10 11:22:16 1996 Anthony Green <[email protected]>
* n32.S, ffi.c: Lots of changes to support passing and
returning structures with the n32 calling convention.
* n32.S: Fixed fn pointer bug.
* ffi.c (ffi_prep_cif_machdep): Fix for o32 structure
return values.
(ffi_prep_args): Fixed n32 structure passing when structures
partially fit in registers.
Wed Oct 9 13:49:25 1996 Anthony Green <[email protected]>
* objects.mak: Added n32.o.
* n32.S: Created.
* ffi.c (ffi_prep_args): Added magic to support proper
n32 processing.
Tue Oct 8 10:37:35 1996 Anthony Green <[email protected]>
* ffi.c: Moved ffi_type definitions to types.c.
(ffi_prep_args): Fixed type promotion bug.
* o32.S: This code is only built for o32 compiles.
A lot of the #define cruft has moved to ffi_mips.h.
* ffi.c (ffi_prep_cif_machdep): Fixed arg flags. Second arg
is only processed if the first is either a float or double.
Mon Oct 7 15:33:59 1996 Anthony Green <[email protected]>
* o32.S: Modified to compile under each of o32, n32 and n64.
* ffi.c (FFI_*_TYPEDEF): Removed redundant ';'
Fri Oct 4 09:53:25 1996 Anthony Green <[email protected]>
* ffi.c (ffi_call): Removed FFI_ABI arg, and swapped
remaining args.
Wed Oct 2 17:41:22 1996 Anthony Green <[email protected]>
* o32.S: Removed crufty definitions.
Wed Oct 2 12:53:42 1996 Anthony Green <[email protected]>
* ffi.c (ffi_prep_cif): cif->rvalue no longer initialized to NULL.
(ffi_prep_cif_machdep): Moved all machine independent cif processing
to src/prep_cif.c. Introduced ffi_prep_cif_machdep. Return types
of FFI_TYPE_STRUCT are no different than FFI_TYPE_INT.
Tue Oct 1 17:11:02 1996 Anthony Green <[email protected]>
* ffi.c, o32.S, object.mak: Created
--- libffi/src/sparc ------------------------------------------------------
Tue Feb 24 16:33:18 1998 Anthony Green <[email protected]>
* ffi.c (ffi_prep_args): Added long double support.
Thu May 8 16:53:58 1997 Anthony Green <[email protected]>
* objects.mak: Removed.
Thu May 1 16:07:56 1997 Anthony Green <[email protected]>
* v8.S: Fixed minor portability problem reported by
Russ McManus <[email protected]>.
Tue Nov 26 14:12:43 1996 Anthony Green <[email protected]>
* v8.S: Used STACKFRAME define elsewhere.
* ffi.c (ffi_prep_args): Zero out space when USING_PURIFY
is set.
(ffi_prep_cif_machdep): Allocate the correct stack frame
space for functions with < 6 args.
Tue Oct 29 15:08:55 1996 Anthony Green <[email protected]>
* ffi.c (ffi_prep_args): int z is now size_t z.
Mon Oct 14 13:31:24 1996 Anthony Green <[email protected]>
* v8.S (ffi_call_V8): Gordon rewrites this again. It looks
great now.
* ffi.c (ffi_call): The comment about hijacked registers
is no longer valid after gordoni hacked v8.S.
* v8.S (ffi_call_V8): Rewrote with gordoni. Much simpler.
* v8.S, ffi.c: ffi_call() had changed to accept more than
two args, so v8.S had to change (because it hijacks incoming
arg registers).
* ffi.c: Interface changes based on feedback from Jim Blandy.
Thu Oct 10 17:48:16 1996 Anthony Green <[email protected]>
* ffi.c, v8.S, objects.mak: Created.
+20
View File
@@ -0,0 +1,20 @@
libffi - Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001 Red Hat, Inc.
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 CYGNUS SOLUTIONS 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.
+72
View File
@@ -0,0 +1,72 @@
## Process this with automake to create Makefile.in
AUTOMAKE_OPTIONS = foreign no-installinfo
if TESTSUBDIR
SUBDIRS = include testsuite
else
SUBDIRS = include
endif
EXTRA_DIST = \
LICENSE \
README \
ChangeLog.v1 \
alpha/ffi.c \
arm/ffi.c \
ia64/ffi.c \
m68k/ffi.c \
mips/ffi.c \
powerpc/ffi.c \
s390/ffi.c \
sparc/ffi.c \
x86/ffi.c \
alpha/osf.S \
arm/sysv.S \
ia64/unix.S \
m68k/sysv.S \
mips/n32.S \
mips/o32.S \
powerpc/sysv.S \
s390/sysv.S \
sparc/v8.S \
sparc/v9.S \
x86/win32.S \
x86/sysv.S
# Multilib support variables.
MULTISRCTOP =
MULTIBUILDTOP =
MULTIDIRS =
MULTISUBDIR =
MULTIDO = true
MULTICLEAN = true
## Install a library built with a cross compiler in tooldir, not
## libdir.
if USE_LIBDIR
toolexeclibdir = $(libdir)$(MULTISUBDIR)
else
toolexecdir = $(exec_prefix)/$(target_alias)
toolexeclibdir = $(toolexecdir)/lib$(MULTISUBDIR)
endif
toolexeclib_LIBRARIES = libffi.a
## Work around automake deficiency
nodist_libffi_a_SOURCES = t-ffi.c t-asm.S
libffi_a_SOURCES = \
debug.c \
prep_cif.c \
types.c \
raw_api.c \
java_raw_api.c \
$(nodist_libffi_a_SOURCES)
AM_CFLAGS = -fexceptions
INCLUDES = \
-I$(top_srcdir)/include \
-Iinclude \
-I$(top_srcdir)/src
+513
View File
@@ -0,0 +1,513 @@
README for libffi-2.00
libffi-2.00 has not been released yet! This is a development snapshot!
libffi-1.20 was released on [SOME FUTURE DAY]. Check the libffi web
page for updates: <URL:http://sourceware.cygnus.com/libffi/>.
What is libffi?
===============
Compilers for high level languages generate code that follow certain
conventions. These conventions are necessary, in part, for separate
compilation to work. One such convention is the "calling
convention". The "calling convention" is essentially a set of
assumptions made by the compiler about where function arguments will
be found on entry to a function. A "calling convention" also specifies
where the return value for a function is found.
Some programs may not know at the time of compilation what arguments
are to be passed to a function. For instance, an interpreter may be
told at run-time about the number and types of arguments used to call
a given function. Libffi can be used in such programs to provide a
bridge from the interpreter program to compiled code.
The libffi library provides a portable, high level programming
interface to various calling conventions. This allows a programmer to
call any function specified by a call interface description at run
time.
Ffi stands for Foreign Function Interface. A foreign function
interface is the popular name for the interface that allows code
written in one language to call code written in another language. The
libffi library really only provides the lowest, machine dependent
layer of a fully featured foreign function interface. A layer must
exist above libffi that handles type conversions for values passed
between the two languages.
Supported Platforms and Prerequisites
=====================================
Libffi has been ported to:
Alpha - Linux and OSF/1
ARM - Linux (System V ABI)
IA-32 - Linux, Solaris, and BeOS (System V ABI)
IA-64 - Linux (System V ABI)
m68k - Linux (System V ABI)
MIPS, 32- and 64-bit - Irix (System V/o32 and n32)
PowerPC - Linux (System V ABI)
S390 - Linux (System V ABI)
Sparc, 32-bit - SunOS 4.1.3, Solaris 2.x, and Linux (Sparc v8)
Sparc, 64-bit - Linux (Sparc v9)
Libffi has been tested with GCC 3.00 and newer compilers.
You must use GNU make to build libffi.
If you port libffi to another platform, please let me know! I assume
that some will be easy (x86 NetBSD), and others will be more difficult
(HP, AIX).
Installing libffi
=================
[Note: before actually performing any of these installation steps,
you may wish to read the "Platform Specific Notes" below.]
First you must configure the distribution for your particular
system. Go to the directory you wish to build libffi in and run the
"configure" program found in the root directory of the libffi source
distribution.
You may want to tell configure where to install the libffi library and
header files. To do that, use the --prefix configure switch. Libffi
will install under /usr/local by default.
If you want to enable extra run-time debugging checks use the the
--enable-debug configure switch. This is useful when your program dies
mysteriously while using libffi.
Another useful configure switch is --enable-purify-safety. Using this
will add some extra code which will suppress certain warnings when you
are using Purify with libffi. Only use this switch when using
Purify, as it will slow down the library.
Configure has many other options. Use "configure --help" to see them all.
Once configure has finished, type "make". Note that you must be using
GNU make. SGI's make will not work. Sun's probably won't either.
You can ftp GNU make from prep.ai.mit.edu:/pub/gnu.
To ensure that libffi is working as advertised, type "make test".
To install the library and header files, type "make install".
Using libffi
============
The Basics
----------
Libffi assumes that you have a pointer to the function you wish to
call and that you know the number and types of arguments to pass it,
as well as the return type of the function.
The first thing you must do is create an ffi_cif object that matches
the signature of the function you wish to call. The cif in ffi_cif
stands for Call InterFace. To prepare a call interface object, use the
following function:
ffi_status ffi_prep_cif(ffi_cif *cif, ffi_abi abi,
unsigned int nargs,
ffi_type *rtype, ffi_type **atypes);
CIF is a pointer to the call interface object you wish
to initialize.
ABI is an enum that specifies the calling convention
to use for the call. FFI_DEFAULT_ABI defaults
to the system's native calling convention. Other
ABI's may be used with care. They are system
specific.
NARGS is the number of arguments this function accepts.
libffi does not yet support vararg functions.
RTYPE is a pointer to an ffi_type structure that represents
the return type of the function. Ffi_type objects
describe the types of values. libffi provides
ffi_type objects for many of the native C types:
signed int, unsigned int, signed char, unsigned char,
etc. There is also a pointer ffi_type object and
a void ffi_type. Use &ffi_type_void for functions that
don't return values.
ATYPES is a vector of ffi_type pointers. ARGS must be NARGS long.
If NARGS is 0, this is ignored.
ffi_prep_cif will return a status code that you are responsible
for checking. It will be one of the following:
FFI_OK - All is good.
FFI_BAD_TYPEDEF - One of the ffi_type objects that ffi_prep_cif
came across is bad.
Before making the call, the VALUES vector should be initialized
with pointers to the appropriate argument values.
To call the the function using the initialized ffi_cif, use the
ffi_call function:
void ffi_call(ffi_cif *cif, void *fn, void *rvalue, void **avalues);
CIF is a pointer to the ffi_cif initialized specifically
for this function.
FN is a pointer to the function you want to call.
RVALUE is a pointer to a chunk of memory that is to hold the
result of the function call. Currently, it must be
at least one word in size (except for the n32 version
under Irix 6.x, which must be a pointer to an 8 byte
aligned value (a long long). It must also be at least
word aligned (depending on the return type, and the
system's alignment requirements). If RTYPE is
&ffi_type_void, this is ignored. If RVALUE is NULL,
the return value is discarded.
AVALUES is a vector of void* that point to the memory locations
holding the argument values for a call.
If NARGS is 0, this is ignored.
If you are expecting a return value from FN it will have been stored
at RVALUE.
An Example
----------
Here is a trivial example that calls puts() a few times.
#include <stdio.h>
#include <ffi.h>
int main()
{
ffi_cif cif;
ffi_type *args[1];
void *values[1];
char *s;
int rc;
/* Initialize the argument info vectors */
args[0] = &ffi_type_uint;
values[0] = &s;
/* Initialize the cif */
if (ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_uint, args) == FFI_OK)
{
s = "Hello World!";
ffi_call(&cif, puts, &rc, values);
/* rc now holds the result of the call to puts */
/* values holds a pointer to the function's arg, so to
call puts() again all we need to do is change the
value of s */
s = "This is cool!";
ffi_call(&cif, puts, &rc, values);
}
return 0;
}
Aggregate Types
---------------
Although libffi has no special support for unions or bit-fields, it is
perfectly happy passing structures back and forth. You must first
describe the structure to libffi by creating a new ffi_type object
for it. Here is the definition of ffi_type:
typedef struct _ffi_type
{
unsigned size;
short alignment;
short type;
struct _ffi_type **elements;
} ffi_type;
All structures must have type set to FFI_TYPE_STRUCT. You may set
size and alignment to 0. These will be calculated and reset to the
appropriate values by ffi_prep_cif().
elements is a NULL terminated array of pointers to ffi_type objects
that describe the type of the structure elements. These may, in turn,
be structure elements.
The following example initializes a ffi_type object representing the
tm struct from Linux's time.h:
struct tm {
int tm_sec;
int tm_min;
int tm_hour;
int tm_mday;
int tm_mon;
int tm_year;
int tm_wday;
int tm_yday;
int tm_isdst;
/* Those are for future use. */
long int __tm_gmtoff__;
__const char *__tm_zone__;
};
{
ffi_type tm_type;
ffi_type *tm_type_elements[12];
int i;
tm_type.size = tm_type.alignment = 0;
tm_type.elements = &tm_type_elements;
for (i = 0; i < 9; i++)
tm_type_elements[i] = &ffi_type_sint;
tm_type_elements[9] = &ffi_type_slong;
tm_type_elements[10] = &ffi_type_pointer;
tm_type_elements[11] = NULL;
/* tm_type can now be used to represent tm argument types and
return types for ffi_prep_cif() */
}
Platform Specific Notes
=======================
Intel x86
---------
There are no known problems with the x86 port.
Sun Sparc - SunOS 4.1.3 & Solaris 2.x
-------------------------------------
There's a bug in the structure passing code for sparc processors.
Struct arguments that are passed in value actually end up being passed
by reference. This will be fixed Real Soon Now.
"long long" values are not supported yet.
You must use GNU Make to build libffi on Sun platforms.
MIPS - Irix 5.3 & 6.x
---------------------
Irix 6.2 and better supports three different calling conventions: o32,
n32 and n64. Currently, libffi only supports both o32 and n32 under
Irix 6.x, but only o32 under Irix 5.3. Libffi will automatically be
configured for whichever calling convention it was built for.
By default, the configure script will try to build libffi with the GNU
development tools. To build libffi with the SGI development tools, set
the environment variable CC to either "cc -32" or "cc -n32" before
running configure under Irix 6.x (depending on whether you want an o32
or n32 library), or just "cc" for Irix 5.3.
With the n32 calling convention, when returning structures smaller
than 16 bytes, be sure to provide an RVALUE that is 8 byte aligned.
Here's one way of forcing this:
double struct_storage[2];
my_small_struct *s = (my_small_struct *) struct_storage;
/* Use s for RVALUE */
If you don't do this you are liable to get spurious bus errors.
"long long" values are not supported yet.
You must use GNU Make to build libffi on SGI platforms.
ARM - System V ABI
------------------
The ARM port was performed on a NetWinder running ARM Linux ELF
(2.0.31) and gcc 2.8.1. config.sub still does not recognize the
machine name sa110-unknown-linux-gnu (currently returned by
NetWinder). In the mean time the package can be configured by running
'configure arm-linux'.
PowerPC System V ABI
--------------------
There are two `System V ABI's which libffi implements for PowerPC.
They differ only in how small structures are returned from functions.
In the FFI_SYSV version, structures that are 8 bytes or smaller are
returned in registers. This is what GCC does when it is configured
for solaris, and is what the System V ABI I have (dated September
1995) says.
In the FFI_GCC_SYSV version, all structures are returned the same way:
by passing a pointer as the first argument to the function. This is
what GCC does when it is configured for linux or a generic sysv
target.
EGCS 1.0.1 (and probably other versions of EGCS/GCC) also has a
inconsistency with the SysV ABI: When a procedure is called with many
floating-point arguments, some of them get put on the stack. They are
all supposed to be stored in double-precision format, even if they are
only single-precision, but EGCS stores single-precision arguments as
single-precision anyway. This causes one test to fail (the `many
arguments' test).
History
=======
2.00 ???-??-01
Many configure improvements. No more run tests required at
configure time so cross compiling should work well. ffi.h is
target independent now and may be installed safely anywhere.
ffitest replaced with dejagnu testing infrastructure. s390
port from ????. ia64 port from Hans Boehm. General
performance improvements. Closures designed and implemented
by Kresten Krab Thorup and ported to ia64 and powerpc by Hans
Boehm and ???? respectively.
1.20 Oct-5-98
Raffaele Sena produces ARM port.
1.19 Oct-5-98
Fixed x86 long double and long long return support.
m68k bug fixes from Andreas Schwab.
Patch for DU assembler compatibility for the Alpha from Richard
Henderson.
1.18 Apr-17-98
Bug fixes and MIPS configuration changes.
1.17 Feb-24-98
Bug fixes and m68k port from Andreas Schwab. PowerPC port from
Geoffrey Keating. Various bug x86, Sparc and MIPS bug fixes.
1.16 Feb-11-98
Richard Henderson produces Alpha port.
1.15 Dec-4-97
Fixed an n32 ABI bug. New libtool, auto* support.
1.14 May-13-97
libtool is now used to generate shared and static libraries.
Fixed a minor portability problem reported by Russ McManus
<[email protected]>.
1.13 Dec-2-96
Added --enable-purify-safety to keep Purify from complaining
about certain low level code.
Sparc fix for calling functions with < 6 args.
Linux x86 a.out fix.
1.12 Nov-22-96
Added missing ffi_type_void, needed for supporting void return
types. Fixed test case for non MIPS machines. Cygnus Support
is now Cygnus Solutions.
1.11 Oct-30-96
Added notes about GNU make.
1.10 Oct-29-96
Added configuration fix for non GNU compilers.
1.09 Oct-29-96
Added --enable-debug configure switch. Clean-ups based on LCLint
feedback. ffi_mips.h is always installed. Many configuration
fixes. Fixed ffitest.c for sparc builds.
1.08 Oct-15-96
Fixed n32 problem. Many clean-ups.
1.07 Oct-14-96
Gordon Irlam rewrites v8.S again. Bug fixes.
1.06 Oct-14-96
Gordon Irlam improved the sparc port.
1.05 Oct-14-96
Interface changes based on feedback.
1.04 Oct-11-96
Sparc port complete (modulo struct passing bug).
1.03 Oct-10-96
Passing struct args, and returning struct values works for
all architectures/calling conventions. Expanded tests.
1.02 Oct-9-96
Added SGI n32 support. Fixed bugs in both o32 and Linux support.
Added "make test".
1.01 Oct-8-96
Fixed float passing bug in mips version. Restructured some
of the code. Builds cleanly with SGI tools.
1.00 Oct-7-96
First release. No public announcement.
Authors & Credits
=================
libffi was written by Anthony Green <[email protected]>.
Portions of libffi were derived from Gianni Mariani's free gencall
library for Silicon Graphics machines.
The closure mechanism was designed and implemented on IA-32 by Kresten
Krab Thorup.
The initial Sparc port was derived from code contributed by the fine
folks at Visible Decisions Inc. Further enhancements were made by
Gordon Irlam while at Cygnus Solutions.
The Alpha port was written by Richard Henderson at Cygnus Solutions/Red Hat.
Andreas Schwab ported libffi to m68k Linux and provided a number of
bug fixes.
Geoffrey Keating ported libffi to the PowerPC.
Raffaele Sena ported libffi to the ARM.
Hans Boehm ported libffi to IA-64.
Jakub Jelinek ported libffi to 32- and 64-bit Sparc Linux.
Jesper Skov and Andrew Haley both did more than their fair share of
stepping through the code and tracking down bugs.
Thanks also to Tom Tromey for bug fixes and configuration help.
Thanks to Jim Blandy, who provided some useful feedback on the libffi
interface.
If you have a problem, or have found a bug, please send a note to
[email protected].
+12
View File
@@ -0,0 +1,12 @@
/* Define this if you want extra debugging */
#undef FFI_DEBUG
/* Define this if you are using Purify and want to suppress
spurious messages. */
#undef USING_PURIFY
/* Define this is you do not want support for aggregate types. */
#undef FFI_NO_STRUCTS
/* Define this is you do not want support for the raw API. */
#undef FFI_NO_RAW_API
+97
View File
@@ -0,0 +1,97 @@
AC_DEFUN([AC_COMPILE_CHECK_SIZEOF],
[changequote(<<, >>)dnl
dnl The name to #define.
define(<<AC_TYPE_NAME>>, translit(sizeof_$1, [a-z *], [A-Z_P]))dnl
dnl The cache variable name.
define(<<AC_CV_NAME>>, translit(ac_cv_sizeof_$1, [ *], [_p]))dnl
changequote([, ])dnl
AC_MSG_CHECKING(size of $1)
AC_CACHE_VAL(AC_CV_NAME,
[for ac_size in 4 8 1 2 16 12 $2 ; do # List sizes in rough order of prevalence.
AC_TRY_COMPILE([#include "confdefs.h"
#include <sys/types.h>
$2
], [switch (0) case 0: case (sizeof ($1) == $ac_size):;], AC_CV_NAME=$ac_size)
if test x$AC_CV_NAME != x ; then break; fi
done
])
if test x$AC_CV_NAME = x ; then
AC_MSG_ERROR([cannot determine a size for $1])
fi
AC_MSG_RESULT($AC_CV_NAME)
AC_DEFINE_UNQUOTED(AC_TYPE_NAME, $AC_CV_NAME, [The number of bytes in type $1])
undefine([AC_TYPE_NAME])dnl
undefine([AC_CV_NAME])dnl
])
AC_DEFUN([AC_C_BIGENDIAN_CROSS],
[AC_CACHE_CHECK(whether byte ordering is bigendian, ac_cv_c_bigendian,
[ac_cv_c_bigendian=unknown
# See if sys/param.h defines the BYTE_ORDER macro.
AC_TRY_COMPILE([#include <sys/types.h>
#include <sys/param.h>], [
#if !BYTE_ORDER || !BIG_ENDIAN || !LITTLE_ENDIAN
bogus endian macros
#endif], [# It does; now see whether it defined to BIG_ENDIAN or not.
AC_TRY_COMPILE([#include <sys/types.h>
#include <sys/param.h>], [
#if BYTE_ORDER != BIG_ENDIAN
not big endian
#endif], ac_cv_c_bigendian=yes, ac_cv_c_bigendian=no)])
if test $ac_cv_c_bigendian = unknown; then
AC_TRY_RUN([main () {
/* Are we little or big endian? From Harbison&Steele. */
union
{
long l;
char c[sizeof (long)];
} u;
u.l = 1;
exit (u.c[sizeof (long) - 1] == 1);
}], ac_cv_c_bigendian=no, ac_cv_c_bigendian=yes,
[ echo $ac_n "cross-compiling... " 2>&AC_FD_MSG ])
fi])
if test $ac_cv_c_bigendian = unknown; then
AC_MSG_CHECKING(to probe for byte ordering)
[
cat >conftest.c <<EOF
short ascii_mm[] = { 0x4249, 0x4765, 0x6E44, 0x6961, 0x6E53, 0x7953, 0 };
short ascii_ii[] = { 0x694C, 0x5454, 0x656C, 0x6E45, 0x6944, 0x6E61, 0 };
void _ascii() { char* s = (char*) ascii_mm; s = (char*) ascii_ii; }
short ebcdic_ii[] = { 0x89D3, 0xE3E3, 0x8593, 0x95C5, 0x89C4, 0x9581, 0 };
short ebcdic_mm[] = { 0xC2C9, 0xC785, 0x95C4, 0x8981, 0x95E2, 0xA8E2, 0 };
void _ebcdic() { char* s = (char*) ebcdic_mm; s = (char*) ebcdic_ii; }
int main() { _ascii (); _ebcdic (); return 0; }
EOF
] if test -f conftest.c ; then
if ${CC-cc} ${CFLAGS} conftest.c -o conftest.o && test -f conftest.o ; then
if test `grep -l BIGenDianSyS conftest.o` ; then
echo $ac_n ' big endian probe OK, ' 1>&AC_FD_MSG
ac_cv_c_bigendian=yes
fi
if test `grep -l LiTTleEnDian conftest.o` ; then
echo $ac_n ' little endian probe OK, ' 1>&AC_FD_MSG
if test $ac_cv_c_bigendian = yes ; then
ac_cv_c_bigendian=unknown;
else
ac_cv_c_bigendian=no
fi
fi
echo $ac_n 'guessing bigendian ... ' >&AC_FD_MSG
fi
fi
AC_MSG_RESULT($ac_cv_c_bigendian)
fi
if test $ac_cv_c_bigendian = yes; then
AC_DEFINE(WORDS_BIGENDIAN, 1, [whether byteorder is bigendian])
BYTEORDER=4321
else
BYTEORDER=1234
fi
AC_DEFINE_UNQUOTED(BYTEORDER, $BYTEORDER, [1234 = LIL_ENDIAN, 4321 = BIGENDIAN])
if test $ac_cv_c_bigendian = unknown; then
AC_MSG_ERROR(unknown endianess - sorry, please pre-set ac_cv_c_bigendian)
fi
])
+247
View File
@@ -0,0 +1,247 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1998 Cygnus Solutions
Alpha Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
extern void ffi_call_osf(void *, unsigned long, unsigned, void *, void (*)());
extern void ffi_closure_osf(void);
ffi_status
ffi_prep_cif_machdep(ffi_cif *cif)
{
/* Adjust cif->bytes to represent a minimum 6 words for the temporary
register argument loading area. */
if (cif->bytes < 6*SIZEOF_ARG)
cif->bytes = 6*SIZEOF_ARG;
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_STRUCT:
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags = cif->rtype->type;
break;
default:
cif->flags = FFI_TYPE_INT;
break;
}
return FFI_OK;
}
void
ffi_call(ffi_cif *cif, void (*fn)(), void *rvalue, void **avalue)
{
unsigned long *stack, *argp;
long i, avn;
ffi_type **arg_types;
FFI_ASSERT (cif->abi == FFI_OSF);
/* If the return value is a struct and we don't have a return
value address then we need to make one. */
if (rvalue == NULL && cif->flags == FFI_TYPE_STRUCT)
rvalue = alloca(cif->rtype->size);
/* Allocate the space for the arguments, plus 4 words of temp
space for ffi_call_osf. */
argp = stack = alloca(cif->bytes + 4*SIZEOF_ARG);
if (cif->flags == FFI_TYPE_STRUCT)
*(void **) argp++ = rvalue;
i = 0;
avn = cif->nargs;
arg_types = cif->arg_types;
while (i < avn)
{
switch ((*arg_types)->type)
{
case FFI_TYPE_SINT8:
*(SINT64 *) argp = *(SINT8 *)(* avalue);
break;
case FFI_TYPE_UINT8:
*(SINT64 *) argp = *(UINT8 *)(* avalue);
break;
case FFI_TYPE_SINT16:
*(SINT64 *) argp = *(SINT16 *)(* avalue);
break;
case FFI_TYPE_UINT16:
*(SINT64 *) argp = *(UINT16 *)(* avalue);
break;
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
/* Note that unsigned 32-bit quantities are sign extended. */
*(SINT64 *) argp = *(SINT32 *)(* avalue);
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
*(UINT64 *) argp = *(UINT64 *)(* avalue);
break;
case FFI_TYPE_FLOAT:
if (argp - stack < 6)
{
/* Note the conversion -- all the fp regs are loaded as
doubles. The in-register format is the same. */
*(double *) argp = *(float *)(* avalue);
}
else
*(float *) argp = *(float *)(* avalue);
break;
case FFI_TYPE_DOUBLE:
*(double *) argp = *(double *)(* avalue);
break;
case FFI_TYPE_STRUCT:
memcpy(argp, *avalue, (*arg_types)->size);
break;
default:
FFI_ASSERT(0);
}
argp += ALIGN((*arg_types)->size, SIZEOF_ARG) / SIZEOF_ARG;
i++, arg_types++, avalue++;
}
ffi_call_osf(stack, cif->bytes, cif->flags, rvalue, fn);
}
ffi_status
ffi_prep_closure (ffi_closure* closure,
ffi_cif* cif,
void (*fun)(ffi_cif*, void*, void**, void*),
void *user_data)
{
unsigned int *tramp;
FFI_ASSERT (cif->abi == FFI_OSF);
tramp = (unsigned int *) &closure->tramp[0];
tramp[0] = 0x47fb0401; /* mov $27,$1 */
tramp[1] = 0xa77b0010; /* ldq $27,16($27) */
tramp[2] = 0x6bfb0000; /* jmp $31,($27),0 */
tramp[3] = 0x47ff041f; /* nop */
*(void **) &tramp[4] = ffi_closure_osf;
closure->cif = cif;
closure->fun = fun;
closure->user_data = user_data;
/* Flush the Icache. */
asm volatile ("imb" : : : "memory");
return FFI_OK;
}
int
ffi_closure_osf_inner(ffi_closure *closure, void *rvalue, unsigned long *argp)
{
ffi_cif *cif;
void **avalue;
ffi_type **arg_types;
long i, avn, argn;
cif = closure->cif;
avalue = alloca(cif->nargs * sizeof(void *));
argn = 0;
/* Copy the caller's structure return address to that the closure
returns the data directly to the caller. */
if (cif->flags == FFI_TYPE_STRUCT)
{
rvalue = (void *) argp[0];
argn = 1;
}
i = 0;
avn = cif->nargs;
arg_types = cif->arg_types;
/* Grab the addresses of the arguments from the stack frame. */
while (i < avn)
{
switch (arg_types[i]->type)
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
case FFI_TYPE_STRUCT:
avalue[i] = &argp[argn];
break;
case FFI_TYPE_FLOAT:
if (argn < 6)
{
/* Floats coming from registers need conversion from double
back to float format. */
*(float *)&argp[argn - 6] = *(double *)&argp[argn - 6];
avalue[i] = &argp[argn - 6];
}
else
avalue[i] = &argp[argn];
break;
case FFI_TYPE_DOUBLE:
avalue[i] = &argp[argn - (argn < 6 ? 6 : 0)];
break;
default:
FFI_ASSERT(0);
}
argn += ALIGN(arg_types[i]->size, SIZEOF_ARG) / SIZEOF_ARG;
i++;
}
/* Invoke the closure. */
(closure->fun) (cif, rvalue, avalue, closure->user_data);
/* Tell ffi_closure_osf how to perform return type promotions. */
return cif->rtype->type;
}
+278
View File
@@ -0,0 +1,278 @@
/* -----------------------------------------------------------------------
osf.S - Copyright (c) 1998 Cygnus Solutions
Alpha/OSF Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
.arch ev6
.text
/* ffi_call_osf (void *args, unsigned long bytes, unsigned flags,
void *raddr, void (*fnaddr)());
Bit o trickiness here -- ARGS+BYTES is the base of the stack frame
for this function. This has been allocated by ffi_call. We also
deallocate some of the stack that has been alloca'd. */
.align 3
.globl ffi_call_osf
.ent ffi_call_osf
ffi_call_osf:
.frame $15, 32, $26, 0
.mask 0x4008000, -32
addq $16,$17,$1
mov $16, $30
stq $26, 0($1)
stq $15, 8($1)
stq $18, 16($1)
mov $1, $15
.prologue 0
stq $19, 24($1)
mov $20, $27
# Load up all of the (potential) argument registers.
ldq $16, 0($30)
ldt $f16, 0($30)
ldt $f17, 8($30)
ldq $17, 8($30)
ldt $f18, 16($30)
ldq $18, 16($30)
ldt $f19, 24($30)
ldq $19, 24($30)
ldt $f20, 32($30)
ldq $20, 32($30)
ldt $f21, 40($30)
ldq $21, 40($30)
# Deallocate the register argument area.
lda $30, 48($30)
jsr $26, ($27), 0
ldgp $29, 0($26)
# If the return value pointer is NULL, assume no return value.
ldq $19, 24($15)
ldq $18, 16($15)
ldq $26, 0($15)
beq $19, $noretval
# Store the return value out in the proper type.
cmpeq $18, FFI_TYPE_INT, $1
bne $1, $retint
cmpeq $18, FFI_TYPE_FLOAT, $2
bne $2, $retfloat
cmpeq $18, FFI_TYPE_DOUBLE, $3
bne $3, $retdouble
$noretval:
ldq $15, 8($15)
ret
$retint:
stq $0, 0($19)
nop
ldq $15, 8($15)
ret
$retfloat:
sts $f0, 0($19)
nop
ldq $15, 8($15)
ret
$retdouble:
stt $f0, 0($19)
nop
ldq $15, 8($15)
ret
.end ffi_call_osf
/* ffi_closure_osf(...)
Receives the closure argument in $1. */
.align 3
.globl ffi_closure_osf
.ent ffi_closure_osf
ffi_closure_osf:
.frame $30, 16*8, $26, 0
.mask 0x4000000, -16*8
ldgp $29, 0($27)
subq $30, 16*8, $30
stq $26, 0($30)
.prologue 1
# Store all of the potential argument registers in va_list format.
stt $f16, 4*8($30)
stt $f17, 5*8($30)
stt $f18, 6*8($30)
stt $f19, 7*8($30)
stt $f20, 8*8($30)
stt $f21, 9*8($30)
stq $16, 10*8($30)
stq $17, 11*8($30)
stq $18, 12*8($30)
stq $19, 13*8($30)
stq $20, 14*8($30)
stq $21, 15*8($30)
# Call ffi_closure_osf_inner to do the bulk of the work.
mov $1, $16
lda $17, 2*8($30)
lda $18, 10*8($30)
jsr $26, ffi_closure_osf_inner
ldgp $29, 0($26)
ldq $26, 0($30)
# Load up the return value in the proper type.
lda $1, $load_table
s4addq $0, $1, $1
ldl $1, 0($1)
addq $1, $29, $1
jmp $31, ($1), $load_32
.align 4
$load_none:
addq $30, 16*8, $30
ret
.align 4
$load_float:
lds $f0, 16($30)
nop
addq $30, 16*8, $30
ret
.align 4
$load_double:
ldt $f0, 16($30)
nop
addq $30, 16*8, $30
ret
.align 4
$load_u8:
#ifdef __alpha_bwx__
ldbu $0, 16($30)
nop
#else
ldq $0, 16($30)
and $0, 255, $0
#endif
addq $30, 16*8, $30
ret
.align 4
$load_s8:
#ifdef __alpha_bwx__
ldbu $0, 16($30)
sextb $0, $0
#else
ldq $0, 16($30)
sll $0, 56, $0
sra $0, 56, $0
#endif
addq $30, 16*8, $30
ret
.align 4
$load_u16:
#ifdef __alpha_bwx__
ldwu $0, 16($30)
nop
#else
ldq $0, 16($30)
zapnot $0, 3, $0
#endif
addq $30, 16*8, $30
ret
.align 4
$load_s16:
#ifdef __alpha_bwx__
ldwu $0, 16($30)
sextw $0, $0
#else
ldq $0, 16($30)
sll $0, 48, $0
sra $0, 48, $0
#endif
addq $30, 16*8, $30
ret
.align 4
$load_32:
ldl $0, 16($30)
nop
addq $30, 16*8, $30
ret
.align 4
$load_64:
ldq $0, 16($30)
nop
addq $30, 16*8, $30
ret
.end ffi_closure_osf
.section .rodata
$load_table:
.gprel32 $load_none # FFI_TYPE_VOID
.gprel32 $load_32 # FFI_TYPE_INT
.gprel32 $load_float # FFI_TYPE_FLOAT
.gprel32 $load_double # FFI_TYPE_DOUBLE
.gprel32 $load_double # FFI_TYPE_LONGDOUBLE
.gprel32 $load_u8 # FFI_TYPE_UINT8
.gprel32 $load_s8 # FFI_TYPE_SINT8
.gprel32 $load_u16 # FFI_TYPE_UINT16
.gprel32 $load_s16 # FFI_TYPE_SINT16
.gprel32 $load_32 # FFI_TYPE_UINT32
.gprel32 $load_32 # FFI_TYPE_SINT32
.gprel32 $load_64 # FFI_TYPE_UINT64
.gprel32 $load_64 # FFI_TYPE_SINT64
.gprel32 $load_none # FFI_TYPE_STRUCT
.gprel32 $load_64 # FFI_TYPE_POINTER
/* Assert that the table above is in sync with ffi.h. */
#if FFI_TYPE_FLOAT != 2 \
|| FFI_TYPE_DOUBLE != 3 \
|| FFI_TYPE_UINT8 != 5 \
|| FFI_TYPE_SINT8 != 6 \
|| FFI_TYPE_UINT16 != 7 \
|| FFI_TYPE_SINT16 != 8 \
|| FFI_TYPE_UINT32 != 9 \
|| FFI_TYPE_SINT32 != 10 \
|| FFI_TYPE_UINT64 != 11 \
|| FFI_TYPE_SINT64 != 12 \
|| FFI_TYPE_STRUCT != 13 \
|| FFI_TYPE_POINTER != 14 \
|| FFI_TYPE_LAST != 14
#error "osf.S out of sync with ffi.h"
#endif
+183
View File
@@ -0,0 +1,183 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1998 Cygnus Solutions
ARM Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments */
/*@-exportheader@*/
void ffi_prep_args(char *stack, extended_cif *ecif)
/*@=exportheader@*/
{
register unsigned int i;
register int tmp;
register unsigned int avn;
register void **p_argv;
register char *argp;
register ffi_type **p_arg;
tmp = 0;
argp = stack;
if ( ecif->cif->rtype->type == FFI_TYPE_STRUCT ) {
*(void **) argp = ecif->rvalue;
argp += 4;
}
avn = ecif->cif->nargs;
p_argv = ecif->avalue;
for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types;
(i != 0) && (avn != 0);
i--, p_arg++)
{
size_t z;
/* Align if necessary */
if (((*p_arg)->alignment - 1) & (unsigned) argp) {
argp = (char *) ALIGN(argp, (*p_arg)->alignment);
}
if (avn != 0)
{
avn--;
z = (*p_arg)->size;
if (z < sizeof(int))
{
z = sizeof(int);
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(signed int *) argp = (signed int)*(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
*(unsigned int *) argp = (unsigned int)*(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
*(signed int *) argp = (signed int)*(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
*(unsigned int *) argp = (unsigned int)*(UINT16 *)(* p_argv);
break;
case FFI_TYPE_STRUCT:
*(unsigned int *) argp = (unsigned int)*(UINT32 *)(* p_argv);
break;
default:
FFI_ASSERT(0);
}
}
else if (z == sizeof(int))
{
*(unsigned int *) argp = (unsigned int)*(UINT32 *)(* p_argv);
}
else
{
memcpy(argp, *p_argv, z);
}
p_argv++;
argp += z;
}
}
return;
}
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
{
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
case FFI_TYPE_STRUCT:
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags = (unsigned) cif->rtype->type;
break;
default:
cif->flags = FFI_TYPE_INT;
break;
}
return FFI_OK;
}
/*@-declundef@*/
/*@-exportheader@*/
extern void ffi_call_SYSV(void (*)(char *, extended_cif *),
/*@out@*/ extended_cif *,
unsigned, unsigned,
/*@out@*/ unsigned *,
void (*fn)());
/*@=declundef@*/
/*@=exportheader@*/
void ffi_call(/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
{
/*@-sysunrecog@*/
ecif.rvalue = alloca(cif->rtype->size);
/*@=sysunrecog@*/
}
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
/*@-usedef@*/
ffi_call_SYSV(ffi_prep_args, &ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
/*@=usedef@*/
break;
default:
FFI_ASSERT(0);
break;
}
}
+111
View File
@@ -0,0 +1,111 @@
/* -----------------------------------------------------------------------
sysv.S - Copyright (c) 1998 Cygnus Solutions
ARM Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#ifdef HAVE_MACHINE_ASM_H
#include <machine/asm.h>
#else
/* XXX these lose for some platforms, I'm sure. */
#define CNAME(x) x
#define ENTRY(x) .globl CNAME(x); .type CNAME(x),%function; CNAME(x):
#endif
.text
# a1: ffi_prep_args
# a2: &ecif
# a3: cif->bytes
# a4: fig->flags
# sp+0: ecif.rvalue
# sp+4: fn
# This assumes we are using gas.
ENTRY(ffi_call_SYSV)
# Save registers
stmfd sp!, {a1-a4, fp, lr}
mov fp, sp
# Make room for all of the new args.
sub sp, fp, a3
# Place all of the ffi_prep_args in position
mov ip, a1
mov a1, sp
# a2 already set
# And call
mov lr, pc
mov pc, ip
# move first 4 parameters in registers
ldr a1, [sp, #0]
ldr a2, [sp, #4]
ldr a3, [sp, #8]
ldr a4, [sp, #12]
# and adjust stack
ldr ip, [fp, #8]
cmp ip, #16
movge ip, #16
add sp, sp, ip
# call function
mov lr, pc
ldr pc, [fp, #28]
# Remove the space we pushed for the args
mov sp, fp
# Load a3 with the pointer to storage for the return value
ldr a3, [sp, #24]
# Load a4 with the return type code
ldr a4, [sp, #12]
# If the return value pointer is NULL, assume no return value.
cmp a3, #0
beq epilogue
# return INT
cmp a4, #FFI_TYPE_INT
streq a1, [a3]
beq epilogue
# return FLOAT
cmp a4, #FFI_TYPE_FLOAT
stfeqs f0, [a3]
beq epilogue
# return DOUBLE or LONGDOUBLE
cmp a4, #FFI_TYPE_DOUBLE
stfeqd f0, [a3]
epilogue:
ldmfd sp!, {a1-a4, fp, pc}
.ffi_call_SYSV_end:
.size CNAME(ffi_call_SYSV),.ffi_call_SYSV_end-CNAME(ffi_call_SYSV)
+6
View File
@@ -0,0 +1,6 @@
libtoolize --automake
automake -a
autoheader
aclocal $ACLOCAL_FLAGS
autoconf
./configure $*
+849
View File
@@ -0,0 +1,849 @@
# Configure fragment invoked in the post-target section for subdirs
# wanting multilib support.
#
# It is advisable to support a few --enable/--disable options to let the
# user select which libraries s/he really wants.
#
# Subdirectories wishing to use multilib should put the following lines
# in the "post-target" section of configure.in.
#
# if [ "${srcdir}" = "." ] ; then
# if [ "${with_target_subdir}" != "." ] ; then
# . ${with_multisrctop}../../config-ml.in
# else
# . ${with_multisrctop}../config-ml.in
# fi
# else
# . ${srcdir}/../config-ml.in
# fi
#
#
# Things are complicated because 6 separate cases must be handled:
# 2 (native, cross) x 3 (absolute-path, relative-not-dot, dot) = 6.
#
# srcdir=. is special. It must handle make programs that don't handle VPATH.
# To implement this, a symlink tree is built for each library and for each
# multilib subdir.
#
# The build tree is layed out as
#
# ./
# newlib
# m68020/
# newlib
# m68881/
# newlib
#
# The nice feature about this arrangement is that inter-library references
# in the build tree work without having to care where you are. Note that
# inter-library references also work in the source tree because symlink trees
# are built when srcdir=.
#
# Unfortunately, trying to access the libraries in the build tree requires
# the user to manually choose which library to use as GCC won't be able to
# find the right one. This is viewed as the lesser of two evils.
#
# Configure variables:
# ${with_target_subdir} = "." for native, or ${target_alias} for cross.
# Set by top level Makefile.
# ${with_multisrctop} = how many levels of multilibs there are in the source
# tree. It exists to handle the case of configuring in the source tree:
# ${srcdir} is not constant.
# ${with_multisubdir} = name of multilib subdirectory (eg: m68020/m68881).
#
# Makefile variables:
# MULTISRCTOP = number of multilib levels in source tree (+1 if cross)
# (FIXME: note that this is different than ${with_multisrctop}. Check out.).
# MULTIBUILDTOP = number of multilib levels in build tree
# MULTIDIRS = list of multilib subdirs (eg: m68000 m68020 ...)
# (only defined in each library's main Makefile).
# MULTISUBDIR = installed subdirectory name with leading '/' (eg: /m68000)
# (only defined in each multilib subdir).
# FIXME: Multilib is currently disabled by default for everything other than
# newlib. It is up to each target to turn on multilib support for the other
# libraries as desired.
# We have to handle being invoked by both Cygnus configure and Autoconf.
#
# Cygnus configure incoming variables:
# srcdir, subdir, host, arguments
#
# Autoconf incoming variables:
# srcdir, host, ac_configure_args
#
# We *could* figure srcdir and host out, but we'd have to do work that
# our caller has already done to figure them out and requiring these two
# seems reasonable.
# Note that `host' in this case is GCC's `target'. Target libraries are
# configured for a particular host.
if [ -n "${ac_configure_args}" ]; then
Makefile=${ac_file-Makefile}
ml_config_shell=${CONFIG_SHELL-/bin/sh}
ml_arguments="${ac_configure_args}"
ml_realsrcdir=${srcdir}
else
Makefile=${Makefile-Makefile}
ml_config_shell=${config_shell-/bin/sh}
ml_arguments="${arguments}"
if [ -n "${subdir}" -a "${subdir}" != "." ] ; then
ml_realsrcdir=${srcdir}/${subdir}
else
ml_realsrcdir=${srcdir}
fi
fi
# Scan all the arguments and set all the ones we need.
ml_verbose=--verbose
for option in ${ml_arguments}
do
case $option in
--*) ;;
-*) option=-$option ;;
esac
case $option in
--*=*)
optarg=`echo $option | sed -e 's/^[^=]*=//'`
;;
esac
case $option in
--disable-*)
enableopt=`echo ${option} | sed 's:^--disable-:enable_:;s:-:_:g'`
eval $enableopt=no
;;
--enable-*)
case "$option" in
*=*) ;;
*) optarg=yes ;;
esac
enableopt=`echo ${option} | sed 's:^--::;s:=.*$::;s:-:_:g'`
eval $enableopt="$optarg"
;;
--norecursion | --no*)
ml_norecursion=yes
;;
--silent | --sil* | --quiet | --q*)
ml_verbose=--silent
;;
--verbose | --v | --verb*)
ml_verbose=--verbose
;;
--with-*)
case "$option" in
*=*) ;;
*) optarg=yes ;;
esac
withopt=`echo ${option} | sed 's:^--::;s:=.*$::;s:-:_:g'`
eval $withopt="$optarg"
;;
--without-*)
withopt=`echo ${option} | sed 's:^--::;s:out::;s:-:_:g'`
eval $withopt=no
;;
esac
done
# Only do this if --enable-multilib.
if [ "${enable_multilib}" = yes ]; then
# Compute whether this is the library's top level directory
# (ie: not a multilib subdirectory, and not a subdirectory like newlib/src).
# ${with_multisubdir} tells us we're in the right branch, but we could be
# in a subdir of that.
# ??? The previous version could void this test by separating the process into
# two files: one that only the library's toplevel configure.in ran (to
# configure the multilib subdirs), and another that all configure.in's ran to
# update the Makefile. It seemed reasonable to collapse all multilib support
# into one file, but it does leave us with having to perform this test.
ml_toplevel_p=no
if [ -z "${with_multisubdir}" ]; then
if [ "${srcdir}" = "." ]; then
# Use ${ml_realsrcdir} instead of ${srcdir} here to account for ${subdir}.
# ${with_target_subdir} = "." for native, otherwise target alias.
if [ "${with_target_subdir}" = "." ]; then
if [ -f ${ml_realsrcdir}/../config-ml.in ]; then
ml_toplevel_p=yes
fi
else
if [ -f ${ml_realsrcdir}/../../config-ml.in ]; then
ml_toplevel_p=yes
fi
fi
else
# Use ${ml_realsrcdir} instead of ${srcdir} here to account for ${subdir}.
if [ -f ${ml_realsrcdir}/../config-ml.in ]; then
ml_toplevel_p=yes
fi
fi
fi
# If this is the library's top level directory, set multidirs to the
# multilib subdirs to support. This lives at the top because we need
# `multidirs' set right away.
if [ "${ml_toplevel_p}" = yes ]; then
multidirs=
for i in `${CC-gcc} --print-multi-lib 2>/dev/null`; do
dir=`echo $i | sed -e 's/;.*$//'`
if [ "${dir}" = "." ]; then
true
else
if [ -z "${multidirs}" ]; then
multidirs="${dir}"
else
multidirs="${multidirs} ${dir}"
fi
fi
done
# Target libraries are configured for the host they run on, so we check
# $host here, not $target.
case "${host}" in
arc-*-elf*)
if [ x$enable_biendian != xyes ]
then
old_multidirs=${multidirs}
multidirs=""
for x in ${old_multidirs}; do
case "${x}" in
*be*) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
;;
arm-*-*)
if [ x"$enable_fpu" = xno ]
then
old_multidirs=${multidirs}
multidirs=""
for x in ${old_multidirs}; do
case "${x}" in
*fpu*) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x"$enable_26bit" = xno ]
then
old_multidirs=${multidirs}
multidirs=""
for x in ${old_multidirs}; do
case "${x}" in
*26bit*) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x"$enable_underscore" = xno ]
then
old_multidirs=${multidirs}
multidirs=""
for x in ${old_multidirs}; do
case "${x}" in
*under*) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x"$enable_interwork" = xno ]
then
old_multidirs=${multidirs}
multidirs=""
for x in ${old_multidirs}; do
case "${x}" in
*interwork*) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_biendian = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*le* ) : ;;
*be* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x"$enable_nofmult" = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*nofmult* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
;;
m68*-*-*)
if [ x$enable_softfloat = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*soft-float* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_m68881 = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*m68881* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_m68000 = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*m68000* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_m68020 = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*m68020* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
;;
mips*-*-*)
if [ x$enable_single_float = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*single* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_biendian = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*el* ) : ;;
*eb* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_softfloat = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*soft-float* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
case " $multidirs " in
*" mabi=64 "*)
# We will not be able to create libraries with -mabi=64 if
# we cannot even link a trivial program. It usually
# indicates the 64bit libraries are missing.
if echo 'main() {}' > conftest.c &&
${CC-gcc} -mabi=64 conftest.c -o conftest; then
:
else
echo Could not link program with -mabi=64, disabling it.
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*mabi=64* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
rm -f conftest.c conftest
;;
esac
;;
powerpc*-*-* | rs6000*-*-*)
if [ x$enable_softfloat = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*soft-float* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_powercpu = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
power | */power | */power/* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_powerpccpu = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*powerpc* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_powerpcos = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*mcall-linux* | *mcall-solaris* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_biendian = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*mlittle* | *mbig* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_sysv = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*mcall-sysv* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
if [ x$enable_aix = xno ]
then
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*mcall-aix* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
;;
sparc*-*-*)
case " $multidirs " in
*" m64 "*)
# We will not be able to create libraries with -m64 if
# we cannot even link a trivial program. It usually
# indicates the 64bit libraries are missing.
if echo 'main() {}' > conftest.c &&
${CC-gcc} -m64 conftest.c -o conftest; then
:
else
echo Could not link program with -m64, disabling it.
old_multidirs="${multidirs}"
multidirs=""
for x in ${old_multidirs}; do
case "$x" in
*m64* ) : ;;
*) multidirs="${multidirs} ${x}" ;;
esac
done
fi
rm -f conftest.c conftest
;;
esac
;;
esac
# Remove extraneous blanks from multidirs.
# Tests like `if [ -n "$multidirs" ]' require it.
multidirs=`echo "$multidirs" | sed -e 's/^[ ][ ]*//' -e 's/[ ][ ]*$//' -e 's/[ ][ ]*/ /g'`
# Add code to library's top level makefile to handle building the multilib
# subdirs.
cat > Multi.tem <<\EOF
# FIXME: There should be an @-sign in front of the `if'.
# Leave out until this is tested a bit more.
multi-do:
if [ -z "$(MULTIDIRS)" ]; then \
true; \
else \
rootpre=`pwd`/; export rootpre; \
srcrootpre=`cd $(srcdir); pwd`/; export srcrootpre; \
lib=`echo $${rootpre} | sed -e 's,^.*/\([^/][^/]*\)/$$,\1,'`; \
compiler="$(CC)"; \
for i in `$${compiler} --print-multi-lib 2>/dev/null`; do \
dir=`echo $$i | sed -e 's/;.*$$//'`; \
if [ "$${dir}" = "." ]; then \
true; \
else \
if [ -d ../$${dir}/$${lib} ]; then \
flags=`echo $$i | sed -e 's/^[^;]*;//' -e 's/@/ -/g'`; \
if (cd ../$${dir}/$${lib}; $(MAKE) $(FLAGS_TO_PASS) \
CFLAGS="$(CFLAGS) $${flags}" \
prefix="$(prefix)" \
exec_prefix="$(exec_prefix)" \
GCJFLAGS="$(GCJFLAGS) $${flags}" \
CXXFLAGS="$(CXXFLAGS) $${flags}" \
LIBCFLAGS="$(LIBCFLAGS) $${flags}" \
LIBCXXFLAGS="$(LIBCXXFLAGS) $${flags}" \
LDFLAGS="$(LDFLAGS) $${flags}" \
$(DO)); then \
true; \
else \
exit 1; \
fi; \
else true; \
fi; \
fi; \
done; \
fi
# FIXME: There should be an @-sign in front of the `if'.
# Leave out until this is tested a bit more.
multi-clean:
if [ -z "$(MULTIDIRS)" ]; then \
true; \
else \
lib=`pwd | sed -e 's,^.*/\([^/][^/]*\)$$,\1,'`; \
for dir in Makefile $(MULTIDIRS); do \
if [ -f ../$${dir}/$${lib}/Makefile ]; then \
if (cd ../$${dir}/$${lib}; $(MAKE) $(FLAGS_TO_PASS) $(DO)); \
then true; \
else exit 1; \
fi; \
else true; \
fi; \
done; \
fi
EOF
cat ${Makefile} Multi.tem > Makefile.tem
rm -f ${Makefile} Multi.tem
mv Makefile.tem ${Makefile}
fi # ${ml_toplevel_p} = yes
if [ "${ml_verbose}" = --verbose ]; then
echo "Adding multilib support to Makefile in ${ml_realsrcdir}"
if [ "${ml_toplevel_p}" = yes ]; then
echo "multidirs=${multidirs}"
fi
echo "with_multisubdir=${with_multisubdir}"
fi
if [ "${srcdir}" = "." ]; then
if [ "${with_target_subdir}" != "." ]; then
ml_srcdotdot="../"
else
ml_srcdotdot=""
fi
else
ml_srcdotdot=""
fi
if [ -z "${with_multisubdir}" ]; then
ml_subdir=
ml_builddotdot=
: # ml_srcdotdot= # already set
else
ml_subdir="/${with_multisubdir}"
# The '[^/][^/]*' appears that way to work around a SunOS sed bug.
ml_builddotdot=`echo ${with_multisubdir} | sed -e 's:[^/][^/]*:..:g'`/
if [ "$srcdir" = "." ]; then
ml_srcdotdot=${ml_srcdotdot}${ml_builddotdot}
else
: # ml_srcdotdot= # already set
fi
fi
if [ "${ml_toplevel_p}" = yes ]; then
ml_do='$(MAKE)'
ml_clean='$(MAKE)'
else
ml_do=true
ml_clean=true
fi
# TOP is used by newlib and should not be used elsewhere for this purpose.
# MULTI{SRC,BUILD}TOP are the proper ones to use. MULTISRCTOP is empty
# when srcdir != builddir. MULTIBUILDTOP is always some number of ../'s.
# FIXME: newlib needs to be updated to use MULTI{SRC,BUILD}TOP so we can
# delete TOP. Newlib may wish to continue to use TOP for its own purposes
# of course.
# MULTIDIRS is non-empty for the cpu top level Makefile (eg: newlib/Makefile)
# and lists the subdirectories to recurse into.
# MULTISUBDIR is non-empty in each cpu subdirectory's Makefile
# (eg: newlib/h8300h/Makefile) and is the installed subdirectory name with
# a leading '/'.
# MULTIDO is used for targets like all, install, and check where
# $(FLAGS_TO_PASS) augmented with the subdir's compiler option is needed.
# MULTICLEAN is used for the *clean targets.
#
# ??? It is possible to merge MULTIDO and MULTICLEAN into one. They are
# currently kept separate because we don't want the *clean targets to require
# the existence of the compiler (which MULTIDO currently requires) and
# therefore we'd have to record the directory options as well as names
# (currently we just record the names and use --print-multi-lib to get the
# options).
sed -e "s:^TOP[ ]*=[ ]*\([./]*\)[ ]*$:TOP = ${ml_builddotdot}\1:" \
-e "s:^MULTISRCTOP[ ]*=.*$:MULTISRCTOP = ${ml_srcdotdot}:" \
-e "s:^MULTIBUILDTOP[ ]*=.*$:MULTIBUILDTOP = ${ml_builddotdot}:" \
-e "s:^MULTIDIRS[ ]*=.*$:MULTIDIRS = ${multidirs}:" \
-e "s:^MULTISUBDIR[ ]*=.*$:MULTISUBDIR = ${ml_subdir}:" \
-e "s:^MULTIDO[ ]*=.*$:MULTIDO = $ml_do:" \
-e "s:^MULTICLEAN[ ]*=.*$:MULTICLEAN = $ml_clean:" \
${Makefile} > Makefile.tem
rm -f ${Makefile}
mv Makefile.tem ${Makefile}
# If this is the library's top level, configure each multilib subdir.
# This is done at the end because this is the loop that runs configure
# in each multilib subdir and it seemed reasonable to finish updating the
# Makefile before going on to configure the subdirs.
if [ "${ml_toplevel_p}" = yes ]; then
# We must freshly configure each subdirectory. This bit of code is
# actually partially stolen from the main configure script. FIXME.
if [ -n "${multidirs}" ] && [ -z "${ml_norecursion}" ]; then
if [ "${ml_verbose}" = --verbose ]; then
echo "Running configure in multilib subdirs ${multidirs}"
echo "pwd: `pwd`"
fi
ml_origdir=`pwd`
ml_libdir=`echo $ml_origdir | sed -e 's,^.*/,,'`
# cd to top-level-build-dir/${with_target_subdir}
cd ..
for ml_dir in ${multidirs}; do
if [ "${ml_verbose}" = --verbose ]; then
echo "Running configure in multilib subdir ${ml_dir}"
echo "pwd: `pwd`"
fi
if [ -d ${ml_dir} ]; then true; else
# ``mkdir -p ${ml_dir}'' See also mkinstalldirs.
pathcomp=""
for d in `echo ":${ml_dir}" | sed -ne 's/^:\//#/;s/^://;s/\// /g;s/^#/\//;p'`; do
pathcomp="$pathcomp$d"
case "$pathcomp" in
-* ) pathcomp=./$pathcomp ;;
esac
if test ! -d "$pathcomp"; then
echo "mkdir $pathcomp" 1>&2
mkdir "$pathcomp" > /dev/null 2>&1 || lasterr=$?
fi
if test ! -d "$pathcomp"; then
exit $lasterr
fi
pathcomp="$pathcomp/"
done
fi
if [ -d ${ml_dir}/${ml_libdir} ]; then true; else mkdir ${ml_dir}/${ml_libdir}; fi
# Eg: if ${ml_dir} = m68000/m68881, dotdot = ../../
dotdot=../`echo ${ml_dir} | sed -e 's|[^/]||g' -e 's|/|../|g'`
case ${srcdir} in
".")
echo Building symlink tree in `pwd`/${ml_dir}/${ml_libdir}
if [ "${with_target_subdir}" != "." ]; then
ml_unsubdir="../"
else
ml_unsubdir=""
fi
(cd ${ml_dir}/${ml_libdir};
../${dotdot}${ml_unsubdir}symlink-tree ../${dotdot}${ml_unsubdir}${ml_libdir} "")
if [ -f ${ml_dir}/${ml_libdir}/Makefile ]; then
if [ x"${MAKE}" = x ]; then
(cd ${ml_dir}/${ml_libdir}; make distclean)
else
(cd ${ml_dir}/${ml_libdir}; ${MAKE} distclean)
fi
fi
ml_newsrcdir="."
ml_srcdiroption=
multisrctop=${dotdot}
;;
*)
case "${srcdir}" in
/* | [A-Za-z]:[\\/]* ) # absolute path
ml_newsrcdir=${srcdir}
;;
*) # otherwise relative
ml_newsrcdir=${dotdot}${srcdir}
;;
esac
ml_srcdiroption="-srcdir=${ml_newsrcdir}"
multisrctop=
;;
esac
case "${progname}" in
/* | [A-Za-z]:[\\/]* ) ml_recprog=${progname} ;;
*) ml_recprog=${dotdot}${progname} ;;
esac
# FIXME: POPDIR=${PWD=`pwd`} doesn't work here.
ML_POPDIR=`pwd`
cd ${ml_dir}/${ml_libdir}
if [ -f ${ml_newsrcdir}/configure ]; then
ml_recprog="${ml_newsrcdir}/configure --cache-file=../config.cache"
fi
# find compiler flag corresponding to ${ml_dir}
for i in `${CC-gcc} --print-multi-lib 2>/dev/null`; do
dir=`echo $i | sed -e 's/;.*$//'`
if [ "${dir}" = "${ml_dir}" ]; then
flags=`echo $i | sed -e 's/^[^;]*;//' -e 's/@/ -/g'`
break
fi
done
ml_config_env='CC="${CC_}$flags" CXX="${CXX_}$flags" GCJ="${GCJ_}$flags"'
if [ "${with_target_subdir}" = "." ]; then
CC_=$CC' '
CXX_=$CXX' '
GCJ_=$GCJ' '
else
# Create a regular expression that matches any string as long
# as ML_POPDIR.
popdir_rx=`echo ${ML_POPDIR} | sed 's,.,.,g'`
CC_=
for arg in ${CC}; do
case $arg in
-[BIL]"${ML_POPDIR}"/*)
CC_="${CC_}"`echo "X${arg}" | sed -n "s/X\\(-[BIL]${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X-[BIL]${popdir_rx}\\(.*\\)/\1/p"`' ' ;;
"${ML_POPDIR}"/*)
CC_="${CC_}"`echo "X${arg}" | sed -n "s/X\\(${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X${popdir_rx}\\(.*\\)/\\1/p"`' ' ;;
*)
CC_="${CC_}${arg} " ;;
esac
done
CXX_=
for arg in ${CXX}; do
case $arg in
-[BIL]"${ML_POPDIR}"/*)
CXX_="${CXX_}"`echo "X${arg}" | sed -n "s/X\\(-[BIL]${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X-[BIL]${popdir_rx}\\(.*\\)/\\1/p"`' ' ;;
"${ML_POPDIR}"/*)
CXX_="${CXX_}"`echo "X${arg}" | sed -n "s/X\\(${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X${popdir_rx}\\(.*\\)/\\1/p"`' ' ;;
*)
CXX_="${CXX_}${arg} " ;;
esac
done
GCJ_=
for arg in ${GCJ}; do
case $arg in
-[BIL]"${ML_POPDIR}"/*)
GCJ_="${GCJ_}"`echo "X${arg}" | sed -n "s/X\\(-[BIL]${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X-[BIL]${popdir_rx}\\(.*\\)/\\1/p"`' ' ;;
"${ML_POPDIR}"/*)
GCJ_="${GCJ_}"`echo "X${arg}" | sed -n "s/X\\(${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X${popdir_rx}\\(.*\\)/\\1/p"`' ' ;;
*)
GCJ_="${GCJ_}${arg} " ;;
esac
done
if test "x${LD_LIBRARY_PATH+set}" = xset; then
LD_LIBRARY_PATH_=
for arg in `echo "$LD_LIBRARY_PATH" | tr ':' ' '`; do
case "$arg" in
"${ML_POPDIR}"/*)
arg=`echo "X${arg}" | sed -n "s/X\\(${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X${popdir_rx}\\(.*\\)/\\1/p"`
;;
esac
if test "x$LD_LIBRARY_PATH_" != x; then
LD_LIBRARY_PATH_=$LD_LIBRARY_PATH_:$arg
else
LD_LIBRARY_PATH_=$arg
fi
done
ml_config_env="$ml_config_env LD_LIBRARY_PATH=$LD_LIBRARY_PATH_"
fi
if test "x${SHLIB_PATH+set}" = xset; then
SHLIB_PATH_=
for arg in `echo "$SHLIB_PATH" | tr ':' ' '`; do
case "$arg" in
"${ML_POPDIR}"/*)
arg=`echo "X${arg}" | sed -n "s/X\\(${popdir_rx}\\).*/\\1/p"`/${ml_dir}`echo "X${arg}" | sed -n "s/X${popdir_rx}\\(.*\\)/\\1/p"`
;;
esac
if test "x$SHLIB_PATH_" != x; then
SHLIB_PATH_=$SHLIB_PATH_:$arg
else
SHLIB_PATH_=$arg
fi
done
ml_config_env="$ml_config_env SHLIB_PATH=$SHLIB_PATH_"
fi
fi
if eval ${ml_config_env} ${ml_config_shell} ${ml_recprog} \
--with-multisubdir=${ml_dir} --with-multisrctop=${multisrctop} \
${ml_arguments} ${ml_srcdiroption} ; then
true
else
exit 1
fi
cd ${ML_POPDIR}
done
cd ${ml_origdir}
fi
fi # ${ml_toplevel_p} = yes
fi # ${enable_multilib} = yes
+166
View File
@@ -0,0 +1,166 @@
dnl Process this with autoconf to create configure
AC_INIT(fficonfig.h.in)
AM_CONFIG_HEADER(fficonfig.h)
dnl Default to --enable-multilib
AC_ARG_ENABLE(multilib,
[ --enable-multilib build many library versions (default)],
[case "${enableval}" in
yes) multilib=yes ;;
no) multilib=no ;;
*) AC_MSG_ERROR(bad value ${enableval} for multilib option) ;;
esac], [multilib=yes])dnl
dnl We may get other options which we don't document:
dnl --with-target-subdir, --with-multisrctop, --with-multisubdir
### I don't understand this, so comment it out for now
#if test "${srcdir}" = "."; then
# if test "${with_target_subdir}" != "."; then
# libffi_basedir="${srcdir}/${with_multisrctop}.."
# else
# libffi_basedir="${srcdir}/${with_multisrctop}"
# fi
#else
libffi_basedir="${srcdir}"
#fi
AC_SUBST(libffi_basedir)
AC_CANONICAL_HOST
AM_INIT_AUTOMAKE(libffi,2.00-beta,no-define)
AC_EXEEXT
AM_MAINTAINER_MODE
AC_PROG_CC
AC_PROG_RANLIB
AC_PROG_LN_S
dnl The -no-testsuite modules omit the test subdir.
AM_CONDITIONAL(TESTSUBDIR, test -d $srcdir/testsuite)
TARGET="unknown"
case "$host" in
alpha*-*-linux* | alpha*-*-osf*)
TARGET=ALPHA;
$LN_S ${srcdir}/alpha/ffi.c t-ffi.c;
$LN_S $srcdir/alpha/osf.S t-asm.S;;
arm*-*-linux-*)
TARGET=ARM;
$LN_S $srcdir/arm/ffi.c t-ffi.c;
$LN_S $srcdir/arm/ffi.S t-asm.S;;
i*86-*-beos*|i*86-*-freebsd*|i*86-*-linux*|i*86-*-solaris*)
TARGET=X86;
$LN_S $srcdir/x86/ffi.c t-ffi.c;
$LN_S $srcdir/x86/sysv.S t-asm.S;;
i*86-*-cygwin*|i*86-*-mingw*|i*86-*-win32*)
TARGET=X86_WIN32;
$LN_S $srcdir/x86/ffi.c t-ffi.c;
$LN_S $srcdir/x86/win32.S t-asm.S;;
ia64*-*-*)
TARGET=IA64;
$LN_S $srcdir/ia64/ffi.c t-ffi.c;
$LN_S $srcdir/ia64/unix.S t-asm.S;;
m68k-*-linux*)
TARGET=M68K;
$LN_S $srcdir/m68k/ffi.c t-ffi.c;
$LN_S $srcdir/m68k/unix.S t-asm.S;;
mips-*-elf)
TARGET=MIPS;
$LN_S $srcdir/mips/ffi.c t-ffi.c;
$LN_S $srcdir/mips/o32.S t-asm.S;;
mips-sgi-irix5.*)
TARGET=MIPS;
$LN_S $srcdir/mips/ffi.c t-ffi.c;
$LN_S $srcdir/mips/o32.S t-asm.S;;
mips-sgi-irix6.*)
TARGET=MIPS;
$LN_S $srcdir/mips/ffi.c t-ffi.c;
$LN_S $srcdir/mips/n32.S t-asm.S;;
powerpc-*-linux* | powerpc-*-sysv* | powerpc-*-beos*)
TARGET=POWERPC;
$LN_S $srcdir/powerpc/ffi.c t-ffi.c;
$LN_S $srcdir/powerpc/sysv.S t-asm.S;;
s390-*-linux*)
TARGET=S390;
$LN_S $srcdir/s390/ffi.c t-ffi.c;
$LN_S $srcdir/s390/v8.S t-asm.S;;
sparc-*-linux*|sparc-sun-*)
TARGET=SPARC;
$LN_S $srcdir/sparc/ffi.c t-ffi.c;
$LN_S $srcdir/sparc/v8.S t-asm.S;;
sparc64-*-linux*)
TARGET=SPARC;
$LN_S $srcdir/sparc/ffi.c t-ffi.c;
$LN_S $srcdir/sparc/v9.S t-asm.S;;
esac
AC_SUBST(AM_RUNTESTFLAGS)
if test ${TARGET} = unknown; then
AC_ERROR("libffi has not been ported to $host.")
fi
AC_HEADER_STDC
AC_CHECK_FUNCS(memcpy)
AC_FUNC_ALLOCA
dnl AC_CHECK_SIZEOF(char)
AC_COMPILE_CHECK_SIZEOF(short)
AC_COMPILE_CHECK_SIZEOF(int)
AC_COMPILE_CHECK_SIZEOF(long)
AC_COMPILE_CHECK_SIZEOF(long long)
AC_COMPILE_CHECK_SIZEOF(float)
AC_COMPILE_CHECK_SIZEOF(double)
AC_COMPILE_CHECK_SIZEOF(long double)
AC_COMPILE_CHECK_SIZEOF(void *)
AC_C_BIGENDIAN_CROSS
AC_SUBST(TARGET)
AC_SUBST(TARGETDIR)
AC_SUBST(SHELL)
AC_ARG_ENABLE(debug,[ --enable-debug debugging mode], AC_DEFINE(FFI_DEBUG))
AC_ARG_ENABLE(debug,[ --disable-structs omit code for struct support], AC_DEFINE(FFI_NO_STRUCTS))
AC_ARG_ENABLE(debug,[ --disable-raw-api make the raw api unavailable], AC_DEFINE(FFI_NO_RAW_API))
AC_ARG_ENABLE(purify-safety,
[ --enable-purify-safety purify-safe mode], AC_DEFINE(USING_PURIFY))
AM_CONDITIONAL(USE_LIBDIR, test -z "$with_cross_host")
if test "${multilib}" = "yes"; then
multilib_arg="--enable-multilib"
else
multilib_arg=
fi
AC_OUTPUT(include/Makefile testsuite/Makefile Makefile,
[
if test -n "$CONFIG_FILES"; then
ac_file=Makefile . ${libffi_basedir}/config-ml.in
fi
],
srcdir=${srcdir}
host=${host}
target=${target}
with_multisubdir=${with_multisubdir}
ac_configure_args="${multilib_arg} ${ac_configure_args}"
CONFIG_SHELL=${CONFIG_SHELL-/bin/sh}
libffi_basedir=${libffi_basedir}
CC="${CC}"
DEFS="$DEFS"
test ! -d include && mkdir include
test ! -f include/fficonfig.h && cp fficonfig.h include/fficonfig.h
if cmp -s fficonfig.h include/fficonfig.h 2>/dev/null; then
echo fficonfig.h unchanged
else
echo Moving fficonfig.h to include/fficonfig.h
cp fficonfig.h include/fficonfig.h
fi
)
+65
View File
@@ -0,0 +1,65 @@
/* -----------------------------------------------------------------------
debug.c - Copyright (c) 1996 Red Hat, Inc.
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 RED HAT 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
#include <stdio.h>
/* General debugging routines */
void ffi_stop_here(void)
{
/* This function is only useful for debugging purposes.
Place a breakpoint on ffi_stop_here to be notified of
significant events. */
}
/* This function should only be called via the FFI_ASSERT() macro */
int ffi_assert(char *file, int line)
{
fprintf(stderr, "ASSERTION FAILURE: %s line %d\n", file, line);
ffi_stop_here();
abort();
/* This has to return something for the compiler not to complain */
/*@notreached@*/
return 0;
}
/* Perform a sanity check on an ffi_type structure */
bool ffi_type_test(ffi_type *a)
{
/*@-usedef@*/
FFI_ASSERT(a->type <= FFI_TYPE_LAST);
FFI_ASSERT(a->type > FFI_TYPE_VOID ? a->size > 0 : 1);
FFI_ASSERT(a->type > FFI_TYPE_VOID ? a->alignment > 0 : 1);
FFI_ASSERT(a->type == FFI_TYPE_STRUCT ? a->elements != NULL : 1);
/*@=usedef@*/
/* This is a silly thing to return, but it keeps the compiler from
issuing warnings about "a" not being used in non-debug builds. */
return (a != NULL);
}
+736
View File
@@ -0,0 +1,736 @@
/* -----------------------------------------------------------------------
ffitest.c - Copyright (c) 1996, 1997, 1998 Cygnus Solutions
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 RED HAT 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 <ffi.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <float.h>
/* This is lame. Long double support is barely there under SunOS 4.x */
#if defined(__sparc__) && (SIZEOF_LONG_DOUBLE != 16)
#define BROKEN_LONG_DOUBLE
#endif
#define CHECK(x) !(x) ? fail(__FILE__, __LINE__) : 0
static int fail(char *file, int line)
{
fprintf(stderr, "Test failure: %s line %d\n", file, line);
exit(EXIT_FAILURE);
/*@notreached@*/
return 0;
}
#define MAX_ARGS 256
static size_t my_strlen(char *s)
{
return (strlen(s));
}
static int promotion(signed char sc, signed short ss,
unsigned char uc, unsigned short us)
{
int r = (int) sc + (int) ss + (int) uc + (int) us;
return r;
}
static signed char return_sc(signed char sc)
{
return sc;
}
static unsigned char return_uc(unsigned char uc)
{
return uc;
}
static long long return_ll(long long ll)
{
return ll;
}
static int floating(int a, float b, double c, long double d, int e)
{
int i;
#if 0
/* This is ifdef'd out for now. long double support under SunOS/gcc
is pretty much non-existent. You'll get the odd bus error in library
routines like printf(). */
printf("%d %f %f %Lf %d\n", a, (double)b, c, d, e);
#endif
i = (int) ((float)a/b + ((float)c/(float)d));
return i;
}
static float many(float f1,
float f2,
float f3,
float f4,
float f5,
float f6,
float f7,
float f8,
float f9,
float f10,
float f11,
float f12,
float f13)
{
#if 0
printf("%f %f %f %f %f %f %f %f %f %f %f %f %f\n",
(double) f1, (double) f2, (double) f3, (double) f4, (double) f5,
(double) f6, (double) f7, (double) f8, (double) f9, (double) f10,
(double) f11, (double) f12, (double) f13);
#endif
return ((f1/f2+f3/f4+f5/f6+f7/f8+f9/f10+f11/f12) * f13);
}
static double dblit(float f)
{
return f/3.0;
}
static long double ldblit(float f)
{
return (long double) (((long double) f)/ (long double) 3.0);
}
typedef struct
{
unsigned char uc;
double d;
unsigned int ui;
} test_structure_1;
typedef struct
{
double d1;
double d2;
} test_structure_2;
typedef struct
{
int si;
} test_structure_3;
typedef struct
{
unsigned ui1;
unsigned ui2;
unsigned ui3;
} test_structure_4;
typedef struct
{
char c1;
char c2;
} test_structure_5;
static test_structure_1 struct1(test_structure_1 ts)
{
/*@-type@*/
ts.uc++;
/*@=type@*/
ts.d--;
ts.ui++;
return ts;
}
static test_structure_2 struct2(test_structure_2 ts)
{
ts.d1--;
ts.d2--;
return ts;
}
static test_structure_3 struct3(test_structure_3 ts)
{
ts.si = -(ts.si*2);
return ts;
}
static test_structure_4 struct4(test_structure_4 ts)
{
ts.ui3 = ts.ui1 * ts.ui2 * ts.ui3;
return ts;
}
static test_structure_5 struct5(test_structure_5 ts1, test_structure_5 ts2)
{
ts1.c1 += ts2.c1;
ts1.c2 -= ts2.c2;
return ts1;
}
/* Take an int and a float argument, together with int userdata, and */
/* return the sum. */
static void closure_test_fn(ffi_cif* cif,void* resp,void** args, void* userdata)
{
*(int*)resp =
*(int *)args[0] + (int)(*(float *)args[1]) + (int)(long)userdata;
}
typedef int (*closure_test_type)(int, float);
int main(/*@unused@*/ int argc, /*@unused@*/ char *argv[])
{
ffi_cif cif;
ffi_type *args[MAX_ARGS];
void *values[MAX_ARGS];
char *s;
signed char sc;
unsigned char uc;
signed short ss;
unsigned short us;
unsigned long ul;
long long ll;
float f;
double d;
long double ld;
signed int si1;
signed int si2;
#if defined(__alpha__) || defined(IA64) || defined(SPARC64) || (defined(__mips__) && (_MIPS_SIM == _ABIN32))
long long rint;
#else
int rint;
#endif
long long rlonglong;
ffi_type ts1_type;
ffi_type ts2_type;
ffi_type ts3_type;
ffi_type ts4_type;
ffi_type ts5_type;
ffi_type *ts1_type_elements[4];
ffi_type *ts2_type_elements[3];
ffi_type *ts3_type_elements[2];
ffi_type *ts4_type_elements[4];
ffi_type *ts5_type_elements[3];
ts1_type.size = 0;
ts1_type.alignment = 0;
ts1_type.type = FFI_TYPE_STRUCT;
ts2_type.size = 0;
ts2_type.alignment = 0;
ts2_type.type = FFI_TYPE_STRUCT;
ts3_type.size = 0;
ts3_type.alignment = 0;
ts3_type.type = FFI_TYPE_STRUCT;
ts4_type.size = 0;
ts4_type.alignment = 0;
ts4_type.type = FFI_TYPE_STRUCT;
ts5_type.size = 0;
ts5_type.alignment = 0;
ts5_type.type = FFI_TYPE_STRUCT;
/*@-immediatetrans@*/
ts1_type.elements = ts1_type_elements;
ts2_type.elements = ts2_type_elements;
ts3_type.elements = ts3_type_elements;
ts4_type.elements = ts4_type_elements;
ts5_type.elements = ts5_type_elements;
/*@=immediatetrans@*/
ts1_type_elements[0] = &ffi_type_uchar;
ts1_type_elements[1] = &ffi_type_double;
ts1_type_elements[2] = &ffi_type_uint;
ts1_type_elements[3] = NULL;
ts2_type_elements[0] = &ffi_type_double;
ts2_type_elements[1] = &ffi_type_double;
ts2_type_elements[2] = NULL;
ts3_type_elements[0] = &ffi_type_sint;
ts3_type_elements[1] = NULL;
ts4_type_elements[0] = &ffi_type_uint;
ts4_type_elements[1] = &ffi_type_uint;
ts4_type_elements[2] = &ffi_type_uint;
ts4_type_elements[3] = NULL;
ts5_type_elements[0] = &ffi_type_schar;
ts5_type_elements[1] = &ffi_type_schar;
ts5_type_elements[2] = NULL;
ul = 0;
/* return value tests */
{
#if defined(__mips__) /* || defined(ARM) */
puts ("long long tests not run. This is a known bug on this architecture.");
#else
args[0] = &ffi_type_sint64;
values[0] = &ll;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_sint64, args) == FFI_OK);
for (ll = 0LL; ll < 100LL; ll++)
{
ul++;
ffi_call(&cif, FFI_FN(return_ll), &rlonglong, values);
CHECK(rlonglong == ll);
}
for (ll = 55555555555000LL; ll < 55555555555100LL; ll++)
{
ul++;
ffi_call(&cif, FFI_FN(return_ll), &rlonglong, values);
CHECK(rlonglong == ll);
}
#endif
args[0] = &ffi_type_schar;
values[0] = &sc;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_schar, args) == FFI_OK);
for (sc = (signed char) -127;
sc < (signed char) 127; /*@-type@*/ sc++ /*@=type@*/)
{
ul++;
ffi_call(&cif, FFI_FN(return_sc), &rint, values);
CHECK(rint == (int) sc);
}
args[0] = &ffi_type_uchar;
values[0] = &uc;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_uchar, args) == FFI_OK);
for (uc = (unsigned char) '\x00';
uc < (unsigned char) '\xff'; /*@-type@*/ uc++ /*@=type@*/)
{
ul++;
ffi_call(&cif, FFI_FN(return_uc), &rint, values);
CHECK(rint == (signed int) uc);
}
printf("%lu return value tests run\n", ul);
}
#ifdef BROKEN_LONG_DOUBLE
printf ("This architecture has broken `long double' support. No floating point\ntests have been run.\n");
#else
/* float arg tests */
{
args[0] = &ffi_type_float;
values[0] = &f;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_longdouble, args) == FFI_OK);
f = 3.14159;
#if 0
/* This is ifdef'd out for now. long double support under SunOS/gcc
is pretty much non-existent. You'll get the odd bus error in library
routines like printf(). */
printf ("%Lf\n", ldblit(f));
#endif
ld = 666;
ffi_call(&cif, FFI_FN(ldblit), &ld, values);
#if 0
/* This is ifdef'd out for now. long double support under SunOS/gcc
is pretty much non-existent. You'll get the odd bus error in library
routines like printf(). */
printf ("%Lf, %Lf, %Lf, %Lf\n", ld, ldblit(f), ld - ldblit(f), LDBL_EPSILON);
#endif
/* These are not always the same!! Check for a reasonable delta */
/*@-realcompare@*/
if (ld - ldblit(f) < LDBL_EPSILON)
/*@=realcompare@*/
puts("long double return value tests ok!");
else
CHECK(0);
}
/* float arg tests */
{
args[0] = &ffi_type_sint;
values[0] = &si1;
args[1] = &ffi_type_float;
values[1] = &f;
args[2] = &ffi_type_double;
values[2] = &d;
args[3] = &ffi_type_longdouble;
values[3] = &ld;
args[4] = &ffi_type_sint;
values[4] = &si2;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 5,
&ffi_type_sint, args) == FFI_OK);
si1 = 6;
f = 3.14159;
d = (double)1.0/(double)3.0;
ld = 2.71828182846L;
si2 = 10;
floating (si1, f, d, ld, si2);
ffi_call(&cif, FFI_FN(floating), &rint, values);
printf ("%d vs %d\n", rint, floating (si1, f, d, ld, si2));
CHECK(rint == floating(si1, f, d, ld, si2));
printf("float arg tests ok!\n");
}
#endif
/* strlen tests */
{
args[0] = &ffi_type_pointer;
values[0] = (void*) &s;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_sint, args) == FFI_OK);
s = "a";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 1);
s = "1234567";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 7);
s = "1234567890123456789012345";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 25);
printf("strlen tests passed\n");
}
/* float arg tests */
{
args[0] = &ffi_type_float;
values[0] = &f;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_double, args) == FFI_OK);
f = 3.14159;
ffi_call(&cif, FFI_FN(dblit), &d, values);
/* These are not always the same!! Check for a reasonable delta */
/*@-realcompare@*/
CHECK(d - dblit(f) < DBL_EPSILON);
/*@=realcompare@*/
printf("double return value tests ok!\n");
}
/* many arg tests */
{
float ff;
float fa[13];
for (ul = 0; ul < 13; ul++)
{
args[ul] = &ffi_type_float;
values[ul] = &fa[ul];
fa[ul] = (float) ul;
}
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 13,
&ffi_type_float, args) == FFI_OK);
/*@-usedef@*/
ff = many(fa[0], fa[1],
fa[2], fa[3],
fa[4], fa[5],
fa[6], fa[7],
fa[8], fa[9],
fa[10],fa[11],fa[12]);
/*@=usedef@*/
ffi_call(&cif, FFI_FN(many), &f, values);
/*@-realcompare@*/
if (f - ff < FLT_EPSILON)
/*@=realcompare@*/
printf("many arg tests ok!\n");
else
#ifdef POWERPC
printf("many arg tests failed! This is a gcc bug.\n");
#else
CHECK(0);
#endif
}
/* promotion tests */
{
args[0] = &ffi_type_schar;
args[1] = &ffi_type_sshort;
args[2] = &ffi_type_uchar;
args[3] = &ffi_type_ushort;
values[0] = &sc;
values[1] = &ss;
values[2] = &uc;
values[3] = &us;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 4,
&ffi_type_sint, args) == FFI_OK);
us = 0;
ul = 0;
for (sc = (signed char) -127;
sc <= (signed char) 120; /*@-type@*/ sc += 1 /*@=type@*/)
for (ss = -30000; ss <= 30000; ss += 10000)
for (uc = (unsigned char) 0;
uc <= (unsigned char) 200; /*@-type@*/ uc += 20 /*@=type@*/)
for (us = 0; us <= 60000; us += 10000)
{
ul++;
ffi_call(&cif, FFI_FN(promotion), &rint, values);
CHECK(rint == (int) sc + (int) ss + (int) uc + (int) us);
}
printf("%lu promotion tests run\n", ul);
}
#ifndef X86_WIN32 /* Structures dont work on Win32 */
/* struct tests */
{
test_structure_1 ts1_arg;
/* This is a hack to get a properly aligned result buffer */
test_structure_1 *ts1_result =
(test_structure_1 *) malloc (sizeof(test_structure_1));
args[0] = &ts1_type;
values[0] = &ts1_arg;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ts1_type, args) == FFI_OK);
ts1_arg.uc = '\x01';
ts1_arg.d = 3.14159;
ts1_arg.ui = 555;
ffi_call(&cif, FFI_FN(struct1), ts1_result, values);
CHECK(ts1_result->ui == 556);
CHECK(ts1_result->d == 3.14159 - 1);
puts ("structure test 1 ok!\n");
free (ts1_result);
}
/* struct tests */
{
test_structure_2 ts2_arg;
/* This is a hack to get a properly aligned result buffer */
test_structure_2 *ts2_result =
(test_structure_2 *) malloc (sizeof(test_structure_2));
args[0] = &ts2_type;
values[0] = &ts2_arg;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ts2_type, args) == FFI_OK);
ts2_arg.d1 = 5.55;
ts2_arg.d2 = 6.66;
printf ("%g\n", ts2_result->d1);
printf ("%g\n", ts2_result->d2);
ffi_call(&cif, FFI_FN(struct2), ts2_result, values);
printf ("%g\n", ts2_result->d1);
printf ("%g\n", ts2_result->d2);
CHECK(ts2_result->d1 == 5.55 - 1);
CHECK(ts2_result->d2 == 6.66 - 1);
printf("structure test 2 ok!\n");
free (ts2_result);
}
/* struct tests */
{
int compare_value;
test_structure_3 ts3_arg;
test_structure_3 *ts3_result =
(test_structure_3 *) malloc (sizeof(test_structure_3));
args[0] = &ts3_type;
values[0] = &ts3_arg;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ts3_type, args) == FFI_OK);
ts3_arg.si = -123;
compare_value = ts3_arg.si;
ffi_call(&cif, FFI_FN(struct3), ts3_result, values);
printf ("%d %d\n", ts3_result->si, -(compare_value*2));
if (ts3_result->si == -(ts3_arg.si*2))
puts ("structure test 3 ok!");
else
{
puts ("Structure test 3 found structure passing bug.");
puts (" Current versions of GCC are not 100% compliant with the");
puts (" n32 ABI. There is a known problem related to passing");
puts (" small structures. Send a bug report to the gcc maintainers.");
}
free (ts3_result);
}
/* struct tests */
{
test_structure_4 ts4_arg;
/* This is a hack to get a properly aligned result buffer */
test_structure_4 *ts4_result =
(test_structure_4 *) malloc (sizeof(test_structure_4));
args[0] = &ts4_type;
values[0] = &ts4_arg;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ts4_type, args) == FFI_OK);
ts4_arg.ui1 = 2;
ts4_arg.ui2 = 3;
ts4_arg.ui3 = 4;
ffi_call (&cif, FFI_FN(struct4), ts4_result, values);
if (ts4_result->ui3 == 2U * 3U * 4U)
puts ("structure test 4 ok!");
else
puts ("Structure test 4 found GCC's structure passing bug.");
free (ts4_result);
}
/* struct tests */
{
test_structure_5 ts5_arg1, ts5_arg2;
/* This is a hack to get a properly aligned result buffer */
test_structure_5 *ts5_result =
(test_structure_5 *) malloc (sizeof(test_structure_5));
args[0] = &ts5_type;
args[1] = &ts5_type;
values[0] = &ts5_arg1;
values[1] = &ts5_arg2;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2,
&ts5_type, args) == FFI_OK);
ts5_arg1.c1 = 2;
ts5_arg1.c2 = 6;
ts5_arg2.c1 = 5;
ts5_arg2.c2 = 3;
ffi_call (&cif, FFI_FN(struct5), ts5_result, values);
if (ts5_result->c1 == 7
&& ts5_result->c2 == 3)
puts ("structure test 5 ok!");
else
puts ("Structure test 5 found GCC's structure passing bug.");
free (ts5_result);
}
#else
printf("Structure passing doesn't work on Win32.\n");
#endif /* X86_WIN32 */
# if FFI_CLOSURES
/* A simple closure test */
{
ffi_closure cl;
ffi_type * cl_arg_types[3];
cl_arg_types[0] = &ffi_type_sint;
cl_arg_types[1] = &ffi_type_float;
cl_arg_types[2] = NULL;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 2,
&ffi_type_sint, cl_arg_types) == FFI_OK);
CHECK(ffi_prep_closure(&cl, &cif, closure_test_fn,
(void *) 3 /* userdata */)
== FFI_OK);
CHECK((*((closure_test_type)(&cl)))(1, 2.0) == 6);
}
# endif
/* If we arrived here, all is good */
(void) puts("\nLooks good. No surprises.\n");
/*@-compdestroy@*/
return 0;
}
+670
View File
@@ -0,0 +1,670 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1998 Cygnus Solutions
Copyright (c) 2000 Hewlett Packard Company
IA64 Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
#include "ia64_flags.h"
/* Memory image of fp register contents. Should eventually be an fp */
/* type long enough to hold an entire register. For now we use double. */
typedef double float80;
/* The stack layout at call to ffi_prep_regs. Other_args will remain */
/* on the stack for the actual call. Everything else we be transferred */
/* to registers and popped by the assembly code. */
struct ia64_args {
long scratch[2]; /* Two scratch words at top of stack. */
/* Allows sp to passed as arg pointer. */
void * r8_contents; /* Value to be passed in r8 */
long spare; /* Not used. */
float80 fp_regs[8]; /* Contents of 8 floating point argument */
/* registers. */
long out_regs[8]; /* Contents of the 8 out registers used */
/* for integer parameters. */
long other_args[0]; /* Arguments passed on stack, variable size */
/* Treated as continuation of out_regs. */
};
static size_t float_type_size(unsigned short tp)
{
switch(tp) {
case FFI_TYPE_FLOAT:
return sizeof(float);
case FFI_TYPE_DOUBLE:
return sizeof(double);
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
return sizeof(long double);
#endif
default:
FFI_ASSERT(0);
}
}
/*
* Is type a struct containing at most n floats, doubles, or extended
* doubles, all of the same fp type?
* If so, set *element_type to the fp type.
*/
static bool is_homogeneous_fp_aggregate(ffi_type * type, int n,
unsigned short * element_type)
{
ffi_type **ptr;
unsigned short element, struct_element;
int type_set = 0;
FFI_ASSERT(type != NULL);
FFI_ASSERT(type->elements != NULL);
ptr = &(type->elements[0]);
while ((*ptr) != NULL)
{
switch((*ptr) -> type) {
case FFI_TYPE_FLOAT:
if (type_set && element != FFI_TYPE_FLOAT) return 0;
if (--n < 0) return FALSE;
type_set = 1;
element = FFI_TYPE_FLOAT;
break;
case FFI_TYPE_DOUBLE:
if (type_set && element != FFI_TYPE_DOUBLE) return 0;
if (--n < 0) return FALSE;
type_set = 1;
element = FFI_TYPE_DOUBLE;
break;
case FFI_TYPE_STRUCT:
if (!is_homogeneous_fp_aggregate(type, n, &struct_element))
return FALSE;
if (type_set && struct_element != element) return FALSE;
n -= (type -> size)/float_type_size(element);
element = struct_element;
if (n < 0) return FALSE;
break;
/* case FFI_TYPE_LONGDOUBLE:
Not yet implemented. */
default:
return FALSE;
}
ptr++;
}
*element_type = element;
return TRUE;
}
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments. Returns nonzero
if fp registers are used for arguments. */
static bool
ffi_prep_args(struct ia64_args *stack, extended_cif *ecif, int bytes)
{
register long i, avn;
register void **p_argv;
register long *argp = stack -> out_regs;
register float80 *fp_argp = stack -> fp_regs;
register ffi_type **p_arg;
/* For big return structs, r8 needs to contain the target address. */
/* Since r8 is otherwise dead, we set it unconditionally. */
stack -> r8_contents = ecif -> rvalue;
i = 0;
avn = ecif->cif->nargs;
p_arg = ecif->cif->arg_types;
p_argv = ecif->avalue;
while (i < avn)
{
size_t z; /* z is in units of arg slots or words, not bytes. */
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
z = 1;
*(SINT64 *) argp = *(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
z = 1;
*(UINT64 *) argp = *(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
z = 1;
*(SINT64 *) argp = *(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
z = 1;
*(UINT64 *) argp = *(UINT16 *)(* p_argv);
break;
case FFI_TYPE_SINT32:
z = 1;
*(SINT64 *) argp = *(SINT32 *)(* p_argv);
break;
case FFI_TYPE_UINT32:
z = 1;
*(UINT64 *) argp = *(UINT32 *)(* p_argv);
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
z = 1;
*(UINT64 *) argp = *(UINT64 *)(* p_argv);
break;
case FFI_TYPE_FLOAT:
z = 1;
if (fp_argp - stack->fp_regs < 8)
{
/* Note the conversion -- all the fp regs are loaded as
doubles. */
*fp_argp++ = *(float *)(* p_argv);
}
/* Also put it into the integer registers or memory: */
*(UINT64 *) argp = *(UINT32 *)(* p_argv);
break;
case FFI_TYPE_DOUBLE:
z = 1;
if (fp_argp - stack->fp_regs < 8)
*fp_argp++ = *(double *)(* p_argv);
/* Also put it into the integer registers or memory: */
*(double *) argp = *(double *)(* p_argv);
break;
case FFI_TYPE_STRUCT:
{
size_t sz = (*p_arg)->size;
unsigned short element_type;
z = ((*p_arg)->size + SIZEOF_ARG - 1)/SIZEOF_ARG;
if (is_homogeneous_fp_aggregate(*p_arg, 8, &element_type)) {
int i;
int nelements = sz/float_type_size(element_type);
for (i = 0; i < nelements; ++i) {
switch (element_type) {
case FFI_TYPE_FLOAT:
if (fp_argp - stack->fp_regs < 8)
*fp_argp++ = ((float *)(* p_argv))[i];
break;
case FFI_TYPE_DOUBLE:
if (fp_argp - stack->fp_regs < 8)
*fp_argp++ = ((double *)(* p_argv))[i];
break;
default:
/* Extended precision not yet implemented. */
abort();
}
}
}
/* And pass it in integer registers as a struct, with */
/* its actual field sizes packed into registers. */
memcpy(argp, *p_argv, (*p_arg)->size);
}
break;
default:
FFI_ASSERT(0);
}
argp += z;
i++, p_arg++, p_argv++;
}
return (fp_argp != stack -> fp_regs);
}
/* Perform machine dependent cif processing */
ffi_status
ffi_prep_cif_machdep(ffi_cif *cif)
{
long i, avn;
bool is_simple = TRUE;
long simple_flag = FFI_SIMPLE_V;
/* Adjust cif->bytes to include space for the 2 scratch words,
r8 register contents, spare word,
the 8 fp register contents, and all 8 integer register contents.
This will be removed before the call, though 2 scratch words must
remain. */
cif->bytes += 4*sizeof(long) + 8 *sizeof(float80);
if (cif->bytes < sizeof(struct ia64_args))
cif->bytes = sizeof(struct ia64_args);
/* The stack must be double word aligned, so round bytes up
appropriately. */
cif->bytes = ALIGN(cif->bytes, 2*sizeof(void*));
avn = cif->nargs;
if (avn <= 2) {
for (i = 0; i < avn; ++i) {
switch(cif -> arg_types[i] -> type) {
case FFI_TYPE_SINT32:
simple_flag = FFI_ADD_INT_ARG(simple_flag);
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
simple_flag = FFI_ADD_LONG_ARG(simple_flag);
break;
default:
is_simple = FALSE;
}
}
} else {
is_simple = FALSE;
}
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
cif->flags = FFI_TYPE_VOID;
break;
case FFI_TYPE_STRUCT:
{
size_t sz = cif -> rtype -> size;
unsigned short element_type;
is_simple = FALSE;
if (is_homogeneous_fp_aggregate(cif -> rtype, 8, &element_type)) {
int nelements = sz/float_type_size(element_type);
if (nelements <= 1) {
if (0 == nelements) {
cif -> flags = FFI_TYPE_VOID;
} else {
cif -> flags = element_type;
}
} else {
switch(element_type) {
case FFI_TYPE_FLOAT:
cif -> flags = FFI_IS_FLOAT_FP_AGGREGATE | nelements;
break;
case FFI_TYPE_DOUBLE:
cif -> flags = FFI_IS_DOUBLE_FP_AGGREGATE | nelements;
break;
default:
/* long double NYI */
abort();
}
}
break;
}
if (sz <= 32) {
if (sz <= 8) {
cif->flags = FFI_TYPE_INT;
} else if (sz <= 16) {
cif->flags = FFI_IS_SMALL_STRUCT2;
} else if (sz <= 24) {
cif->flags = FFI_IS_SMALL_STRUCT3;
} else {
cif->flags = FFI_IS_SMALL_STRUCT4;
}
} else {
cif->flags = FFI_TYPE_STRUCT;
}
}
break;
case FFI_TYPE_FLOAT:
is_simple = FALSE;
cif->flags = FFI_TYPE_FLOAT;
break;
case FFI_TYPE_DOUBLE:
is_simple = FALSE;
cif->flags = FFI_TYPE_DOUBLE;
break;
default:
cif->flags = FFI_TYPE_INT;
/* This seems to depend on little endian mode, and the fact that */
/* the return pointer always points to at least 8 bytes. But */
/* that also seems to be true for other platforms. */
break;
}
if (is_simple) cif -> flags |= simple_flag;
return FFI_OK;
}
extern int ffi_call_unix(bool (*)(struct ia64_args *, extended_cif *, int),
extended_cif *, unsigned,
unsigned, unsigned *, void (*)());
void
ffi_call(ffi_cif *cif, void (*fn)(), void *rvalue, void **avalue)
{
extended_cif ecif;
long simple = cif -> flags & FFI_SIMPLE;
/* Should this also check for Unix ABI? */
/* This is almost, but not quite, machine independent. Note that */
/* we can get away with not caring about length of the result because */
/* we assume we are little endian, and the result buffer is large */
/* enough. */
/* This needs work for HP/UX. */
if (simple) {
long (*lfn)() = (long (*)())fn;
long result;
switch(simple) {
case FFI_SIMPLE_V:
result = lfn();
break;
case FFI_SIMPLE_I:
result = lfn(*(int *)avalue[0]);
break;
case FFI_SIMPLE_L:
result = lfn(*(long *)avalue[0]);
break;
case FFI_SIMPLE_II:
result = lfn(*(int *)avalue[0], *(int *)avalue[1]);
break;
case FFI_SIMPLE_IL:
result = lfn(*(int *)avalue[0], *(long *)avalue[1]);
break;
case FFI_SIMPLE_LI:
result = lfn(*(long *)avalue[0], *(int *)avalue[1]);
break;
case FFI_SIMPLE_LL:
result = lfn(*(long *)avalue[0], *(long *)avalue[1]);
break;
}
if ((cif->flags & ~FFI_SIMPLE) != FFI_TYPE_VOID && 0 != rvalue) {
* (long *)rvalue = result;
}
return;
}
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return
value address then we need to make one. */
if (rvalue == NULL && cif->rtype->type == FFI_TYPE_STRUCT)
ecif.rvalue = alloca(cif->rtype->size);
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_UNIX:
ffi_call_unix(ffi_prep_args, &ecif, cif->bytes,
cif->flags, rvalue, fn);
break;
default:
FFI_ASSERT(0);
break;
}
}
/*
* Closures represent a pair consisting of a function pointer, and
* some user data. A closure is invoked by reinterpreting the closure
* as a function pointer, and branching to it. Thus we can make an
* interpreted function callable as a C function: We turn the interpreter
* itself, together with a pointer specifying the interpreted procedure,
* into a closure.
* On X86, the first few words of the closure structure actually contain code,
* which will do the right thing. On most other architectures, this
* would raise some Icache/Dcache coherence issues (which can be solved, but
* often not cheaply).
* For IA64, function pointer are already pairs consisting of a code
* pointer, and a gp pointer. The latter is needed to access global variables.
* Here we set up such a pair as the first two words of the closure (in
* the "trampoline" area), but we replace the gp pointer with a pointer
* to the closure itself. We also add the real gp pointer to the
* closure. This allows the function entry code to both retrieve the
* user data, and to restire the correct gp pointer.
*/
static void
ffi_prep_incoming_args_UNIX(struct ia64_args *args, void **rvalue,
void **avalue, ffi_cif *cif);
/* This function is entered with the doctored gp (r1) value.
* This code is extremely gcc specific. There is some argument that
* it should really be written in assembly code, since it depends on
* gcc properties that might change over time.
*/
/* ffi_closure_UNIX is an assembly routine, which copies the register */
/* state into s struct ia64_args, and the invokes */
/* ffi_closure_UNIX_inner. It also recovers the closure pointer */
/* from its fake gp pointer. */
void ffi_closure_UNIX();
#ifndef __GNUC__
# error This requires gcc
#endif
void
ffi_closure_UNIX_inner (ffi_closure *closure, struct ia64_args * args)
/* Hopefully declarint this as a varargs function will force all args */
/* to memory. */
{
// this is our return value storage
long double res;
// our various things...
ffi_cif *cif;
unsigned short rtype;
void *resp;
void **arg_area;
resp = (void*)&res;
cif = closure->cif;
arg_area = (void**) alloca (cif->nargs * sizeof (void*));
/* this call will initialize ARG_AREA, such that each
* element in that array points to the corresponding
* value on the stack; and if the function returns
* a structure, it will re-set RESP to point to the
* structure return address. */
ffi_prep_incoming_args_UNIX(args, (void**)&resp, arg_area, cif);
(closure->fun) (cif, resp, arg_area, closure->user_data);
rtype = cif->flags;
/* now, do a generic return based on the value of rtype */
if (rtype == FFI_TYPE_INT)
{
asm volatile ("ld8 r8=[%0]" : : "r" (resp) : "r8");
}
else if (rtype == FFI_TYPE_FLOAT)
{
asm volatile ("ldfs f8=[%0]" : : "r" (resp) : "f8");
}
else if (rtype == FFI_TYPE_DOUBLE)
{
asm volatile ("ldfd f8=[%0]" : : "r" (resp) : "f8");
}
else if (rtype == FFI_IS_SMALL_STRUCT2)
{
asm volatile ("ld8 r8=[%0]; ld8 r9=[%1]"
: : "r" (resp), "r" (resp+8) : "r8","r9");
}
else if (rtype == FFI_IS_SMALL_STRUCT3)
{
asm volatile ("ld8 r8=[%0]; ld8 r9=[%1]; ld8 r10=[%2]"
: : "r" (resp), "r" (resp+8), "r" (resp+16)
: "r8","r9","r10");
}
else if (rtype == FFI_IS_SMALL_STRUCT4)
{
asm volatile ("ld8 r8=[%0]; ld8 r9=[%1]; ld8 r10=[%2]; ld8 r11=[%3]"
: : "r" (resp), "r" (resp+8), "r" (resp+16), "r" (resp+24)
: "r8","r9","r10","r11");
}
else if (rtype != FFI_TYPE_VOID && rtype != FFI_TYPE_STRUCT)
{
/* Can only happen for homogeneous FP aggregates? */
abort();
}
}
static void
ffi_prep_incoming_args_UNIX(struct ia64_args *args, void **rvalue,
void **avalue, ffi_cif *cif)
{
register unsigned int i;
register unsigned int avn;
register void **p_argv;
register unsigned long *argp = args -> out_regs;
unsigned fp_reg_num = 0;
register ffi_type **p_arg;
avn = cif->nargs;
p_argv = avalue;
for (i = cif->nargs, p_arg = cif->arg_types; i != 0; i--, p_arg++)
{
size_t z; /* In units of words or argument slots. */
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
z = 1;
*p_argv = (void *)argp;
break;
case FFI_TYPE_FLOAT:
z = 1;
/* Convert argument back to float in place from the saved value */
if (fp_reg_num < 8) {
*(float *)argp = args -> fp_regs[fp_reg_num++];
} else {
*(float *)argp = *(double *)argp;
}
*p_argv = (void *)argp;
break;
case FFI_TYPE_DOUBLE:
z = 1;
if (fp_reg_num < 8) {
*p_argv = args -> fp_regs + fp_reg_num++;
} else {
*p_argv = (void *)argp;
}
break;
case FFI_TYPE_STRUCT:
{
size_t sz = (*p_arg)->size;
unsigned short element_type;
z = ((*p_arg)->size + SIZEOF_ARG - 1)/SIZEOF_ARG;
if (is_homogeneous_fp_aggregate(*p_arg, 8, &element_type)) {
int nelements = sz/float_type_size(element_type);
if (nelements + fp_reg_num >= 8) {
/* hard case NYI. */
abort();
}
if (element_type == FFI_TYPE_DOUBLE) {
*p_argv = args -> fp_regs + fp_reg_num;
fp_reg_num += nelements;
break;
}
if (element_type == FFI_TYPE_FLOAT) {
int j;
for (j = 0; j < nelements; ++ j) {
((float *)argp)[j] = args -> fp_regs[fp_reg_num + j];
}
*p_argv = (void *)argp;
fp_reg_num += nelements;
break;
}
abort(); /* Other fp types NYI */
}
}
break;
default:
FFI_ASSERT(0);
}
argp += z;
p_argv++;
}
return;
}
/* Fill in a closure to refer to the specified fun and user_data. */
/* cif specifies the argument and result types for fun. */
/* the cif must already be prep'ed */
/* The layout of a function descriptor. A C function pointer really */
/* points to one of these. */
typedef struct ia64_fd_struct {
void *code_pointer;
void *gp;
} ia64_fd;
ffi_status
ffi_prep_closure (ffi_closure* closure,
ffi_cif* cif,
void (*fun)(ffi_cif*,void*,void**,void*),
void *user_data)
{
struct ffi_ia64_trampoline_struct *tramp =
(struct ffi_ia64_trampoline_struct *) (closure -> tramp);
ia64_fd *fd = (ia64_fd *)(void *)ffi_closure_UNIX;
FFI_ASSERT (cif->abi == FFI_UNIX);
tramp -> code_pointer = fd -> code_pointer;
tramp -> real_gp = fd -> gp;
tramp -> fake_gp = closure;
closure->cif = cif;
closure->user_data = user_data;
closure->fun = fun;
return FFI_OK;
}
+62
View File
@@ -0,0 +1,62 @@
/* -----------------------------------------------------------------------
ia64_flags.h - Copyright (c) 2000 Hewlett Packard Company
IA64/unix Foreign Function Interface
Original author: Hans Boehm, HP Labs
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
/* Homogeneous Floating Point Aggregates (HFAs) which are returned */
/* in FP registers. The least significant bits specify the size in */
/* words. */
#define FFI_IS_FLOAT_FP_AGGREGATE 0x1000
#define FFI_IS_DOUBLE_FP_AGGREGATE 0x0800
#define FLOAT_FP_AGGREGATE_BIT 12
#define DOUBLE_FP_AGGREGATE_BIT 11
/* Small structures containing N words. If N=1, they are returned */
/* as though they were integers. */
#define FFI_IS_SMALL_STRUCT2 0x40 /* Struct > 8, <=16 bytes */
#define FFI_IS_SMALL_STRUCT3 0x41 /* Struct > 16 <= 24 bytes */
#define FFI_IS_SMALL_STRUCT4 0x42 /* Struct > 24, <=32 bytes */
/* Flag values identifying particularly simple cases, which are */
/* handled specially. We treat functions as simple if they take all */
/* arguments can be passed as 32 or 64 bit integer quantities, there is */
/* either no return value or it can be treated as a 64bit integer, and */
/* if there are at most 2 arguments. */
/* This is OR'ed with the normal flag values. */
#define FFI_SIMPLE_V 0x10000 /* () -> X */
#define FFI_SIMPLE_I 0x20000 /* (int) -> X */
#define FFI_SIMPLE_L 0x30000 /* (long) -> X */
#define FFI_SIMPLE_II 0x40000 /* (int,int) -> X */
#define FFI_SIMPLE_IL 0x50000 /* (int,long) -> X */
#define FFI_SIMPLE_LI 0x60000 /* (long,int) -> X */
#define FFI_SIMPLE_LL 0x70000 /* (long,long) -> X */
/* Mask for all of the FFI_SIMPLE bits: */
#define FFI_SIMPLE 0xf0000
/* An easy way to build FFI_SIMPLE flags from FFI_SIMPLE_V: */
#define FFI_ADD_LONG_ARG(flag) (((flag) << 1) | 0x10000)
#define FFI_ADD_INT_ARG(flag) ((flag) << 1)
+301
View File
@@ -0,0 +1,301 @@
/* -----------------------------------------------------------------------
unix.S - Copyright (c) 1998 Cygnus Solutions
Copyright (c) 2000 Hewlett Packard Company
IA64/unix Foreign Function Interface
Primary author: Hans Boehm, HP Labs
Loosely modeled on Cygnus code for other platforms.
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include "ia64_flags.h"
/* parameters: */
#define callback in0
#define ecifp in1
#define bytes in2
#define flags in3
#define raddr in4
#define fn in5
#define FLOAT_SZ 8 /* in-memory size of fp operands */
.text
.align 16
.global ffi_call_unix#
.proc ffi_call_unix#
ffi_call_unix:
alloc loc0=ar.pfs,6,5,8,0
mov loc1=b0;
sub sp=sp,bytes
mov loc4=r1 /* Save gp */
ld8 r8=[callback],8 /* code address of callback */
;;
mov out0=sp
mov out1=ecifp
mov out2=bytes
ld8 r1=[callback] /* Set up gp for callback. Unnecessary? */
mov b6=r8
;;
br.call.sptk.many b0 = b6 /* call ffi_prep_args */
cmp.eq p6,p0=0,r8 /* r8 nonzero ==> need fp regs */
;;
(p6) add loc2=32+8*FLOAT_SZ,sp
(p6) br.cond.dptk.many fp_done
;; /* Quiets warning; needed? */
add loc2=32,sp
add loc3=32+FLOAT_SZ,sp
;;
ldfd f8=[loc2],2*FLOAT_SZ
ldfd f9=[loc3],2*FLOAT_SZ
;;
ldfd f10=[loc2],2*FLOAT_SZ
ldfd f11=[loc3],2*FLOAT_SZ
;;
ldfd f12=[loc2],2*FLOAT_SZ
ldfd f13=[loc3],2*FLOAT_SZ
;;
ldfd f14=[loc2],2*FLOAT_SZ
ldfd f15=[loc3]
fp_done:
add r9=16,sp /* Pointer to r8_contents */
/* loc2 points at first integer register value. */
add loc3=8,loc2
;;
ld8 r8=[r9] /* Just in case we return large struct */
ld8 out0=[loc2],16
ld8 out1=[loc3],16
;;
ld8 out2=[loc2],16
ld8 out3=[loc3],16
;;
ld8 out4=[loc2],16
ld8 out5=[loc3],16
;;
ld8 out6=[loc2],16
ld8 out7=[loc3]
/* loc2 points at first stack parameter. Set sp to 16 bytes */
/* below that. */
add sp=-16,loc2
ld8 r8=[fn],8
;;
ld8 r1=[fn] /* Set up gp */
mov b6=r8;;
br.call.sptk.many b0 = b6 /* call ffi_prep_args */
/* Handle return value. */
cmp.eq p6,p0=0,raddr
cmp.eq p7,p0=FFI_TYPE_INT,flags
cmp.eq p10,p0=FFI_IS_SMALL_STRUCT2,flags
cmp.eq p11,p0=FFI_IS_SMALL_STRUCT3,flags
cmp.eq p12,p0=FFI_IS_SMALL_STRUCT4,flags
;;
(p6) br.cond.dpnt.few done /* Dont copy ret values if raddr = 0 */
(p7) br.cond.dptk.few copy1
(p10) br.cond.dpnt.few copy2
(p11) br.cond.dpnt.few copy3
(p12) br.cond.dpnt.few copy4
cmp.eq p8,p0=FFI_TYPE_FLOAT,flags
cmp.eq p9,p0=FFI_TYPE_DOUBLE,flags
tbit.nz p6,p0=flags,FLOAT_FP_AGGREGATE_BIT
tbit.nz p7,p0=flags,DOUBLE_FP_AGGREGATE_BIT
;;
(p8) stfs [raddr]=f8
(p9) stfd [raddr]=f8
;;
(p6) br.cond.dpnt.few handle_float_hfa
(p7) br.cond.dpnt.few handle_double_hfa
br done
copy4:
add loc3=24,raddr
;;
st8 [loc3]=r11
copy3:
add loc3=16,raddr
;;
st8 [loc3]=r10
copy2:
add loc3=8,raddr
;;
st8 [loc3]=r9
copy1:
st8 [raddr]=r8
/* In the big struct case, raddr was passed as an argument. */
/* In the void case there was nothing to do. */
done:
mov r1=loc4 /* Restore gp */
mov ar.pfs = loc0
mov b0 = loc1
br.ret.sptk.many b0
handle_double_hfa:
/* Homogeneous floating point array of doubles is returned in */
/* registers f8-f15. Save one at a time to return area. */
and flags=0xf,flags /* Retrieve size */
;;
cmp.eq p6,p0=2,flags
cmp.eq p7,p0=3,flags
cmp.eq p8,p0=4,flags
cmp.eq p9,p0=5,flags
cmp.eq p10,p0=6,flags
cmp.eq p11,p0=7,flags
cmp.eq p12,p0=8,flags
;;
(p6) br.cond.dptk.few dhfa2
(p7) br.cond.dptk.few dhfa3
(p8) br.cond.dptk.few dhfa4
(p9) br.cond.dptk.few dhfa5
(p10) br.cond.dptk.few dhfa6
(p11) br.cond.dptk.few dhfa7
dhfa8: add loc3=7*8,raddr
;;
stfd [loc3]=f15
dhfa7: add loc3=6*8,raddr
;;
stfd [loc3]=f14
dhfa6: add loc3=5*8,raddr
;;
stfd [loc3]=f13
dhfa5: add loc3=4*8,raddr
;;
stfd [loc3]=f12
dhfa4: add loc3=3*8,raddr
;;
stfd [loc3]=f11
dhfa3: add loc3=2*8,raddr
;;
stfd [loc3]=f10
dhfa2: add loc3=1*8,raddr
;;
stfd [loc3]=f9
stfd [raddr]=f8
br done
handle_float_hfa:
/* Homogeneous floating point array of floats is returned in */
/* registers f8-f15. Save one at a time to return area. */
and flags=0xf,flags /* Retrieve size */
;;
cmp.eq p6,p0=2,flags
cmp.eq p7,p0=3,flags
cmp.eq p8,p0=4,flags
cmp.eq p9,p0=5,flags
cmp.eq p10,p0=6,flags
cmp.eq p11,p0=7,flags
cmp.eq p12,p0=8,flags
;;
(p6) br.cond.dptk.few shfa2
(p7) br.cond.dptk.few shfa3
(p8) br.cond.dptk.few shfa4
(p9) br.cond.dptk.few shfa5
(p10) br.cond.dptk.few shfa6
(p11) br.cond.dptk.few shfa7
shfa8: add loc3=7*4,raddr
;;
stfd [loc3]=f15
shfa7: add loc3=6*4,raddr
;;
stfd [loc3]=f14
shfa6: add loc3=5*4,raddr
;;
stfd [loc3]=f13
shfa5: add loc3=4*4,raddr
;;
stfd [loc3]=f12
shfa4: add loc3=3*4,raddr
;;
stfd [loc3]=f11
shfa3: add loc3=2*4,raddr
;;
stfd [loc3]=f10
shfa2: add loc3=1*4,raddr
;;
stfd [loc3]=f9
stfd [raddr]=f8
br done
.endp ffi_call_unix
.text
.align 16
.global ffi_closure_UNIX
.proc ffi_closure_UNIX
ffi_closure_UNIX:
alloc loc0=ar.pfs,8,2,2,0
mov loc1=b0
/* Retrieve closure pointer and real gp. */
mov out0=gp
add gp=16,gp
;;
ld8 gp=[gp]
/* Reserve a structia64_args on the stack such that arguments */
/* past the first 8 are automatically placed in the right */
/* slot. Note that when we start the sp points at 2 8-byte */
/* scratch words, followed by the extra arguments. */
# define BASIC_ARGS_SZ (8*FLOAT_SZ+8*8+2*8)
# define FIRST_FP_OFFSET (4*8)
add r14=-(BASIC_ARGS_SZ-FIRST_FP_OFFSET),sp
add r15=-(BASIC_ARGS_SZ-FIRST_FP_OFFSET-FLOAT_SZ),sp
add sp=-BASIC_ARGS_SZ,sp
/* r14 points to fp_regs[0], r15 points to fp_regs[1] */
;;
stfd [r14]=f8,2*FLOAT_SZ
stfd [r15]=f9,2*FLOAT_SZ
;;
stfd [r14]=f10,2*FLOAT_SZ
stfd [r15]=f11,2*FLOAT_SZ
;;
stfd [r14]=f12,2*FLOAT_SZ
stfd [r15]=f13,2*FLOAT_SZ
;;
stfd [r14]=f14,FLOAT_SZ+8
stfd [r15]=f15,2*8
;;
/* r14 points to first parameter register area, r15 to second. */
st8 [r14]=in0,2*8
st8 [r15]=in1,2*8
;;
st8 [r14]=in2,2*8
st8 [r15]=in3,2*8
;;
st8 [r14]=in4,2*8
st8 [r15]=in5,2*8
;;
st8 [r14]=in6,2*8
st8 [r15]=in7,2*8
/* Call ffi_closure_UNIX_inner */
mov out1=sp
br.call.sptk.many b0=ffi_closure_UNIX_inner
;;
mov b0=loc1
mov ar.pfs=loc0
br.ret.sptk.many b0
.endp ffi_closure_UNIX
@@ -1,4 +1,3 @@
.deps
.libs
fficonfig.h
Makefile
Makefile.in
+6
View File
@@ -0,0 +1,6 @@
## Process this with automake to create Makefile.in
AUTOMAKE_OPTIONS = foreign
noinst_HEADERS=ffi_private.h
include_HEADERS=ffi.h
+393
View File
@@ -0,0 +1,393 @@
/* -----------------------------------------------------------------*-C-*-
libffi 2.0.0 - Copyright (C) 1996, 1997, 1998, 1999, 2000,
2001 Red Hat, Inc.
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 RED HAT 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.
----------------------------------------------------------------------- */
/* -------------------------------------------------------------------
The basic API is described in the README file.
The raw API is designed to bypass some of the argument packing
and unpacking on architectures for which it can be avoided.
The closure API allows interpreted functions to be packaged up
inside a C function pointer, so that they can be called as C functions,
with no understanding on the client side that they are interpreted.
It can also be used in other cases in which it is necessary to package
up a user specified parameter and a function pointer as a single
function pointer.
The closure API must be implemented in order to get its functionality,
e.g. for use by gij. Routines are provided to emulate the raw API
if the underlying platform doesn't allow faster implementation.
More details on the raw and cloure API can be found in:
http://gcc.gnu.org/ml/java/1999-q3/msg00138.html
and
http://gcc.gnu.org/ml/java/1999-q3/msg00174.html
-------------------------------------------------------------------- */
#ifndef FFI_H
#define FFI_H
#if !defined(__ASSEMBLER__) && !defined(__GNUC__)
#error --- ffi.h requires GNU C ---
#endif
#ifdef __cplusplus
extern "C" {
#endif
#if !defined(LIBFFI_ASM)
#include <stddef.h>
#if defined(FFI_DEBUG)
#include <stdio.h>
#endif
#endif
#ifndef LIBFFI_ASM
/* ---- Generic type definitions ----------------------------------------- */
typedef enum ffi_abi {
/* Leave this for debugging purposes */
FFI_FIRST_ABI = 0,
/* ---- Sparc -------------------- */
#ifdef __sparc__
FFI_V8,
FFI_V8PLUS,
FFI_V9,
#if defined(__arch64__) || defined(__sparcv9)
FFI_DEFAULT_ABI = FFI_V9,
#else
FFI_DEFAULT_ABI = FFI_V8,
#endif
#endif
/* ---- Intel x86 ---------------- */
#ifdef __i386__
FFI_SYSV,
FFI_DEFAULT_ABI = FFI_SYSV,
#endif
/* ---- Intel ia64 ---------------- */
#ifdef __ia64__
FFI_UNIX, /* Linux and all Unix variants use the same conventions */
FFI_DEFAULT_ABI = FFI_UNIX,
#endif
/* ---- Mips --------------------- */
#ifdef __mips__
FFI_O32,
FFI_N32,
FFI_N64,
# if defined(__mips_eabi)
# define FFI_MIPS_EABI
# define FFI_MIPS_O32
# else
# if !defined(_MIPS_SIM)
#error -- something is very wrong --
# else
# if _MIPS_SIM==_ABIN32 && defined(_ABIN32)
# define FFI_MIPS_N32
# else
# define FFI_MIPS_O32
# endif
# endif
# endif
# if defined(FFI_MIPS_O32)
FFI_DEFAULT_ABI = FFI_O32,
# else
FFI_DEFAULT_ABI = FFI_N32,
# endif
#endif
/* ---- Alpha -------------------- */
#ifdef __alpha__
FFI_OSF,
FFI_DEFAULT_ABI = FFI_OSF,
#endif
/* ---- Motorola m68k ------------ */
#ifdef __m68k__
FFI_SYSV,
FFI_DEFAULT_ABI = FFI_SYSV,
#endif
/* ---- PowerPC ------------------ */
#ifdef __powerpc__
FFI_SYSV,
FFI_GCC_SYSV,
FFI_DEFAULT_ABI = FFI_GCC_SYSV,
#endif
/* ---- ARM --------------------- */
#ifdef __arm__
FFI_SYSV,
FFI_DEFAULT_ABI = FFI_SYSV,
#endif
/* ---- S390 --------------------- */
#ifdef __S390__
FFI_SYSV,
FFI_DEFAULT_ABI = FFI_SYSV,
#endif
/* Leave this for debugging purposes */
FFI_LAST_ABI
} ffi_abi;
typedef struct _ffi_type
{
size_t size;
unsigned short alignment;
unsigned short type;
struct _ffi_type **elements;
} ffi_type;
/* These are defined in ffi.c */
extern ffi_type ffi_type_void;
extern ffi_type ffi_type_uint8;
extern ffi_type ffi_type_sint8;
extern ffi_type ffi_type_uint16;
extern ffi_type ffi_type_sint16;
extern ffi_type ffi_type_uint32;
extern ffi_type ffi_type_sint32;
extern ffi_type ffi_type_uint64;
extern ffi_type ffi_type_sint64;
extern ffi_type ffi_type_float;
extern ffi_type ffi_type_double;
extern ffi_type ffi_type_longdouble;
extern ffi_type ffi_type_pointer;
extern ffi_type ffi_type_ushort;
extern ffi_type ffi_type_sint;
extern ffi_type ffi_type_uint;
extern ffi_type ffi_type_slong;
extern ffi_type ffi_type_ulong;
typedef enum {
FFI_OK = 0,
FFI_BAD_TYPEDEF,
FFI_BAD_ABI
} ffi_status;
typedef unsigned FFI_TYPE;
typedef struct {
ffi_abi abi;
unsigned nargs;
ffi_type **arg_types;
ffi_type *rtype;
unsigned bytes;
unsigned flags;
#ifdef __mips__
#if _MIPS_SIM == _ABIN32
unsigned rstruct_flag;
#endif
#endif
} ffi_cif;
/* ---- Definitions for the raw API -------------------------------------- */
#if !FFI_NO_RAW_API
typedef union {
#if _MIPS_SIM==_ABIN32 && defined(_ABIN32)
long sint;
unsigned long uint;
#else
int sint;
unsigned uint;
#endif
float flt;
char data[sizeof(void*)];
void* ptr;
} ffi_raw;
void ffi_raw_call (/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ ffi_raw *avalue);
void ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw);
void ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args);
size_t ffi_raw_size (ffi_cif *cif);
#if !NO_JAVA_RAW_API
/* This is analogous to the raw API, except it uses Java parameter */
/* packing, even on 64-bit machines. I.e. on 64-bit machines */
/* longs and doubles are followed by an empty 64-bit word. */
void ffi_java_raw_call (/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ ffi_raw *avalue);
void ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw);
void ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args);
size_t ffi_java_raw_size (ffi_cif *cif);
#endif /* !NO_JAVA_RAW_API */
#endif /* !FFI_NO_RAW_API */
/* ---- Definitions for closures ----------------------------------------- */
#ifdef __i386__
#define FFI_CLOSURES 1 /* x86 supports closures */
#define FFI_TRAMPOLINE_SIZE 10
#define FFI_NATIVE_RAW_API 1 /* and has native raw api support */
#elif defined(X86_WIN32)
#define FFI_CLOSURES 1 /* x86 supports closures */
#define FFI_TRAMPOLINE_SIZE 10
#define FFI_NATIVE_RAW_API 1 /* and has native raw api support */
#elif defined(IA64)
#define FFI_CLOSURES 1
#define FFI_TRAMPOLINE_SIZE 24 /* Really the following struct, which */
/* can be interpreted as a C function */
/* decriptor: */
struct ffi_ia64_trampoline_struct {
void * code_pointer; /* Pointer to ffi_closure_UNIX */
void * fake_gp; /* Pointer to closure, installed as gp */
void * real_gp; /* Real gp value, reinstalled by */
/* ffi_closure_UNIX. */
};
#define FFI_NATIVE_RAW_API 0
#elif defined(__alpha__)
#define FFI_CLOSURES 1
#define FFI_TRAMPOLINE_SIZE 24
#define FFI_NATIVE_RAW_API 0
#elif defined(POWERPC)
#define FFI_CLOSURES 1
#define FFI_TRAMPOLINE_SIZE 40
#define FFI_NATIVE_RAW_API 0
#else
#define FFI_CLOSURES 0
#define FFI_NATIVE_RAW_API 0
#endif
#if FFI_CLOSURES
typedef struct {
char tramp[FFI_TRAMPOLINE_SIZE];
ffi_cif *cif;
void (*fun)(ffi_cif*,void*,void**,void*);
void *user_data;
} ffi_closure;
ffi_status
ffi_prep_closure (ffi_closure*,
ffi_cif *,
void (*fun)(ffi_cif*,void*,void**,void*),
void *user_data);
#if !FFI_NO_RAW_API
typedef struct {
char tramp[FFI_TRAMPOLINE_SIZE];
ffi_cif *cif;
#if !FFI_NATIVE_RAW_API
/* if this is enabled, then a raw closure has the same layout
as a regular closure. We use this to install an intermediate
handler to do the transaltion, void** -> ffi_raw*. */
void (*translate_args)(ffi_cif*,void*,void**,void*);
void *this_closure;
#endif
void (*fun)(ffi_cif*,void*,ffi_raw*,void*);
void *user_data;
} ffi_raw_closure;
ffi_status
ffi_prep_raw_closure (ffi_raw_closure*,
ffi_cif *cif,
void (*fun)(ffi_cif*,void*,ffi_raw*,void*),
void *user_data);
#ifndef NO_JAVA_RAW_API
ffi_status
ffi_prep_java_raw_closure (ffi_raw_closure*,
ffi_cif *cif,
void (*fun)(ffi_cif*,void*,ffi_raw*,void*),
void *user_data);
#endif
#endif /* !FFI_NO_RAW_API */
#endif /* FFI_CLOSURES */
/* ---- Public interface definition -------------------------------------- */
ffi_status ffi_prep_cif (ffi_cif *cif,
ffi_abi abi,
unsigned int nargs,
ffi_type *rtype,
ffi_type **atypes);
void ffi_call (ffi_cif *cif,
void (*fn)(),
void *rvalue,
void **avalue);
/* Useful for eliminating compiler warnings */
#define FFI_FN(f) ((void (*)())f)
#endif
#ifdef __cplusplus
}
#endif
#endif
+168
View File
@@ -0,0 +1,168 @@
/* -----------------------------------------------------------------*-C-*-
libffi 2.0.0 - Copyright (C) 1996, 1997, 1998, 1999, 2000,
2001 Red Hat, Inc.
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 RED HAT 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 <fficonfig.h>
#define ALIGN(v, a) (((((size_t) (v))-1) | ((a)-1))+1)
/* ---- Generic type definitions ----------------------------------------- */
#define FLOAT32 float
#define FLOAT64 double
#define FLOAT80 long double
#define UINT8 unsigned char
#define SINT8 signed char
#if SIZEOF_INT == 2
#define UINT16 unsigned int
#define SINT16 int
#else
#if SIZEOF_SHORT == 2
#define UINT16 unsigned short
#define SINT16 short
#endif
#endif
#if SIZEOF_INT == 4
#define UINT32 unsigned int
#define SINT32 int
#else
#if SIZEOF_SHORT == 4
#define UINT32 unsigned short
#define SINT32 short
#else
#if SIZEOF_LONG == 4
#define UINT32 unsigned long
#define SINT32 long
#endif
#endif
#endif
#if SIZEOF_INT == 8
#define UINT64 unsigned int
#define SINT64 int
#else
#if SIZEOF_LONG == 8
#define UINT64 unsigned long
#define SINT64 long
#else
#if SIZEOF_LONG_LONG == 8
#define UINT64 unsigned long long
#define SINT64 long long
#endif
#endif
#endif
#define FFI_TYPE_VOID 0
#define FFI_TYPE_INT 1
#define FFI_TYPE_FLOAT 2
#define FFI_TYPE_DOUBLE 3
#if SIZEOF_LONG_DOUBLE == SIZEOF_DOUBLE
#define FFI_TYPE_LONGDOUBLE FFI_TYPE_DOUBLE
#else
#define FFI_TYPE_LONGDOUBLE 4
#endif
#define FFI_TYPE_UINT8 5 /* If this changes, update ffi_mips.h. */
#define FFI_TYPE_SINT8 6 /* If this changes, update ffi_mips.h. */
#define FFI_TYPE_UINT16 7
#define FFI_TYPE_SINT16 8
#define FFI_TYPE_UINT32 9
#define FFI_TYPE_SINT32 10
#define FFI_TYPE_UINT64 11
#define FFI_TYPE_SINT64 12
#define FFI_TYPE_STRUCT 13 /* If this changes, update ffi_mips.h. */
#define FFI_TYPE_POINTER 14
#define FFI_TYPE_LAST 14
#if _MIPS_SIM==_ABIN32 && defined(_ABIN32)
#define SIZEOF_ARG 8
#else
#define SIZEOF_ARG SIZEOF_VOID_P
#endif
#ifndef __ASSEMBLER__
/* This part of the private header file is only for C code. */
/* Check for the existence of memcpy. */
#if STDC_HEADERS
# include <string.h>
#else
# ifndef HAVE_MEMCPY
# define memcpy(d, s, n) bcopy ((s), (d), (n))
# endif
#endif
#ifndef FALSE
#define FALSE 0
#endif
#ifndef TRUE
#define TRUE (!FALSE)
#endif
#ifndef __cplusplus
/* bool is a keyword in C++ */
typedef int bool;
#endif
#ifdef FFI_DEBUG
/* Debugging functions */
void ffi_stop_here(void);
bool ffi_type_test(ffi_type *a);
#define FFI_ASSERT(x) ((x) ? 0 : ffi_assert(__FILE__,__LINE__))
#else
#define FFI_ASSERT(x)
#endif
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif);
/* Extended cif, used in callback from assembly routine */
typedef struct
{
ffi_cif *cif;
void *rvalue;
void **avalue;
} extended_cif;
#endif /* __ASSEMBLER__ */
+269
View File
@@ -0,0 +1,269 @@
/* -----------------------------------------------------------------------
java_raw_api.c - Copyright (c) 1999 Cygnus Solutions
Cloned from raw_api.c
Raw_api.c author: Kresten Krab Thorup <[email protected]>
Java_raw_api.c author: Hans-J. Boehm <[email protected]>
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 RED HAT 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.
----------------------------------------------------------------------- */
/* This defines a Java- and 64-bit specific variant of the raw API. */
/* It assumes that "raw" argument blocks look like Java stacks on a */
/* 64-bit machine. Arguments that can be stored in a single stack */
/* stack slots (longs, doubles) occupy 128 bits, but only the first */
/* 64 bits are actually used. */
#include <ffi.h>
#include <ffi_private.h>
#if !defined(NO_JAVA_RAW_API) && !defined(FFI_NO_RAW_API)
size_t
ffi_java_raw_size (ffi_cif *cif)
{
size_t result = 0;
int i;
ffi_type **at = cif->arg_types;
for (i = cif->nargs-1; i >= 0; i--, at++)
{
switch((*at) -> type) {
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
result += 2 * SIZEOF_ARG;
break;
case FFI_TYPE_STRUCT:
/* No structure parameters in Java. */
abort();
default:
result += SIZEOF_ARG;
}
}
return result;
}
void
ffi_java_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args)
{
unsigned i;
ffi_type **tp = cif->arg_types;
#if WORDS_BIGENDIAN
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
switch ((*tp)->type)
{
case FFI_TYPE_UINT8:
case FFI_TYPE_SINT8:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 1);
break;
case FFI_TYPE_UINT16:
case FFI_TYPE_SINT16:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 2);
break;
#if SIZEOF_ARG >= 4
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT32:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 4);
break;
#endif
#if SIZEOF_ARG == 8
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
case FFI_TYPE_DOUBLE:
*args = (void *)raw;
raw += 2;
break;
#endif
case FFI_TYPE_POINTER:
*args = (void*) &(raw++)->ptr;
break;
default:
*args = raw;
raw += ALIGN ((*tp)->size, SIZEOF_ARG) / SIZEOF_ARG;
}
}
#else /* WORDS_BIGENDIAN */
#if !PDP
/* then assume little endian */
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
#if SIZEOF_ARG == 8
switch((*tp)->type) {
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
case FFI_TYPE_DOUBLE:
*args = (void*) raw;
raw += 2;
break;
default:
*args = (void*) raw++;
}
#else /* SIZEOF_ARG != 8 */
*args = (void*) raw;
raw += ALIGN ((*tp)->size, sizeof (void*)) / sizeof (void*);
#endif /* SIZEOF_ARG == 8 */
}
#else
#error "pdp endian not supported"
#endif /* ! PDP */
#endif /* WORDS_BIGENDIAN */
}
void
ffi_java_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw)
{
unsigned i;
ffi_type **tp = cif->arg_types;
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
switch ((*tp)->type)
{
case FFI_TYPE_UINT8:
(raw++)->uint = *(UINT8*) (*args);
break;
case FFI_TYPE_SINT8:
(raw++)->sint = *(SINT8*) (*args);
break;
case FFI_TYPE_UINT16:
(raw++)->uint = *(UINT16*) (*args);
break;
case FFI_TYPE_SINT16:
(raw++)->sint = *(SINT16*) (*args);
break;
#if SIZEOF_ARG >= 4
case FFI_TYPE_UINT32:
(raw++)->uint = *(UINT32*) (*args);
break;
case FFI_TYPE_SINT32:
(raw++)->sint = *(SINT32*) (*args);
break;
#endif
case FFI_TYPE_FLOAT:
(raw++)->flt = *(FLOAT32*) (*args);
break;
#if SIZEOF_ARG == 8
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
case FFI_TYPE_DOUBLE:
raw->uint = *(UINT64*) (*args);
raw += 2;
break;
#endif
case FFI_TYPE_POINTER:
(raw++)->ptr = **(void***) args;
break;
default:
#if SIZEOF_ARG == 8
FFI_ASSERT(FALSE); /* Should have covered all cases */
#else
memcpy ((void*) raw->data, (void*)*args, (*tp)->size);
raw += ALIGN ((*tp)->size, SIZEOF_ARG) / SIZEOF_ARG;
#endif
}
}
}
#if !FFI_NATIVE_RAW_API
/* This is a generic definition of ffi_raw_call, to be used if the
* native system does not provide a machine-specific implementation.
* Having this, allows code to be written for the raw API, without
* the need for system-specific code to handle input in that format;
* these following couple of functions will handle the translation forth
* and back automatically. */
void ffi_java_raw_call (/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ ffi_raw *raw)
{
void **avalue = (void**) alloca (cif->nargs * sizeof (void*));
ffi_java_raw_to_ptrarray (cif, raw, avalue);
ffi_call (cif, fn, rvalue, avalue);
}
#if FFI_CLOSURES /* base system provides closures */
static void
ffi_java_translate_args (ffi_cif *cif, void *rvalue,
void **avalue, void *user_data)
{
ffi_raw *raw = (ffi_raw*)alloca (ffi_java_raw_size (cif));
ffi_raw_closure *cl = (ffi_raw_closure*)user_data;
ffi_java_ptrarray_to_raw (cif, avalue, raw);
(*cl->fun) (cif, rvalue, raw, cl->user_data);
}
/* Again, here is the generic version of ffi_prep_raw_closure, which
* will install an intermediate "hub" for translation of arguments from
* the pointer-array format, to the raw format */
ffi_status
ffi_prep_java_raw_closure (ffi_raw_closure* cl,
ffi_cif *cif,
void (*fun)(ffi_cif*,void*,ffi_raw*,void*),
void *user_data)
{
ffi_status status;
status = ffi_prep_closure ((ffi_closure*) cl,
cif,
&ffi_java_translate_args,
(void*)cl);
if (status == FFI_OK)
{
cl->fun = fun;
cl->user_data = user_data;
}
return status;
}
#endif /* FFI_CLOSURES */
#endif /* !FFI_NATIVE_RAW_API */
#endif /* !FFI_NO_RAW_API */
+184
View File
@@ -0,0 +1,184 @@
/* -----------------------------------------------------------------------
ffi.c
m68k Foreign Function Interface
----------------------------------------------------------------------- */
#include <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
/* ffi_prep_args is called by the assembly routine once stack space has
been allocated for the function's arguments. */
static void *
ffi_prep_args (void *stack, extended_cif *ecif)
{
unsigned int i;
int tmp;
unsigned int avn;
void **p_argv;
char *argp;
ffi_type **p_arg;
void *struct_value_ptr;
tmp = 0;
argp = stack;
if (ecif->cif->rtype->type == FFI_TYPE_STRUCT
&& ecif->cif->rtype->size > 8)
struct_value_ptr = ecif->rvalue;
else
struct_value_ptr = NULL;
avn = ecif->cif->nargs;
p_argv = ecif->avalue;
for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types;
i != 0 && avn != 0;
i--, p_arg++)
{
size_t z;
/* Align if necessary. */
if (((*p_arg)->alignment - 1) & (unsigned) argp)
argp = (char *) ALIGN (argp, (*p_arg)->alignment);
if (avn != 0)
{
avn--;
z = (*p_arg)->size;
if (z < sizeof (int))
{
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(signed int *) argp = (signed int) *(SINT8 *) *p_argv;
break;
case FFI_TYPE_UINT8:
*(unsigned int *) argp = (unsigned int) *(UINT8 *) *p_argv;
break;
case FFI_TYPE_SINT16:
*(signed int *) argp = (signed int) *(SINT16 *) *p_argv;
break;
case FFI_TYPE_UINT16:
*(unsigned int *) argp = (unsigned int) *(UINT16 *) *p_argv;
break;
case FFI_TYPE_STRUCT:
memcpy (argp + sizeof (int) - z, *p_argv, z);
break;
default:
FFI_ASSERT (0);
}
z = sizeof (int);
}
else
memcpy (argp, *p_argv, z);
p_argv++;
argp += z;
}
}
return struct_value_ptr;
}
#define CIF_FLAGS_INT 1
#define CIF_FLAGS_DINT 2
#define CIF_FLAGS_FLOAT 4
#define CIF_FLAGS_DOUBLE 8
#define CIF_FLAGS_LDOUBLE 16
#define CIF_FLAGS_POINTER 32
#define CIF_FLAGS_STRUCT 64
/* Perform machine dependent cif processing */
ffi_status
ffi_prep_cif_machdep (ffi_cif *cif)
{
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
cif->flags = 0;
break;
case FFI_TYPE_STRUCT:
if (cif->rtype->size > 4 && cif->rtype->size <= 8)
cif->flags = CIF_FLAGS_DINT;
else if (cif->rtype->size <= 4)
cif->flags = CIF_FLAGS_STRUCT;
else
cif->flags = 0;
break;
case FFI_TYPE_FLOAT:
cif->flags = CIF_FLAGS_FLOAT;
break;
case FFI_TYPE_DOUBLE:
cif->flags = CIF_FLAGS_DOUBLE;
break;
case FFI_TYPE_LONGDOUBLE:
cif->flags = CIF_FLAGS_LDOUBLE;
break;
case FFI_TYPE_POINTER:
cif->flags = CIF_FLAGS_POINTER;
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
cif->flags = CIF_FLAGS_DINT;
break;
default:
cif->flags = CIF_FLAGS_INT;
break;
}
return FFI_OK;
}
extern void ffi_call_SYSV (void *(*) (void *, extended_cif *),
extended_cif *,
unsigned, unsigned, unsigned,
void *, void (*fn) ());
void
ffi_call (ffi_cif *cif, void (*fn) (), void *rvalue, void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return value
address then we need to make one. */
if (rvalue == NULL
&& cif->rtype->type == FFI_TYPE_STRUCT
&& cif->rtype->size > 8)
ecif.rvalue = alloca (cif->rtype->size);
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
ffi_call_SYSV (ffi_prep_args, &ecif, cif->bytes,
cif->flags, cif->rtype->size * 8,
ecif.rvalue, fn);
break;
default:
FFI_ASSERT (0);
break;
}
}
+96
View File
@@ -0,0 +1,96 @@
/* -----------------------------------------------------------------------
sysv.S
m68k Foreign Function Interface
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
.text
.globl ffi_call_SYSV
.type ffi_call_SYSV,@function
ffi_call_SYSV:
link %fp,#0
move.l %d2,-(%sp)
| Make room for all of the new args.
sub.l 16(%fp),%sp
| Call ffi_prep_args
move.l 12(%fp),-(%sp)
pea 4(%sp)
move.l 8(%fp),%a0
jsr (%a0)
addq.l #8,%sp
| Pass pointer to struct value, if any
move.l %a0,%a1
| Call the function
move.l 32(%fp),%a0
jsr (%a0)
| Remove the space we pushed for the args
add.l 16(%fp),%sp
| Load the pointer to storage for the return value
move.l 28(%fp),%a1
| Load the return type code
move.l 20(%fp),%d2
| If the return value pointer is NULL, assume no return value.
tst.l %a1
jbeq noretval
btst #0,%d2
jbeq retlongint
move.l %d0,(%a1)
jbra epilogue
retlongint:
btst #1,%d2
jbeq retfloat
move.l %d0,(%a1)
move.l %d1,4(%a1)
jbra epilogue
retfloat:
btst #2,%d2
jbeq retdouble
fmove.s %fp0,(%a1)
jbra epilogue
retdouble:
btst #3,%d2
jbeq retlongdouble
fmove.d %fp0,(%a1)
jbra epilogue
retlongdouble:
btst #4,%d2
jbeq retpointer
fmove.x %fp0,(%a1)
jbra epilogue
retpointer:
btst #5,%d2
jbeq retstruct
move.l %a0,(%a1)
jbra epilogue
retstruct:
btst #6,%d2
jbeq noretval
move.l 24(%fp),%d2
bfins %d0,(%a1){#0,%d2}
noretval:
epilogue:
move.l (%sp)+,%d2
unlk %a6
rts
.size ffi_call_SYSV,.-ffi_call_SYSV
+469
View File
@@ -0,0 +1,469 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1996, 2001 Red Hat, Inc.
MIPS Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <mips/mips.h>
#include <stdlib.h>
#if _MIPS_SIM == _MIPS_SIM_NABI32
#define FIX_ARGP \
FFI_ASSERT(argp <= &stack[bytes]); \
if (argp == &stack[bytes]) \
{ \
argp = stack; \
}
#else
#define FIX_ARGP
#endif
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments */
static void ffi_prep_args(char *stack,
extended_cif *ecif,
int bytes,
int flags)
{
register int i;
register int avn;
register void **p_argv;
register char *argp;
register ffi_type **p_arg;
#if _MIPS_SIM == _MIPS_SIM_NABI32
/* If more than 8 double words are used, the remainder go
on the stack. We reorder stuff on the stack here to
support this easily. */
if (bytes > 8 * SIZEOF_ARG)
argp = &stack[bytes - (8 * SIZEOF_ARG)];
else
argp = stack;
#else
argp = stack;
#endif
memset(stack, 0, bytes);
#if _MIPS_SIM == _MIPS_SIM_NABI32
if ( ecif->cif->rstruct_flag != 0 )
#else
if ( ecif->cif->rtype->type == FFI_TYPE_STRUCT )
#endif
{
*(SLOT_TYPE_UNSIGNED *) argp = (SLOT_TYPE_UNSIGNED) ecif->rvalue;
argp += sizeof(SLOT_TYPE_UNSIGNED);
FIX_ARGP;
}
avn = ecif->cif->nargs;
p_argv = ecif->avalue;
for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types;
i && avn;
i--, p_arg++)
{
size_t z;
/* Align if necessary */
if (((*p_arg)->alignment - 1) & (unsigned) argp) {
argp = (char *) ALIGN(argp, (*p_arg)->alignment);
FIX_ARGP;
}
#if _MIPS_SIM == _MIPS_SIM_ABI32
#define OFFSET 0
#else
#define OFFSET sizeof(int)
#endif
if (avn)
{
avn--;
z = (*p_arg)->size;
if (z < sizeof(SLOT_TYPE_UNSIGNED))
{
z = sizeof(SLOT_TYPE_UNSIGNED);
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(SINT32 *) &argp[OFFSET] = (SINT32)*(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
*(UINT32 *) &argp[OFFSET] = (UINT32)*(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
*(SINT32 *) &argp[OFFSET] = (SINT32)*(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
*(UINT32 *) &argp[OFFSET] = (UINT32)*(UINT16 *)(* p_argv);
break;
case FFI_TYPE_SINT32:
*(SINT32 *) &argp[OFFSET] = (SINT32)*(SINT32 *)(* p_argv);
break;
case FFI_TYPE_UINT32:
case FFI_TYPE_POINTER:
*(UINT32 *) &argp[OFFSET] = (UINT32)*(UINT32 *)(* p_argv);
break;
/* This can only happen with 64bit slots */
case FFI_TYPE_FLOAT:
*(float *) argp = *(float *)(* p_argv);
break;
/* Handle small structures */
case FFI_TYPE_STRUCT:
memcpy(argp, *p_argv, (*p_arg)->size);
break;
default:
FFI_ASSERT(0);
}
}
else
{
#if _MIPS_SIM == _MIPS_SIM_ABI32
memcpy(argp, *p_argv, z);
#else
{
unsigned end = (unsigned) argp+z;
unsigned cap = (unsigned) stack+bytes;
/* Check if the data will fit within the register
space. Handle it if it doesn't. */
if (end <= cap)
memcpy(argp, *p_argv, z);
else
{
unsigned portion = end - cap;
memcpy(argp, *p_argv, portion);
argp = stack;
memcpy(argp,
(void*)((unsigned)(*p_argv)+portion), z - portion);
}
}
#endif
}
p_argv++;
argp += z;
FIX_ARGP;
}
}
return;
}
#if _MIPS_SIM == _MIPS_SIM_NABI32
/* The n32 spec says that if "a chunk consists solely of a double
float field (but not a double, which is part of a union), it
is passed in a floating point register. Any other chunk is
passed in an integer register". This code traverses structure
definitions and generates the appropriate flags. */
static unsigned calc_n32_struct_flags(ffi_type *arg, unsigned *shift)
{
unsigned flags = 0;
unsigned index = 0;
ffi_type *e;
while (e = arg->elements[index])
{
if (e->type == FFI_TYPE_DOUBLE)
{
flags += (FFI_TYPE_DOUBLE << *shift);
*shift += FFI_FLAG_BITS;
}
else if (e->type == FFI_TYPE_STRUCT)
flags += calc_n32_struct_flags(e, shift);
else
*shift += FFI_FLAG_BITS;
index++;
}
return flags;
}
unsigned calc_n32_return_struct_flags(ffi_type *arg)
{
unsigned flags = 0;
unsigned index = 0;
unsigned small = FFI_TYPE_SMALLSTRUCT;
ffi_type *e;
/* Returning structures under n32 is a tricky thing.
A struct with only one or two floating point fields
is returned in $f0 (and $f2 if necessary). Any other
struct results at most 128 bits are returned in $2
(the first 64 bits) and $3 (remainder, if necessary).
Larger structs are handled normally. */
if (arg->size > 16)
return 0;
if (arg->size > 8)
small = FFI_TYPE_SMALLSTRUCT2;
e = arg->elements[0];
if (e->type == FFI_TYPE_DOUBLE)
flags = FFI_TYPE_DOUBLE << FFI_FLAG_BITS;
else if (e->type == FFI_TYPE_FLOAT)
flags = FFI_TYPE_FLOAT << FFI_FLAG_BITS;
if (flags && (e = arg->elements[1]))
{
if (e->type == FFI_TYPE_DOUBLE)
flags += FFI_TYPE_DOUBLE;
else if (e->type == FFI_TYPE_FLOAT)
flags += FFI_TYPE_FLOAT;
else
return small;
if (flags && (arg->elements[2]))
{
/* There are three arguments and the first two are
floats! This must be passed the old way. */
return small;
}
}
else
if (!flags)
return small;
return flags;
}
#endif
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
{
cif->flags = 0;
#if _MIPS_SIM == _MIPS_SIM_ABI32
/* Set the flags necessary for O32 processing */
if (cif->rtype->type != FFI_TYPE_STRUCT)
{
if (cif->nargs > 0)
{
switch ((cif->arg_types)[0]->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags += (cif->arg_types)[0]->type;
break;
default:
break;
}
if (cif->nargs > 1)
{
/* Only handle the second argument if the first
is a float or double. */
if (cif->flags)
{
switch ((cif->arg_types)[1]->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags += (cif->arg_types)[1]->type << FFI_FLAG_BITS;
break;
default:
break;
}
}
}
}
}
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
case FFI_TYPE_STRUCT:
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags += cif->rtype->type << (FFI_FLAG_BITS * 2);
break;
default:
cif->flags += FFI_TYPE_INT << (FFI_FLAG_BITS * 2);
break;
}
#endif
#if _MIPS_SIM == _MIPS_SIM_NABI32
/* Set the flags necessary for N32 processing */
{
unsigned shift = 0;
unsigned count = (cif->nargs < 8) ? cif->nargs : 8;
unsigned index = 0;
unsigned struct_flags = 0;
if (cif->rtype->type == FFI_TYPE_STRUCT)
{
struct_flags = calc_n32_return_struct_flags(cif->rtype);
if (struct_flags == 0)
{
/* This means that the structure is being passed as
a hidden argument */
shift = FFI_FLAG_BITS;
count = (cif->nargs < 7) ? cif->nargs : 7;
cif->rstruct_flag = !0;
}
else
cif->rstruct_flag = 0;
}
else
cif->rstruct_flag = 0;
while (count-- > 0)
{
switch ((cif->arg_types)[index]->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags += ((cif->arg_types)[index]->type << shift);
shift += FFI_FLAG_BITS;
break;
case FFI_TYPE_STRUCT:
cif->flags += calc_n32_struct_flags((cif->arg_types)[index],
&shift);
break;
default:
shift += FFI_FLAG_BITS;
}
index++;
}
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_STRUCT:
{
if (struct_flags == 0)
{
/* The structure is returned through a hidden
first argument. Do nothing, 'cause FFI_TYPE_VOID
is 0 */
}
else
{
/* The structure is returned via some tricky
mechanism */
cif->flags += FFI_TYPE_STRUCT << (FFI_FLAG_BITS * 8);
cif->flags += struct_flags << (4 + (FFI_FLAG_BITS * 8));
}
break;
}
case FFI_TYPE_VOID:
/* Do nothing, 'cause FFI_TYPE_VOID is 0 */
break;
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
cif->flags += cif->rtype->type << (FFI_FLAG_BITS * 8);
break;
default:
cif->flags += FFI_TYPE_INT << (FFI_FLAG_BITS * 8);
break;
}
}
#endif
return FFI_OK;
}
/* Low level routine for calling O32 functions */
extern int ffi_call_O32(void (*)(char *, extended_cif *, int, int),
extended_cif *, unsigned,
unsigned, unsigned *, void (*)());
/* Low level routine for calling N32 functions */
extern int ffi_call_N32(void (*)(char *, extended_cif *, int, int),
extended_cif *, unsigned,
unsigned, unsigned *, void (*)());
void ffi_call(ffi_cif *cif, void (*fn)(), void *rvalue, void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
ecif.rvalue = alloca(cif->rtype->size);
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
#if _MIPS_SIM == _MIPS_SIM_ABI32
case FFI_O32:
ffi_call_O32(ffi_prep_args, &ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
break;
#endif
#if _MIPS_SIM == _MIPS_SIM_NABI32
case FFI_N32:
ffi_call_N32(ffi_prep_args, &ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
break;
#endif
default:
FFI_ASSERT(0);
break;
}
}
+144
View File
@@ -0,0 +1,144 @@
/* -----------------------------------------------------------------------
ffi-mips.h - Copyright (c) 1996, 2001 Red Hat, Inc.
MIPS FFI Definitions
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 CYGNUS SUPPORT 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 MIPS_H
#if defined(__mips_eabi)
# define FFI_MIPS_EABI
# define FFI_MIPS_O32
#else
# if !defined(_MIPS_SIM)
-- something is very wrong --
# else
# if _MIPS_SIM==_ABIN32 && defined(_ABIN32)
# define FFI_MIPS_N32
# else
# if defined(__GNUC__)
# define FFI_MIPS_O32
# else
# if _MIPS_SIM==_ABIO32
# define FFI_MIPS_O32
# else
-- this is an unsupported platform --
# endif
# endif
# endif
# endif
#endif
#define v0 $2
#define v1 $3
#define a0 $4
#define a1 $5
#define a2 $6
#define a3 $7
#define a4 $8
#define a5 $9
#define a6 $10
#define a7 $11
#define t0 $8
#define t1 $9
#define t2 $10
#define t3 $11
#define t4 $12
#define t5 $13
#define t6 $14
#define t7 $15
#define t8 $24
#define t9 $25
#define ra $31
#if defined(FFI_MIPS_O32)
#define FFI_DEFAULT_ABI FFI_O32
/* O32 stack frames have 32bit integer args */
#define SLOT_TYPE_UNSIGNED UINT32
#define SLOT_TYPE_SIGNED SINT32
#define SIZEOF_ARG 4
#define REG_L lw
#define REG_S sw
#define SUBU subu
#define ADDU addu
#define SRL srl
#define LI li
#else
#define FFI_DEFAULT_ABI FFI_N32
/* N32 and N64 frames have 64bit integer args */
#define SLOT_TYPE_UNSIGNED UINT64
#define SLOT_TYPE_SIGNED SINT64
#define SIZEOF_ARG 8
#define REG_L ld
#define REG_S sd
#define SUBU dsubu
#define ADDU daddu
#define SRL dsrl
#define LI dli
#endif
#define FFI_FLAG_BITS 2
/* SGI's strange assembler requires that we multiply by 4 rather
than shift left by FFI_FLAG_BITS */
#define FFI_ARGS_D FFI_TYPE_DOUBLE
#define FFI_ARGS_F FFI_TYPE_FLOAT
#define FFI_ARGS_DD FFI_TYPE_DOUBLE * 4 + FFI_TYPE_DOUBLE
#define FFI_ARGS_FF FFI_TYPE_FLOAT * 4 + FFI_TYPE_FLOAT
#define FFI_ARGS_FD FFI_TYPE_DOUBLE * 4 + FFI_TYPE_FLOAT
#define FFI_ARGS_DF FFI_TYPE_FLOAT * 4 + FFI_TYPE_DOUBLE
/* Needed for N32 structure returns */
#define FFI_TYPE_SMALLSTRUCT FFI_TYPE_UINT8
#define FFI_TYPE_SMALLSTRUCT2 FFI_TYPE_SINT8
#if 0
/* The SGI assembler can't handle this.. */
#define FFI_TYPE_STRUCT_DD (( FFI_ARGS_DD ) << 4) + FFI_TYPE_STRUCT
#else
/* ...so we calculate these by hand! */
#define FFI_TYPE_STRUCT_D 61
#define FFI_TYPE_STRUCT_F 45
#define FFI_TYPE_STRUCT_DD 253
#define FFI_TYPE_STRUCT_FF 173
#define FFI_TYPE_STRUCT_FD 237
#define FFI_TYPE_STRUCT_DF 189
#define FFI_TYPE_STRUCT_SMALL 93
#define FFI_TYPE_STRUCT_SMALL2 109
#endif
#endif
+320
View File
@@ -0,0 +1,320 @@
/* -----------------------------------------------------------------------
n32.S - Copyright (c) 1996, 1998, 2001 Red Hat, Inc.
MIPS Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
#include <mips/mips.h>
/* Only build this code if we are compiling for n32 */
#if defined(FFI_MIPS_N32)
#define callback a0
#define bytes a2
#define flags a3
#define raddr a4
#define fn a5
#define SIZEOF_FRAME ( 8 * SIZEOF_ARG )
.text
.align 2
.globl ffi_call_N32
.ent ffi_call_N32
ffi_call_N32:
# Prologue
SUBU $sp, SIZEOF_FRAME # Frame size
REG_S $fp, SIZEOF_FRAME - 2*SIZEOF_ARG($sp) # Save frame pointer
REG_S ra, SIZEOF_FRAME - 1*SIZEOF_ARG($sp) # Save return address
move $fp, $sp
move t9, callback # callback function pointer
REG_S bytes, 2*SIZEOF_ARG($fp) # bytes
REG_S flags, 3*SIZEOF_ARG($fp) # flags
REG_S raddr, 4*SIZEOF_ARG($fp) # raddr
REG_S fn, 5*SIZEOF_ARG($fp) # fn
# Allocate at least 4 words in the argstack
move v0, bytes
bge bytes, 4 * SIZEOF_ARG, bigger
LI v0, 4 * SIZEOF_ARG
b sixteen
bigger:
ADDU t4, v0, 2 * SIZEOF_ARG -1 # make sure it is aligned
and v0, t4, -2 * SIZEOF_ARG # to a proper boundry.
sixteen:
SUBU $sp, $sp, v0 # move the stack pointer to reflect the
# arg space
ADDU a0, $sp, 0 # 4 * SIZEOF_ARG
ADDU a3, $fp, 3 * SIZEOF_ARG
# Call ffi_prep_args
jal t9
# ADDU $sp, $sp, 4 * SIZEOF_ARG # adjust $sp to new args
# Copy the stack pointer to t9
move t9, $sp
# Fix the stack if there are more than 8 64bit slots worth
# of arguments.
# Load the number of bytes
REG_L t6, 2*SIZEOF_ARG($fp)
# Is it bigger than 8 * SIZEOF_ARG?
dadd t7, $0, 8 * SIZEOF_ARG
dsub t8, t6, t7
bltz t8, loadregs
add t9, t9, t8
loadregs:
REG_L t4, 3*SIZEOF_ARG($fp) # load the flags word
add t6, t4, 0 # and copy it into t6
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg1_floatp
REG_L a0, 0*SIZEOF_ARG(t9)
b arg1_next
arg1_floatp:
bne t4, FFI_TYPE_FLOAT, arg1_doublep
l.s $f12, 0*SIZEOF_ARG(t9)
b arg1_next
arg1_doublep:
l.d $f12, 0*SIZEOF_ARG(t9)
arg1_next:
add t4, t6, 0
SRL t4, 1*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg2_floatp
REG_L a1, 1*SIZEOF_ARG(t9)
b arg2_next
arg2_floatp:
bne t4, FFI_TYPE_FLOAT, arg2_doublep
l.s $f13, 1*SIZEOF_ARG(t9)
b arg2_next
arg2_doublep:
l.d $f13, 1*SIZEOF_ARG(t9)
arg2_next:
add t4, t6, 0
SRL t4, 2*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg3_floatp
REG_L a2, 2*SIZEOF_ARG(t9)
b arg3_next
arg3_floatp:
bne t4, FFI_TYPE_FLOAT, arg3_doublep
l.s $f14, 2*SIZEOF_ARG(t9)
b arg3_next
arg3_doublep:
l.d $f14, 2*SIZEOF_ARG(t9)
arg3_next:
add t4, t6, 0
SRL t4, 3*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg4_floatp
REG_L a3, 3*SIZEOF_ARG(t9)
b arg4_next
arg4_floatp:
bne t4, FFI_TYPE_FLOAT, arg4_doublep
l.s $f15, 3*SIZEOF_ARG(t9)
b arg4_next
arg4_doublep:
l.d $f15, 3*SIZEOF_ARG(t9)
arg4_next:
add t4, t6, 0
SRL t4, 4*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg5_floatp
REG_L a4, 4*SIZEOF_ARG(t9)
b arg5_next
arg5_floatp:
bne t4, FFI_TYPE_FLOAT, arg5_doublep
l.s $f16, 4*SIZEOF_ARG(t9)
b arg5_next
arg5_doublep:
l.d $f16, 4*SIZEOF_ARG(t9)
arg5_next:
add t4, t6, 0
SRL t4, 5*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg6_floatp
REG_L a5, 5*SIZEOF_ARG(t9)
b arg6_next
arg6_floatp:
bne t4, FFI_TYPE_FLOAT, arg6_doublep
l.s $f17, 5*SIZEOF_ARG(t9)
b arg6_next
arg6_doublep:
l.d $f17, 5*SIZEOF_ARG(t9)
arg6_next:
add t4, t6, 0
SRL t4, 6*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg7_floatp
REG_L a6, 6*SIZEOF_ARG(t9)
b arg7_next
arg7_floatp:
bne t4, FFI_TYPE_FLOAT, arg7_doublep
l.s $f18, 6*SIZEOF_ARG(t9)
b arg7_next
arg7_doublep:
l.d $f18, 6*SIZEOF_ARG(t9)
arg7_next:
add t4, t6, 0
SRL t4, 7*FFI_FLAG_BITS
and t4, ((1<<FFI_FLAG_BITS)-1)
bnez t4, arg8_floatp
REG_L a7, 7*SIZEOF_ARG(t9)
b arg8_next
arg8_floatp:
bne t4, FFI_TYPE_FLOAT, arg8_doublep
l.s $f19, 7*SIZEOF_ARG(t9)
b arg8_next
arg8_doublep:
l.d $f19, 7*SIZEOF_ARG(t9)
arg8_next:
callit:
# Load the function pointer
REG_L t9, 5*SIZEOF_ARG($fp)
# If the return value pointer is NULL, assume no return value.
REG_L t5, 4*SIZEOF_ARG($fp)
beqz t5, noretval
# Shift the return type flag over
SRL t6, 8*FFI_FLAG_BITS
bne t6, FFI_TYPE_INT, retfloat
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
REG_S v0, 0(t4)
b epilogue
retfloat:
bne t6, FFI_TYPE_FLOAT, retdouble
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.s $f0, 0(t4)
b epilogue
retdouble:
bne t6, FFI_TYPE_DOUBLE, retstruct_d
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.d $f0, 0(t4)
b epilogue
retstruct_d:
bne t6, FFI_TYPE_STRUCT_D, retstruct_f
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.d $f0, 0(t4)
b epilogue
retstruct_f:
bne t6, FFI_TYPE_STRUCT_F, retstruct_d_d
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.s $f0, 0(t4)
b epilogue
retstruct_d_d:
bne t6, FFI_TYPE_STRUCT_DD, retstruct_f_f
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.d $f0, 0(t4)
s.d $f2, 8(t4)
b epilogue
retstruct_f_f:
bne t6, FFI_TYPE_STRUCT_FF, retstruct_d_f
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.s $f0, 0(t4)
s.s $f2, 4(t4)
b epilogue
retstruct_d_f:
bne t6, FFI_TYPE_STRUCT_DF, retstruct_f_d
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.d $f0, 0(t4)
s.s $f2, 8(t4)
b epilogue
retstruct_f_d:
bne t6, FFI_TYPE_STRUCT_FD, retstruct_small
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
s.s $f0, 0(t4)
s.d $f2, 8(t4)
b epilogue
retstruct_small:
bne t6, FFI_TYPE_STRUCT_SMALL, retstruct_small2
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
REG_S v0, 0(t4)
b epilogue
retstruct_small2:
bne t6, FFI_TYPE_STRUCT_SMALL2, retstruct
jal t9
REG_L t4, 4*SIZEOF_ARG($fp)
REG_S v0, 0(t4)
REG_S v1, 8(t4)
b epilogue
retstruct:
noretval:
jal t9
# Epilogue
epilogue:
move $sp, $fp
REG_L $fp, SIZEOF_FRAME - 2*SIZEOF_ARG($sp) # Restore frame pointer
REG_L ra, SIZEOF_FRAME - 1*SIZEOF_ARG($sp) # Restore return address
ADDU $sp, SIZEOF_FRAME # Fix stack pointer
j ra
.end ffi_call_N32
#endif
+173
View File
@@ -0,0 +1,173 @@
/* -----------------------------------------------------------------------
o32.S - Copyright (c) 1996, 1998, 2001 Red Hat, Inc.
MIPS Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
#include <mips/mips.h>
/* Only build this code if we are compiling for o32 */
#if defined(FFI_MIPS_O32)
#define callback a0
#define bytes a2
#define flags a3
#define SIZEOF_FRAME ( 4 * SIZEOF_ARG + 2 * SIZEOF_ARG )
.text
.align 2
.globl ffi_call_O32
.ent ffi_call_O32
ffi_call_O32:
# Prologue
SUBU $sp, SIZEOF_FRAME # Frame size
REG_S $fp, SIZEOF_FRAME - 2*SIZEOF_ARG($sp) # Save frame pointer
REG_S ra, SIZEOF_FRAME - 1*SIZEOF_ARG($sp) # Save return address
move $fp, $sp
move t9, callback # callback function pointer
REG_S flags, SIZEOF_FRAME + 3*SIZEOF_ARG($fp) # flags
# Allocate at least 4 words in the argstack
move v0, bytes
bge bytes, 4 * SIZEOF_ARG, bigger
LI v0, 4 * SIZEOF_ARG
b sixteen
bigger:
ADDU t0, v0, 2 * SIZEOF_ARG -1 # make sure it is aligned
and v0, t0, -2 * SIZEOF_ARG # to an 8 byte boundry
sixteen:
SUBU $sp, $sp, v0 # move the stack pointer to reflect the
# arg space
ADDU a0, $sp, 4 * SIZEOF_ARG
ADDU a3, $fp, SIZEOF_FRAME + 3*SIZEOF_ARG
jal t9
REG_L t0, SIZEOF_FRAME + 3*SIZEOF_ARG($fp) # load the flags word
add t2, t0, 0 # and copy it into t2
and t0, ((1<<4)-1) # mask out the return type
SRL t2, 4 # shift our arg info
ADDU $sp, $sp, 4 * SIZEOF_ARG # adjust $sp to new args
bnez t0, pass_d # make it quick for int
REG_L a0, 0*SIZEOF_ARG($sp) # just go ahead and load the
REG_L a1, 1*SIZEOF_ARG($sp) # four regs.
REG_L a2, 2*SIZEOF_ARG($sp)
REG_L a3, 3*SIZEOF_ARG($sp)
b call_it
pass_d:
bne t0, FFI_ARGS_D, pass_f
l.d $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
REG_L a2, 2*SIZEOF_ARG($sp) # passing a double
REG_L a3, 3*SIZEOF_ARG($sp)
b call_it
pass_f:
bne t0, FFI_ARGS_F, pass_d_d
l.s $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
REG_L a1, 1*SIZEOF_ARG($sp) # passing a float
REG_L a2, 2*SIZEOF_ARG($sp)
REG_L a3, 3*SIZEOF_ARG($sp)
b call_it
pass_d_d:
bne t0, FFI_ARGS_DD, pass_f_f
l.d $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
l.d $f14, 2*SIZEOF_ARG($sp) # passing two doubles
b call_it
pass_f_f:
bne t0, FFI_ARGS_FF, pass_d_f
l.s $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
l.s $f14, 1*SIZEOF_ARG($sp) # passing two floats
REG_L a2, 2*SIZEOF_ARG($sp)
REG_L a3, 3*SIZEOF_ARG($sp)
b call_it
pass_d_f:
bne t0, FFI_ARGS_DF, pass_f_d
l.d $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
l.s $f14, 2*SIZEOF_ARG($sp) # passing double and float
REG_L a3, 3*SIZEOF_ARG($sp)
b call_it
pass_f_d:
# assume that the only other combination must be float then double
# bne t0, FFI_ARGS_F_D, call_it
l.s $f12, 0*SIZEOF_ARG($sp) # load $fp regs from args
l.d $f14, 2*SIZEOF_ARG($sp) # passing double and float
call_it:
# Load the function pointer
REG_L t9, SIZEOF_FRAME + 5*SIZEOF_ARG($fp)
# If the return value pointer is NULL, assume no return value.
REG_L t1, SIZEOF_FRAME + 4*SIZEOF_ARG($fp)
beqz t1, noretval
bne t2, FFI_TYPE_INT, retfloat
jal t9
REG_L t0, SIZEOF_FRAME + 4*SIZEOF_ARG($fp)
REG_S v0, 0(t0)
b epilogue
retfloat:
bne t2, FFI_TYPE_FLOAT, retdouble
jal t9
REG_L t0, SIZEOF_FRAME + 4*SIZEOF_ARG($fp)
s.s $f0, 0(t0)
b epilogue
retdouble:
bne t2, FFI_TYPE_DOUBLE, noretval
jal t9
REG_L t0, SIZEOF_FRAME + 4*SIZEOF_ARG($fp)
s.d $f0, 0(t0)
b epilogue
noretval:
jal t9
# Epilogue
epilogue:
move $sp, $fp
REG_L $fp, SIZEOF_FRAME - 2*SIZEOF_ARG($sp) # Restore frame pointer
REG_L ra, SIZEOF_FRAME - 1*SIZEOF_ARG($sp) # Restore return address
ADDU $sp, SIZEOF_FRAME # Fix stack pointer
j ra
.end ffi_call_O32
#endif
+128
View File
@@ -0,0 +1,128 @@
/* -----------------------------------------------------------------------
asm.h - Copyright (c) 1998 Geoffrey Keating
PowerPC Assembly glue.
$Id: asm.h,v 1.1 2001/08/02 19:33:38 orph Exp $
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 AUTHOR 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.
----------------------------------------------------------------------- */
#define ASM_GLOBAL_DIRECTIVE .globl
#define C_SYMBOL_NAME(name) name
/* Macro for a label. */
#ifdef __STDC__
#define C_LABEL(name) name##:
#else
#define C_LABEL(name) name/**/:
#endif
/* This seems to always be the case on PPC. */
#define ALIGNARG(log2) log2
/* For ELF we need the `.type' directive to make shared libs work right. */
#define ASM_TYPE_DIRECTIVE(name,typearg) .type name,typearg;
#define ASM_SIZE_DIRECTIVE(name) .size name,.-name
/* If compiled for profiling, call `_mcount' at the start of each function. */
#ifdef PROF
/* The mcount code relies on a the return address being on the stack
to locate our caller and so it can restore it; so store one just
for its benefit. */
#ifdef PIC
#define CALL_MCOUNT \
.pushsection; \
.section ".data"; \
.align ALIGNARG(2); \
0:.long 0; \
.previous; \
mflr %r0; \
stw %r0,4(%r1); \
bl _GLOBAL_OFFSET_TABLE_@local-4; \
mflr %r11; \
lwz %r0,0b@got(%r11); \
bl JUMPTARGET(_mcount);
#else /* PIC */
#define CALL_MCOUNT \
.section ".data"; \
.align ALIGNARG(2); \
0:.long 0; \
.previous; \
mflr %r0; \
lis %r11,0b@ha; \
stw %r0,4(%r1); \
addi %r0,%r11,0b@l; \
bl JUMPTARGET(_mcount);
#endif /* PIC */
#else /* PROF */
#define CALL_MCOUNT /* Do nothing. */
#endif /* PROF */
#define ENTRY(name) \
ASM_GLOBAL_DIRECTIVE C_SYMBOL_NAME(name); \
ASM_TYPE_DIRECTIVE (C_SYMBOL_NAME(name),@function) \
.align ALIGNARG(2); \
C_LABEL(name) \
CALL_MCOUNT
#define EALIGN_W_0 /* No words to insert. */
#define EALIGN_W_1 nop
#define EALIGN_W_2 nop;nop
#define EALIGN_W_3 nop;nop;nop
#define EALIGN_W_4 EALIGN_W_3;nop
#define EALIGN_W_5 EALIGN_W_4;nop
#define EALIGN_W_6 EALIGN_W_5;nop
#define EALIGN_W_7 EALIGN_W_6;nop
/* EALIGN is like ENTRY, but does alignment to 'words'*4 bytes
past a 2^align boundary. */
#ifdef PROF
#define EALIGN(name, alignt, words) \
ASM_GLOBAL_DIRECTIVE C_SYMBOL_NAME(name); \
ASM_TYPE_DIRECTIVE (C_SYMBOL_NAME(name),@function) \
.align ALIGNARG(2); \
C_LABEL(name) \
CALL_MCOUNT \
b 0f; \
.align ALIGNARG(alignt); \
EALIGN_W_##words; \
0:
#else /* PROF */
#define EALIGN(name, alignt, words) \
ASM_GLOBAL_DIRECTIVE C_SYMBOL_NAME(name); \
ASM_TYPE_DIRECTIVE (C_SYMBOL_NAME(name),@function) \
.align ALIGNARG(alignt); \
EALIGN_W_##words; \
C_LABEL(name)
#endif
#define END(name) \
ASM_SIZE_DIRECTIVE(name)
#ifdef PIC
#define JUMPTARGET(name) name##@plt
#else
#define JUMPTARGET(name) name
#endif
/* Local labels stripped out by the linker. */
#define L(x) .L##x
+680
View File
@@ -0,0 +1,680 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1998 Geoffrey Keating
PowerPC Foreign Function Interface
$Id: ffi.c,v 1.1 2001/08/02 19:31:55 orph Exp $
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 AUTHOR 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
#include <stdio.h>
extern void ffi_closure_SYSV(void);
enum {
/* The assembly depends on these exact flags. */
FLAG_RETURNS_NOTHING = 1 << (31-30), /* These go in cr7 */
FLAG_RETURNS_FP = 1 << (31-29),
FLAG_RETURNS_64BITS = 1 << (31-28),
FLAG_ARG_NEEDS_COPY = 1 << (31- 7),
FLAG_FP_ARGUMENTS = 1 << (31- 6), /* cr1.eq; specified by ABI */
FLAG_4_GPR_ARGUMENTS = 1 << (31- 5),
FLAG_RETVAL_REFERENCE = 1 << (31- 4)
};
/* About the SYSV ABI. */
enum {
NUM_GPR_ARG_REGISTERS = 8,
NUM_FPR_ARG_REGISTERS = 8
};
enum { ASM_NEEDS_REGISTERS = 4 };
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments.
The stack layout we want looks like this:
| Return address from ffi_call_SYSV 4bytes | higher addresses
|--------------------------------------------|
| Previous backchain pointer 4 | stack pointer here
|--------------------------------------------|<+ <<< on entry to
| Saved r28-r31 4*4 | | ffi_call_SYSV
|--------------------------------------------| |
| GPR registers r3-r10 8*4 | | ffi_call_SYSV
|--------------------------------------------| |
| FPR registers f1-f8 (optional) 8*8 | |
|--------------------------------------------| | stack |
| Space for copied structures | | grows |
|--------------------------------------------| | down V
| Parameters that didn't fit in registers | |
|--------------------------------------------| | lower addresses
| Space for callee's LR 4 | |
|--------------------------------------------| | stack pointer here
| Current backchain pointer 4 |-/ during
|--------------------------------------------| <<< ffi_call_SYSV
*/
/*@-exportheader@*/
void ffi_prep_args(extended_cif *ecif, unsigned *const stack)
/*@=exportheader@*/
{
const unsigned bytes = ecif->cif->bytes;
const unsigned flags = ecif->cif->flags;
/* 'stacktop' points at the previous backchain pointer. */
unsigned *const stacktop = stack + (ecif->cif->bytes / sizeof(unsigned));
/* 'gpr_base' points at the space for gpr3, and grows upwards as
we use GPR registers. */
unsigned *gpr_base = stacktop - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS;
int intarg_count = 0;
/* 'fpr_base' points at the space for fpr1, and grows upwards as
we use FPR registers. */
double *fpr_base = (double *)gpr_base - NUM_FPR_ARG_REGISTERS;
int fparg_count = 0;
/* 'copy_space' grows down as we put structures in it. It should
stay 16-byte aligned. */
char *copy_space = ((flags & FLAG_FP_ARGUMENTS)
? (char *)fpr_base
: (char *)gpr_base);
/* 'next_arg' grows up as we put parameters in it. */
unsigned *next_arg = stack + 2;
int i;
ffi_type **ptr;
double double_tmp;
void **p_argv;
size_t struct_copy_size;
unsigned gprvalue;
/* Check that everything starts aligned properly. */
FFI_ASSERT(((unsigned)(char *)stack & 0xF) == 0);
FFI_ASSERT(((unsigned)(char *)copy_space & 0xF) == 0);
FFI_ASSERT(((unsigned)(char *)stacktop & 0xF) == 0);
FFI_ASSERT((bytes & 0xF) == 0);
FFI_ASSERT(copy_space >= (char *)next_arg);
/* Deal with return values that are actually pass-by-reference. */
if (flags & FLAG_RETVAL_REFERENCE)
{
*gpr_base++ = (unsigned)(char *)ecif->rvalue;
intarg_count++;
}
/* Now for the arguments. */
p_argv = ecif->avalue;
for (ptr = ecif->cif->arg_types, i = ecif->cif->nargs;
i > 0;
i--, ptr++, p_argv++)
{
switch ((*ptr)->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
if ((*ptr)->type == FFI_TYPE_FLOAT)
double_tmp = *(float *)*p_argv;
else
double_tmp = *(double *)*p_argv;
if (fparg_count >= NUM_FPR_ARG_REGISTERS)
{
if (intarg_count%2 != 0)
{
intarg_count++;
next_arg++;
}
*(double *)next_arg = double_tmp;
next_arg += 2;
}
else
*fpr_base++ = double_tmp;
fparg_count++;
FFI_ASSERT(flags & FLAG_FP_ARGUMENTS);
break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
if (intarg_count == NUM_GPR_ARG_REGISTERS-1)
intarg_count++;
if (intarg_count >= NUM_GPR_ARG_REGISTERS)
{
if (intarg_count%2 != 0)
{
intarg_count++;
next_arg++;
}
*(long long *)next_arg = *(long long *)*p_argv;
next_arg += 2;
}
else
{
/* whoops: abi states only certain register pairs
* can be used for passing long long int
* specifically (r3,r4), (r5,r6), (r7,r8),
* (r9,r10) and if next arg is long long but
* not correct starting register of pair then skip
* until the proper starting register
*/
if (intarg_count%2 != 0)
{
intarg_count ++;
gpr_base++;
}
*(long long *)gpr_base = *(long long *)*p_argv;
gpr_base += 2;
}
intarg_count += 2;
break;
case FFI_TYPE_STRUCT:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
struct_copy_size = ((*ptr)->size + 15) & ~0xF;
copy_space -= struct_copy_size;
memcpy(copy_space, (char *)*p_argv, (*ptr)->size);
gprvalue = (unsigned)copy_space;
FFI_ASSERT(copy_space > (char *)next_arg);
FFI_ASSERT(flags & FLAG_ARG_NEEDS_COPY);
goto putgpr;
case FFI_TYPE_UINT8:
gprvalue = *(unsigned char *)*p_argv;
goto putgpr;
case FFI_TYPE_SINT8:
gprvalue = *(signed char *)*p_argv;
goto putgpr;
case FFI_TYPE_UINT16:
gprvalue = *(unsigned short *)*p_argv;
goto putgpr;
case FFI_TYPE_SINT16:
gprvalue = *(signed short *)*p_argv;
goto putgpr;
case FFI_TYPE_INT:
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT32:
case FFI_TYPE_POINTER:
gprvalue = *(unsigned *)*p_argv;
putgpr:
if (intarg_count >= NUM_GPR_ARG_REGISTERS)
*next_arg++ = gprvalue;
else
*gpr_base++ = gprvalue;
intarg_count++;
break;
}
}
/* Check that we didn't overrun the stack... */
FFI_ASSERT(copy_space >= (char *)next_arg);
FFI_ASSERT(gpr_base <= stacktop - ASM_NEEDS_REGISTERS);
FFI_ASSERT((unsigned *)fpr_base
<= stacktop - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS);
FFI_ASSERT(flags & FLAG_4_GPR_ARGUMENTS || intarg_count <= 4);
}
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
{
/* All this is for the SYSV ABI. */
int i;
ffi_type **ptr;
unsigned bytes;
int fparg_count = 0, intarg_count = 0;
unsigned flags = 0;
unsigned struct_copy_size = 0;
/* All the machine-independent calculation of cif->bytes will be wrong.
Redo the calculation for SYSV. */
/* Space for the frame pointer, callee's LR, and the asm's temp regs. */
bytes = (2 + ASM_NEEDS_REGISTERS) * sizeof(int);
/* Space for the GPR registers. */
bytes += NUM_GPR_ARG_REGISTERS * sizeof(int);
/* Return value handling. The rules are as follows:
- 32-bit (or less) integer values are returned in gpr3;
- Structures of size <= 4 bytes also returned in gpr3;
- 64-bit integer values and structures between 5 and 8 bytes are returned
in gpr3 and gpr4;
- Single/double FP values are returned in fpr1;
- Larger structures and long double (if not equivalent to double) values
are allocated space and a pointer is passed as the first argument. */
switch (cif->rtype->type)
{
case FFI_TYPE_DOUBLE:
flags |= FLAG_RETURNS_64BITS;
/* Fall through. */
case FFI_TYPE_FLOAT:
flags |= FLAG_RETURNS_FP;
break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
flags |= FLAG_RETURNS_64BITS;
break;
case FFI_TYPE_STRUCT:
if (cif->abi != FFI_GCC_SYSV)
if (cif->rtype->size <= 4)
break;
else if (cif->rtype->size <= 8)
{
flags |= FLAG_RETURNS_64BITS;
break;
}
/* else fall through. */
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
intarg_count++;
flags |= FLAG_RETVAL_REFERENCE;
/* Fall through. */
case FFI_TYPE_VOID:
flags |= FLAG_RETURNS_NOTHING;
break;
default:
/* Returns 32-bit integer, or similar. Nothing to do here. */
break;
}
/* The first NUM_GPR_ARG_REGISTERS words of integer arguments, and the
first NUM_FPR_ARG_REGISTERS fp arguments, go in registers; the rest
goes on the stack. Structures and long doubles (if not equivalent
to double) are passed as a pointer to a copy of the structure.
Stuff on the stack needs to keep proper alignment. */
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
switch ((*ptr)->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
fparg_count++;
/* If this FP arg is going on the stack, it must be
8-byte-aligned. */
if (fparg_count > NUM_FPR_ARG_REGISTERS
&& intarg_count%2 != 0)
intarg_count++;
break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
/* 'long long' arguments are passed as two words, but
either both words must fit in registers or both go
on the stack. If they go on the stack, they must
be 8-byte-aligned. */
if (intarg_count == NUM_GPR_ARG_REGISTERS-1
|| intarg_count >= NUM_GPR_ARG_REGISTERS && intarg_count%2 != 0)
intarg_count++;
intarg_count += 2;
break;
case FFI_TYPE_STRUCT:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
/* We must allocate space for a copy of these to enforce
pass-by-value. Pad the space up to a multiple of 16
bytes (the maximum alignment required for anything under
the SYSV ABI). */
struct_copy_size += ((*ptr)->size + 15) & ~0xF;
/* Fall through (allocate space for the pointer). */
default:
/* Everything else is passed as a 4-byte word in a GPR, either
the object itself or a pointer to it. */
intarg_count++;
break;
}
}
if (fparg_count != 0)
flags |= FLAG_FP_ARGUMENTS;
if (intarg_count > 4)
flags |= FLAG_4_GPR_ARGUMENTS;
if (struct_copy_size != 0)
flags |= FLAG_ARG_NEEDS_COPY;
/* Space for the FPR registers, if needed. */
if (fparg_count != 0)
bytes += NUM_FPR_ARG_REGISTERS * sizeof(double);
/* Stack space. */
if (intarg_count > NUM_GPR_ARG_REGISTERS)
bytes += (intarg_count - NUM_GPR_ARG_REGISTERS) * sizeof(int);
if (fparg_count > NUM_FPR_ARG_REGISTERS)
bytes += (fparg_count - NUM_FPR_ARG_REGISTERS) * sizeof(double);
/* The stack space allocated needs to be a multiple of 16 bytes. */
bytes = (bytes + 15) & ~0xF;
/* Add in the space for the copied structures. */
bytes += struct_copy_size;
cif->flags = flags;
cif->bytes = bytes;
return FFI_OK;
}
/*@-declundef@*/
/*@-exportheader@*/
extern void ffi_call_SYSV(/*@out@*/ extended_cif *,
unsigned, unsigned,
/*@out@*/ unsigned *,
void (*fn)());
/*@=declundef@*/
/*@=exportheader@*/
void ffi_call(/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
{
/*@-sysunrecog@*/
ecif.rvalue = alloca(cif->rtype->size);
/*@=sysunrecog@*/
}
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
case FFI_GCC_SYSV:
/*@-usedef@*/
ffi_call_SYSV(&ecif, -cif->bytes,
cif->flags, ecif.rvalue, fn);
/*@=usedef@*/
break;
default:
FFI_ASSERT(0);
break;
}
}
static void flush_icache(char *, int);
ffi_status
ffi_prep_closure (ffi_closure* closure,
ffi_cif* cif,
void (*fun)(ffi_cif*, void*, void**, void*),
void *user_data)
{
unsigned int *tramp;
FFI_ASSERT (cif->abi == FFI_GCC_SYSV);
tramp = (unsigned int *) &closure->tramp[0];
tramp[0] = 0x7c0802a6; /* mflr r0 */
tramp[1] = 0x4800000d; /* bl 10 <trampoline_initial+0x10> */
tramp[4] = 0x7d6802a6; /* mflr r11 */
tramp[5] = 0x7c0803a6; /* mtlr r0 */
tramp[6] = 0x800b0000; /* lwz r0,0(r11) */
tramp[7] = 0x816b0004; /* lwz r11,4(r11) */
tramp[8] = 0x7c0903a6; /* mtctr r0 */
tramp[9] = 0x4e800420; /* bctr */
*(void **) &tramp[2] = (void *)ffi_closure_SYSV; /* function */
*(void **) &tramp[3] = (void *)closure; /* context */
closure->cif = cif;
closure->fun = fun;
closure->user_data = user_data;
/* Flush the icache. */
flush_icache(&closure->tramp[0],FFI_TRAMPOLINE_SIZE);
return FFI_OK;
}
#define MIN_CACHE_LINE_SIZE 8
static void flush_icache(char * addr1, int size)
{
int i;
char * addr;
for (i = 0; i < size; i += MIN_CACHE_LINE_SIZE) {
addr = addr1 + i;
__asm__ volatile ("icbi 0,%0;" "dcbf 0,%0;" : : "r"(addr) : "memory");
}
addr = addr1 + size - 1;
__asm__ volatile ("icbi 0,%0;" "dcbf 0,%0;" "sync;" "isync;" : : "r"(addr) : "memory");
}
int ffi_closure_helper_SYSV (ffi_closure*, void*, unsigned long*,
unsigned long*, unsigned long*);
/* Basically the trampoline invokes ffi_closure_SYSV, and on
* entry, r11 holds the address of the closure.
* After storing the registers that could possibly contain
* parameters to be passed into the stack frame and setting
* up space for a return value, ffi_closure_SYSV invokes the
* following helper function to do most of the work
*/
int
ffi_closure_helper_SYSV (ffi_closure* closure, void * rvalue,
unsigned long * pgr, unsigned long * pfr,
unsigned long * pst)
{
/* rvalue is the pointer to space for return value in closure assembly */
/* pgr is the pointer to where r3-r10 are stored in ffi_closure_SYSV */
/* pfr is the pointer to where f1-f8 are stored in ffi_closure_SYSV */
/* pst is the pointer to outgoing parameter stack in original caller */
void ** avalue;
ffi_type ** arg_types;
long i, avn;
long nf; /* number of floating registers already used */
long ng; /* number of general registers already used */
ffi_cif * cif;
double temp;
cif = closure->cif;
avalue = alloca(cif->nargs * sizeof(void *));
nf = 0;
ng = 0;
/* Copy the caller's structure return value address so that the closure
returns the data directly to the caller. */
if (cif->rtype->type == FFI_TYPE_STRUCT)
{
rvalue = *pgr;
ng++;
pgr++;
}
i = 0;
avn = cif->nargs;
arg_types = cif->arg_types;
/* Grab the addresses of the arguments from the stack frame. */
while (i < avn)
{
switch (arg_types[i]->type)
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
/* there are 8 gpr registers used to pass values */
if (ng < 8) {
avalue[i] = (((char *)pgr)+3);
ng++;
pgr++;
} else {
avalue[i] = (((char *)pst)+3);
pst++;
}
break;
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
/* there are 8 gpr registers used to pass values */
if (ng < 8) {
avalue[i] = (((char *)pgr)+2);
ng++;
pgr++;
} else {
avalue[i] = (((char *)pst)+2);
pst++;
}
break;
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
case FFI_TYPE_POINTER:
case FFI_TYPE_STRUCT:
/* there are 8 gpr registers used to pass values */
if (ng < 8) {
avalue[i] = pgr;
ng++;
pgr++;
} else {
avalue[i] = pst;
pst++;
}
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
/* passing long long ints are complex, they must
* be passed in suitable register pairs such as
* (r3,r4) or (r5,r6) or (r6,r7), or (r7,r8) or (r9,r10)
* and if the entire pair aren't available then the outgoing
* parameter stack is used for both but an alignment of 8
* must will be kept. So we must either look in pgr
* or pst to find the correct address for this type
* of parameter.
*/
if (ng < 7) {
if (ng & 0x01) {
/* skip r4, r6, r8 as starting points */
ng++;
pgr++;
}
avalue[i] = pgr;
ng+=2;
pgr+=2;
} else {
if (((long)pst) & 4) pst++;
avalue[i] = pst;
pst+=2;
}
break;
case FFI_TYPE_FLOAT:
/* unfortunately float values are stored as doubles
* in the ffi_closure_SYSV code (since we don't check
* the type in that routine). This is also true
* of floats passed on the outgoing parameter stack.
* Also, on the outgoing stack all values are aligned
* to 8
*
* Don't you just love the simplicity of this ABI!
*/
/* there are 8 64bit floating point registers */
if (nf < 8) {
temp = *(double*)pfr;
*(float*)pfr = (float)temp;
avalue[i] = pfr;
nf++;
pfr+=2;
} else {
/* FIXME? here we are really changing the values
* stored in the original calling routines outgoing
* parameter stack. This is probably a really
* naughty thing to do but...
*/
if (((long)pst) & 4) pst++;
temp = *(double*)pst;
*(float*)pst = (float)temp;
avalue[i] = pst;
nf++;
pst+=2;
}
break;
case FFI_TYPE_DOUBLE:
/* On the outgoing stack all values are aligned to 8 */
/* there are 8 64bit floating point registers */
if (nf < 8) {
avalue[i] = pfr;
nf++;
pfr+=2;
} else {
if (((long)pst) & 4) pst++;
avalue[i] = pst;
nf++;
pst+=2;
}
break;
default:
FFI_ASSERT(0);
}
i++;
}
(closure->fun) (cif, rvalue, avalue, closure->user_data);
/* Tell ffi_closure_osf how to perform return type promotions. */
return cif->rtype->type;
}
+148
View File
@@ -0,0 +1,148 @@
#define LIBFFI_ASM
#include <powerpc/asm.h>
.globl ffi_closure_helper_SYSV
ENTRY(ffi_closure_SYSV)
stwu %r1,-144(%r1)
mflr %r0
stw %r31,140(%r1)
stw %r0,148(%r1)
# we want to build up an areas for the parameters passed
# in registers (both floating point and integer)
# so first save gpr 3 to gpr 10 (aligned to 4)
stw %r3, 16(%r1)
stw %r4, 20(%r1)
stw %r5, 24(%r1)
stw %r6, 28(%r1)
stw %r7, 32(%r1)
stw %r8, 36(%r1)
stw %r9, 40(%r1)
stw %r10,44(%r1)
# next save fpr 1 to fpr 8 (aligned to 8)
stfd %f1, 48(%r1)
stfd %f2, 56(%r1)
stfd %f3, 64(%r1)
stfd %f4, 72(%r1)
stfd %f5, 80(%r1)
stfd %f6, 88(%r1)
stfd %f7, 96(%r1)
stfd %f8, 104(%r1)
# set up registers for the routine that actually does the work
# get the context pointer from the trampoline
mr %r3,%r11
# now load up the pointer to the result storage
addi %r4,%r1,112
# now load up the pointer to the saved gpr registers
addi %r5,%r1,16
# now load up the pointer to the saved fpr registers */
addi %r6,%r1,48
# now load up the pointer to the outgoing parameter
# stack in the previous frame
# i.e. the previous frame pointer + 8
addi %r7,%r1,152
# make the call
bl JUMPTARGET(ffi_closure_helper_SYSV)
# now r3 contains the return type
# so use it to look up in a table
# so we know how to deal with each type
# look up the proper starting point in table
# by using return type as offset
addi %r5,%r1,112 # get pointer to results area
addis %r4,0,.L60@ha # get address of jump table
addi %r4,%r4,.L60@l
slwi %r3,%r3,2 # now multiply return type by 4
lwzx %r3,%r4,%r3 # get the contents of that table value
add %r3,%r3,%r4 # add contents of table to table address
mtctr %r3
bctr # jump to it
.align 2
.L60:
.long .L44-.L60 # FFI_TYPE_VOID
.long .L50-.L60 # FFI_TYPE_INT
.long .L47-.L60 # FFI_TYPE_FLOAT
.long .L46-.L60 # FFI_TYPE_DOUBLE
.long .L46-.L60 # FFI_TYPE_LONGDOUBLE
.long .L56-.L60 # FFI_TYPE_UINT8
.long .L55-.L60 # FFI_TYPE_SINT8
.long .L58-.L60 # FFI_TYPE_UINT16
.long .L57-.L60 # FFI_TYPE_SINT16
.long .L50-.L60 # FFI_TYPE_UINT32
.long .L50-.L60 # FFI_TYPE_SINT32
.long .L48-.L60 # FFI_TYPE_UINT64
.long .L48-.L60 # FFI_TYPE_SINT64
.long .L44-.L60 # FFI_TYPE_STRUCT
.long .L50-.L60 # FFI_TYPE_POINTER
# case double
.L46:
lfd %f1,0(%r5)
b .L44
# case float
.L47:
lfs %f1,0(%r5)
b .L44
# case long long
.L48:
lwz %r3,0(%r5)
lwz %r4,4(%r5)
b .L44
# case default / int32 / pointer
.L50:
lwz %r3,0(%r5)
b .L44
# case signed int8
.L55:
addi %r5,%r5,3
lbz %r3,0(%r5)
extsb %r3,%r3
b .L44
# case unsigned int8
.L56:
addi %r5,%r5,3
lbz %r3,0(%r5)
b .L44
# case signed int16
.L57:
addi %r5,%r5,2
lhz %r3,0(%r5)
extsh %r3,%r3
b .L44
#case unsigned int16
.L58:
addi %r5,%r5,2
lhz %r3,0(%r5)
# case void / done
.L44:
lwz %r11,0(%r1)
lwz %r0,4(%r11)
mtlr %r0
lwz %r31,-4(%r11)
mr %r1,%r11
blr
END(ffi_closure_SYSV)
+119
View File
@@ -0,0 +1,119 @@
/* -----------------------------------------------------------------------
sysv.h - Copyright (c) 1998 Geoffrey Keating
PowerPC Assembly glue.
$Id: sysv.S,v 1.1 2001/08/02 19:31:55 orph Exp $
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 AUTHOR 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <powerpc/asm.h>
.globl ffi_prep_args
ENTRY(ffi_call_SYSV)
/* Save the old stack pointer as AP. */
mr %r8,%r1
/* Allocate the stack space we need. */
stwux %r1,%r1,%r4
/* Save registers we use. */
mflr %r9
stw %r28,-16(%r8)
stw %r29,-12(%r8)
stw %r30, -8(%r8)
stw %r31, -4(%r8)
stw %r9, 4(%r8)
/* Save arguments over call... */
mr %r31,%r5 /* flags, */
mr %r30,%r6 /* rvalue, */
mr %r29,%r7 /* function address, */
mr %r28,%r8 /* our AP. */
/* Call ffi_prep_args. */
mr %r4,%r1
bl JUMPTARGET(ffi_prep_args)
/* Now do the call. */
/* Set up cr1 with bits 4-7 of the flags. */
mtcrf 0x40,%r31
/* Get the address to call into CTR. */
mtctr %r29
/* Load all those argument registers. */
lwz %r3,-16-(8*4)(%r28)
lwz %r4,-16-(7*4)(%r28)
lwz %r5,-16-(6*4)(%r28)
lwz %r6,-16-(5*4)(%r28)
bf- 5,1f
nop
lwz %r7,-16-(4*4)(%r28)
lwz %r8,-16-(3*4)(%r28)
lwz %r9,-16-(2*4)(%r28)
lwz %r10,-16-(1*4)(%r28)
nop
1:
/* Load all the FP registers. */
bf- 6,2f
lfd %f1,-16-(8*4)-(8*8)(%r28)
lfd %f2,-16-(8*4)-(7*8)(%r28)
lfd %f3,-16-(8*4)-(6*8)(%r28)
lfd %f4,-16-(8*4)-(5*8)(%r28)
nop
lfd %f5,-16-(8*4)-(4*8)(%r28)
lfd %f6,-16-(8*4)-(3*8)(%r28)
lfd %f7,-16-(8*4)-(2*8)(%r28)
lfd %f8,-16-(8*4)-(1*8)(%r28)
2:
/* Make the call. */
bctrl
/* Now, deal with the return value. */
mtcrf 0x01,%r31
bt- 30,L(done_return_value)
bt- 29,L(fp_return_value)
stw %r3,0(%r30)
bf+ 28,L(done_return_value)
stw %r4,4(%r30)
/* Fall through... */
L(done_return_value):
/* Restore the registers we used and return. */
lwz %r9, 4(%r28)
lwz %r31, -4(%r28)
mtlr %r9
lwz %r30, -8(%r28)
lwz %r29,-12(%r28)
lwz %r28,-16(%r28)
lwz %r1,0(%r1)
blr
L(fp_return_value):
bf 28,L(float_return_value)
stfd %f1,0(%r30)
b L(done_return_value)
L(float_return_value):
stfs %f1,0(%r30)
b L(done_return_value)
END(ffi_call_SYSV)
+146
View File
@@ -0,0 +1,146 @@
/* -----------------------------------------------------------------------
prep_cif.c - Copyright (c) 1996, 1998 Cygnus Solutions
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 RED HAT 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
/* Round up to SIZEOF_ARG. */
#define STACK_ARG_SIZE(x) ALIGN(x, SIZEOF_ARG)
/* Perform machine independent initialization of aggregate type
specifications. */
static ffi_status initialize_aggregate(/*@out@*/ ffi_type *arg)
{
ffi_type **ptr;
FFI_ASSERT(arg != NULL);
/*@-usedef@*/
FFI_ASSERT(arg->elements != NULL);
FFI_ASSERT(arg->size == 0);
FFI_ASSERT(arg->alignment == 0);
ptr = &(arg->elements[0]);
while ((*ptr) != NULL)
{
if (((*ptr)->size == 0) && (initialize_aggregate((*ptr)) != FFI_OK))
return FFI_BAD_TYPEDEF;
/* Perform a sanity check on the argument type */
FFI_ASSERT(ffi_type_test((*ptr)));
arg->size = ALIGN(arg->size, (*ptr)->alignment);
arg->size += (*ptr)->size;
arg->alignment = (arg->alignment > (*ptr)->alignment) ?
arg->alignment : (*ptr)->alignment;
ptr++;
}
if (arg->size == 0)
return FFI_BAD_TYPEDEF;
else
return FFI_OK;
/*@=usedef@*/
}
/* Perform machine independent ffi_cif preparation, then call
machine dependent routine. */
ffi_status ffi_prep_cif(/*@out@*/ /*@partial@*/ ffi_cif *cif,
ffi_abi abi, unsigned int nargs,
/*@dependent@*/ /*@out@*/ /*@partial@*/ ffi_type *rtype,
/*@dependent@*/ ffi_type **atypes)
{
unsigned bytes = 0;
unsigned int i;
ffi_type **ptr;
FFI_ASSERT(cif != NULL);
FFI_ASSERT((abi > FFI_FIRST_ABI) && (abi < FFI_LAST_ABI));
cif->abi = abi;
cif->arg_types = atypes;
cif->nargs = nargs;
cif->rtype = rtype;
cif->flags = 0;
/* Initialize the return type if necessary */
/*@-usedef@*/
if ((cif->rtype->size == 0) && (initialize_aggregate(cif->rtype) != FFI_OK))
return FFI_BAD_TYPEDEF;
/*@=usedef@*/
/* Perform a sanity check on the return type */
FFI_ASSERT(ffi_type_test(cif->rtype));
#ifndef M68K
/* Make space for the return structure pointer */
if (cif->rtype->type == FFI_TYPE_STRUCT
#ifdef __sparc__
&& (cif->abi != FFI_V9 || cif->rtype->size > 32)
#endif
)
bytes = STACK_ARG_SIZE(sizeof(void*));
#endif
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
/* Perform a sanity check on the argument type */
FFI_ASSERT(ffi_type_test(*ptr));
/* Initialize any uninitialized aggregate type definitions */
if (((*ptr)->size == 0) && (initialize_aggregate((*ptr)) != FFI_OK))
return FFI_BAD_TYPEDEF;
#ifdef __sparc__
if (((*ptr)->type == FFI_TYPE_STRUCT
&& ((*ptr)->size > 16 || cif->abi != FFI_V9))
|| ((*ptr)->type == FFI_TYPE_LONGDOUBLE
&& cif->abi != FFI_V9))
bytes += sizeof(void*);
else
#endif
{
/* Add any padding if necessary */
if (((*ptr)->alignment - 1) & bytes)
bytes = ALIGN(bytes, (*ptr)->alignment);
bytes += STACK_ARG_SIZE((*ptr)->size);
}
}
cif->bytes = bytes;
/* Perform machine dependent cif processing */
return ffi_prep_cif_machdep(cif);
}
+240
View File
@@ -0,0 +1,240 @@
/* -----------------------------------------------------------------------
raw_api.c - Copyright (c) 1999 Cygnus Solutions
Author: Kresten Krab Thorup <[email protected]>
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 RED HAT 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.
----------------------------------------------------------------------- */
/* This file defines generic functions for use with the raw api. */
#include <ffi.h>
#include <ffi_private.h>
#if !FFI_NO_RAW_API
size_t
ffi_raw_size (ffi_cif *cif)
{
size_t result = 0;
int i;
ffi_type **at = cif->arg_types;
for (i = cif->nargs-1; i >= 0; i--, at++)
{
#if !FFI_NO_STRUCTS
if ((*at)->type == FFI_TYPE_STRUCT)
result += ALIGN (sizeof (void*), SIZEOF_ARG);
else
#endif
result += ALIGN ((*at)->size, SIZEOF_ARG);
}
return result;
}
void
ffi_raw_to_ptrarray (ffi_cif *cif, ffi_raw *raw, void **args)
{
unsigned i;
ffi_type **tp = cif->arg_types;
#if WORDS_BIGENDIAN
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
switch ((*tp)->type)
{
case FFI_TYPE_UINT8:
case FFI_TYPE_SINT8:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 1);
break;
case FFI_TYPE_UINT16:
case FFI_TYPE_SINT16:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 2);
break;
#if SIZEOF_ARG >= 4
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT32:
*args = (void*) ((char*)(raw++) + SIZEOF_ARG - 4);
break;
#endif
#if !FFI_NO_STRUCTS
case FFI_TYPE_STRUCT:
*args = (raw++)->ptr;
break;
#endif
case FFI_TYPE_POINTER:
*args = (void*) &(raw++)->ptr;
break;
default:
*args = raw;
raw += ALIGN ((*tp)->size, SIZEOF_ARG) / SIZEOF_ARG;
}
}
#else /* WORDS_BIGENDIAN */
#if !PDP
/* then assume little endian */
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
#if !FFI_NO_STRUCTS
if ((*tp)->type == FFI_TYPE_STRUCT)
{
*args = (raw++)->ptr;
}
else
#endif
{
*args = (void*) raw;
raw += ALIGN ((*tp)->size, sizeof (void*)) / sizeof (void*);
}
}
#else
#error "pdp endian not supported"
#endif /* ! PDP */
#endif /* WORDS_BIGENDIAN */
}
void
ffi_ptrarray_to_raw (ffi_cif *cif, void **args, ffi_raw *raw)
{
unsigned i;
ffi_type **tp = cif->arg_types;
for (i = 0; i < cif->nargs; i++, tp++, args++)
{
switch ((*tp)->type)
{
case FFI_TYPE_UINT8:
(raw++)->uint = *(UINT8*) (*args);
break;
case FFI_TYPE_SINT8:
(raw++)->sint = *(SINT8*) (*args);
break;
case FFI_TYPE_UINT16:
(raw++)->uint = *(UINT16*) (*args);
break;
case FFI_TYPE_SINT16:
(raw++)->sint = *(SINT16*) (*args);
break;
#if SIZEOF_ARG >= 4
case FFI_TYPE_UINT32:
(raw++)->uint = *(UINT32*) (*args);
break;
case FFI_TYPE_SINT32:
(raw++)->sint = *(SINT32*) (*args);
break;
#endif
#if !FFI_NO_STRUCTS
case FFI_TYPE_STRUCT:
(raw++)->ptr = *args;
break;
#endif
case FFI_TYPE_POINTER:
(raw++)->ptr = **(void***) args;
break;
default:
memcpy ((void*) raw->data, (void*)*args, (*tp)->size);
raw += ALIGN ((*tp)->size, SIZEOF_ARG) / SIZEOF_ARG;
}
}
}
#if !FFI_NATIVE_RAW_API
/* This is a generic definition of ffi_raw_call, to be used if the
* native system does not provide a machine-specific implementation.
* Having this, allows code to be written for the raw API, without
* the need for system-specific code to handle input in that format;
* these following couple of functions will handle the translation forth
* and back automatically. */
void ffi_raw_call (/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ ffi_raw *raw)
{
void **avalue = (void**) alloca (cif->nargs * sizeof (void*));
ffi_raw_to_ptrarray (cif, raw, avalue);
ffi_call (cif, fn, rvalue, avalue);
}
#if FFI_CLOSURES /* base system provides closures */
static void
ffi_translate_args (ffi_cif *cif, void *rvalue,
void **avalue, void *user_data)
{
ffi_raw *raw = (ffi_raw*)alloca (ffi_raw_size (cif));
ffi_raw_closure *cl = (ffi_raw_closure*)user_data;
ffi_ptrarray_to_raw (cif, avalue, raw);
(*cl->fun) (cif, rvalue, raw, cl->user_data);
}
/* Again, here is the generic version of ffi_prep_raw_closure, which
* will install an intermediate "hub" for translation of arguments from
* the pointer-array format, to the raw format */
ffi_status
ffi_prep_raw_closure (ffi_raw_closure* cl,
ffi_cif *cif,
void (*fun)(ffi_cif*,void*,ffi_raw*,void*),
void *user_data)
{
ffi_status status;
status = ffi_prep_closure ((ffi_closure*) cl,
cif,
&ffi_translate_args,
(void*)cl);
if (status == FFI_OK)
{
cl->fun = fun;
cl->user_data = user_data;
}
return status;
}
#endif /* FFI_CLOSURES */
#endif /* !FFI_NATIVE_RAW_API */
#endif /* !FFI_NO_RAW_API */
+589
View File
@@ -0,0 +1,589 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 2000 Software AG
S390 Foreign Function Interface
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 AUTHOR 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.
----------------------------------------------------------------------- */
/*====================================================================*/
/* Includes */
/* -------- */
/*====================================================================*/
#include <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
#include <stdio.h>
/*====================== End of Includes =============================*/
/*====================================================================*/
/* Defines */
/* ------- */
/*====================================================================*/
#define MAX_GPRARGS 5 /* Max. no. of GPR available */
#define MAX_FPRARGS 2 /* Max. no. of FPR available */
#define STR_GPR 1 /* Structure will fit in 1 or 2 GPR */
#define STR_FPR 2 /* Structure will fit in a FPR */
#define STR_STACK 3 /* Structure needs to go on stack */
/*===================== End of Defines ===============================*/
/*====================================================================*/
/* Types */
/* ----- */
/*====================================================================*/
typedef struct stackLayout
{
int *backChain;
int *endOfStack;
int glue[2];
int scratch[2];
int gprArgs[MAX_GPRARGS];
int notUsed;
union
{
float f;
double d;
} fprArgs[MAX_FPRARGS];
int unUsed[8];
int outArgs[100];
} stackLayout;
/*======================== End of Types ==============================*/
/*====================================================================*/
/* Prototypes */
/* ---------- */
/*====================================================================*/
void ffi_prep_args(stackLayout *, extended_cif *);
static int ffi_check_struct(ffi_type *, unsigned int *);
static void ffi_insert_int(int, stackLayout *, int *, int *);
static void ffi_insert_int64(long long, stackLayout *, int *, int *);
static void ffi_insert_double(double, stackLayout *, int *, int *);
/*====================== End of Prototypes ===========================*/
/*====================================================================*/
/* Externals */
/* --------- */
/*====================================================================*/
extern void ffi_call_SYSV(void (*)(stackLayout *, extended_cif *),
extended_cif *,
unsigned, unsigned,
unsigned *,
void (*fn)());
/*====================== End of Externals ============================*/
/*====================================================================*/
/* */
/* Name - ffi_check_struct. */
/* */
/* Function - Determine if a structure can be passed within a */
/* general or floating point register. */
/* */
/*====================================================================*/
int
ffi_check_struct(ffi_type *arg, unsigned int *strFlags)
{
ffi_type *element;
int i_Element;
for (i_Element = 0; arg->elements[i_Element]; i_Element++) {
element = arg->elements[i_Element];
switch (element->type) {
case FFI_TYPE_DOUBLE :
*strFlags |= STR_FPR;
break;
case FFI_TYPE_STRUCT :
*strFlags |= ffi_check_struct(element, strFlags);
break;
default :
*strFlags |= STR_GPR;
}
}
return (*strFlags);
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_insert_int. */
/* */
/* Function - Insert an integer parameter in a register if there are */
/* spares else on the stack. */
/* */
/*====================================================================*/
void
ffi_insert_int(int gprValue, stackLayout *stack,
int *intArgC, int *outArgC)
{
if (*intArgC < MAX_GPRARGS) {
stack->gprArgs[*intArgC] = gprValue;
*intArgC += 1;
}
else {
stack->outArgs[*outArgC++] = gprValue;
*outArgC += 1;
}
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_insert_int64. */
/* */
/* Function - Insert a long long parameter in registers if there are */
/* spares else on the stack. */
/* */
/*====================================================================*/
void
ffi_insert_int64(long long llngValue, stackLayout *stack,
int *intArgC, int *outArgC)
{
if (*intArgC < (MAX_GPRARGS-1)) {
memcpy(&stack->gprArgs[*intArgC],
&llngValue, sizeof(long long));
*intArgC += 2;
}
else {
memcpy(&stack->outArgs[*outArgC],
&llngValue, sizeof(long long));
*outArgC += 2;
}
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_insert_double. */
/* */
/* Function - Insert a double parameter in a FP register if there is */
/* a spare else on the stack. */
/* */
/*====================================================================*/
void
ffi_insert_double(double dblValue, stackLayout *stack,
int *fprArgC, int *outArgC)
{
if (*fprArgC < MAX_FPRARGS) {
stack->fprArgs[*fprArgC].d = dblValue;
*fprArgC += 1;
}
else {
memcpy(&stack->outArgs[*outArgC],
&dblValue,sizeof(double));
*outArgC += 2;
}
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_prep_args. */
/* */
/* Function - Prepare parameters for call to function. */
/* */
/* ffi_prep_args is called by the assembly routine once stack space */
/* has been allocated for the function's arguments. */
/* */
/* The stack layout we want looks like this: */
/* *------------------------------------------------------------* */
/* | 0 | Back chain (a 0 here signifies end of back chain) | */
/* +--------+---------------------------------------------------+ */
/* | 4 | EOS (end of stack, not used on Linux for S390) | */
/* +--------+---------------------------------------------------+ */
/* | 8 | Glue used in other linkage formats | */
/* +--------+---------------------------------------------------+ */
/* | 12 | Glue used in other linkage formats | */
/* +--------+---------------------------------------------------+ */
/* | 16 | Scratch area | */
/* +--------+---------------------------------------------------+ */
/* | 20 | Scratch area | */
/* +--------+---------------------------------------------------+ */
/* | 24 | GPR parameter register 1 | */
/* +--------+---------------------------------------------------+ */
/* | 28 | GPR parameter register 2 | */
/* +--------+---------------------------------------------------+ */
/* | 32 | GPR parameter register 3 | */
/* +--------+---------------------------------------------------+ */
/* | 36 | GPR parameter register 4 | */
/* +--------+---------------------------------------------------+ */
/* | 40 | GPR parameter register 5 | */
/* +--------+---------------------------------------------------+ */
/* | 44 | Unused | */
/* +--------+---------------------------------------------------+ */
/* | 48 | FPR parameter register 1 | */
/* +--------+---------------------------------------------------+ */
/* | 56 | FPR parameter register 2 | */
/* +--------+---------------------------------------------------+ */
/* | 64 | Unused | */
/* +--------+---------------------------------------------------+ */
/* | 96 | Outgoing args (length x) | */
/* +--------+---------------------------------------------------+ */
/* | 96+x | Copy area for structures (length y) | */
/* +--------+---------------------------------------------------+ */
/* | 96+x+y | Possible stack alignment | */
/* *------------------------------------------------------------* */
/* */
/*====================================================================*/
void
ffi_prep_args(stackLayout *stack, extended_cif *ecif)
{
const unsigned bytes = ecif->cif->bytes;
const unsigned flags = ecif->cif->flags;
/*----------------------------------------------------------*/
/* Pointer to the copy area on stack for structures */
/*----------------------------------------------------------*/
char *copySpace = (char *) stack + bytes + sizeof(stackLayout);
/*----------------------------------------------------------*/
/* Count of general and floating point register usage */
/*----------------------------------------------------------*/
int intArgC = 0,
fprArgC = 0,
outArgC = 0;
int i;
ffi_type **ptr;
void **p_argv;
size_t structCopySize;
unsigned gprValue, strFlags = 0;
unsigned long long llngValue;
double dblValue;
/* Now for the arguments. */
p_argv = ecif->avalue;
/*----------------------------------------------------------------------*/
/* If we returning a structure then we set the first parameter register */
/* to the address of where we are returning this structure */
/*----------------------------------------------------------------------*/
if (flags == FFI_TYPE_STRUCT)
stack->gprArgs[intArgC++] = (int) ecif->rvalue;
for (ptr = ecif->cif->arg_types, i = ecif->cif->nargs;
i > 0;
i--, ptr++, p_argv++)
{
switch ((*ptr)->type) {
case FFI_TYPE_FLOAT:
if (fprArgC < MAX_FPRARGS)
stack->fprArgs[fprArgC++].f = *(float *) *p_argv;
else
stack->outArgs[outArgC++] = *(int *) *p_argv;
break;
case FFI_TYPE_DOUBLE:
dblValue = *(double *) *p_argv;
ffi_insert_double(dblValue, stack, &fprArgC, &outArgC);
break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
llngValue = *(unsigned long long *) *p_argv;
ffi_insert_int64(llngValue, stack, &intArgC, &outArgC);
break;
case FFI_TYPE_UINT8:
gprValue = *(unsigned char *)*p_argv;
ffi_insert_int(gprValue, stack, &intArgC, &outArgC);
break;
case FFI_TYPE_SINT8:
gprValue = *(signed char *)*p_argv;
ffi_insert_int(gprValue, stack, &intArgC, &outArgC);
break;
case FFI_TYPE_UINT16:
gprValue = *(unsigned short *)*p_argv;
ffi_insert_int(gprValue, stack, &intArgC, &outArgC);
break;
case FFI_TYPE_SINT16:
gprValue = *(signed short *)*p_argv;
ffi_insert_int(gprValue, stack, &intArgC, &outArgC);
break;
case FFI_TYPE_STRUCT:
/*--------------------------------------------------*/
/* If structure > 8 bytes then it goes on the stack */
/*--------------------------------------------------*/
if (((*ptr)->size > 8) ||
((*ptr)->size > 4 &&
(*ptr)->size < 8))
strFlags = STR_STACK;
else
strFlags = ffi_check_struct((ffi_type *) *ptr, &strFlags);
switch (strFlags) {
/*-------------------------------------------*/
/* Structure that will fit in one or two GPR */
/*-------------------------------------------*/
case STR_GPR :
if ((*ptr)->size <= 4) {
gprValue = *(unsigned int *) *p_argv;
gprValue = gprValue >> ((4 - (*ptr)->size) * 8);
ffi_insert_int(gprValue, stack, &intArgC, &outArgC);
}
else {
llngValue = *(unsigned long long *) *p_argv;
ffi_insert_int64(llngValue, stack, &intArgC, &outArgC);
}
break;
/*-------------------------------------------*/
/* Structure that will fit in one FPR */
/*-------------------------------------------*/
case STR_FPR :
dblValue = *(double *) *p_argv;
ffi_insert_double(dblValue, stack, &fprArgC, &outArgC);
break;
/*-------------------------------------------*/
/* Structure that must be copied to stack */
/*-------------------------------------------*/
default :
structCopySize = (((*ptr)->size + 15) & ~0xF);
copySpace -= structCopySize;
memcpy(copySpace, (char *)*p_argv, (*ptr)->size);
gprValue = (unsigned) copySpace;
if (intArgC < MAX_GPRARGS)
stack->gprArgs[intArgC++] = gprValue;
else
stack->outArgs[outArgC++] = gprValue;
}
break;
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
structCopySize = (((*ptr)->size + 15) & ~0xF);
copySpace -= structCopySize;
memcpy(copySpace, (char *)*p_argv, (*ptr)->size);
gprValue = (unsigned) copySpace;
if (intArgC < MAX_GPRARGS)
stack->gprArgs[intArgC++] = gprValue;
else
stack->outArgs[outArgC++] = gprValue;
break;
#endif
case FFI_TYPE_INT:
case FFI_TYPE_UINT32:
case FFI_TYPE_SINT32:
case FFI_TYPE_POINTER:
gprValue = *(unsigned *)*p_argv;
if (intArgC < MAX_GPRARGS)
stack->gprArgs[intArgC++] = gprValue;
else
stack->outArgs[outArgC++] = gprValue;
break;
}
}
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_prep_cif_machdep. */
/* */
/* Function - Perform machine dependent CIF processing. */
/* */
/*====================================================================*/
ffi_status
ffi_prep_cif_machdep(ffi_cif *cif)
{
int i;
ffi_type **ptr;
unsigned bytes;
int fpArgC = 0,
intArgC = 0;
unsigned flags = 0;
unsigned structCopySize = 0;
/*-----------------------------------------------------------------*/
/* Extra space required in stack for overflow parameters. */
/*-----------------------------------------------------------------*/
bytes = 0;
/*--------------------------------------------------------*/
/* Return value handling. The rules are as follows: */
/* - 32-bit (or less) integer values are returned in gpr2 */
/* - Structures are returned as pointers in gpr2 */
/* - 64-bit integer values are returned in gpr2 and 3 */
/* - Single/double FP values are returned in fpr0 */
/*--------------------------------------------------------*/
flags = cif->rtype->type;
/*------------------------------------------------------------------------*/
/* The first MAX_GPRARGS words of integer arguments, and the */
/* first MAX_FPRARGS floating point arguments, go in registers; the rest */
/* goes on the stack. Structures and long doubles (if not equivalent */
/* to double) are passed as a pointer to a copy of the structure. */
/* Stuff on the stack needs to keep proper alignment. */
/*------------------------------------------------------------------------*/
for (ptr = cif->arg_types, i = cif->nargs; i > 0; i--, ptr++)
{
switch ((*ptr)->type)
{
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
fpArgC++;
if (fpArgC > MAX_FPRARGS && intArgC%2 != 0)
intArgC++;
break;
case FFI_TYPE_UINT64:
case FFI_TYPE_SINT64:
/*----------------------------------------------------*/
/* 'long long' arguments are passed as two words, but */
/* either both words must fit in registers or both go */
/* on the stack. If they go on the stack, they must */
/* be 8-byte-aligned. */
/*----------------------------------------------------*/
if ((intArgC == MAX_GPRARGS-1) ||
(intArgC >= MAX_GPRARGS) &&
(intArgC%2 != 0))
intArgC++;
intArgC += 2;
break;
case FFI_TYPE_STRUCT:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
/*----------------------------------------------------*/
/* We must allocate space for a copy of these to */
/* enforce pass-by-value. Pad the space up to a */
/* multiple of 16 bytes (the maximum alignment */
/* required for anything under the SYSV ABI). */
/*----------------------------------------------------*/
structCopySize += ((*ptr)->size + 15) & ~0xF;
/*----------------------------------------------------*/
/* Fall through (allocate space for the pointer). */
/*----------------------------------------------------*/
default:
/*----------------------------------------------------*/
/* Everything else is passed as a 4-byte word in a */
/* GPR either the object itself or a pointer to it. */
/*----------------------------------------------------*/
intArgC++;
break;
}
}
/*-----------------------------------------------------------------*/
/* Stack space. */
/*-----------------------------------------------------------------*/
if (intArgC > MAX_GPRARGS)
bytes += (intArgC - MAX_GPRARGS) * sizeof(int);
if (fpArgC > MAX_FPRARGS)
bytes += (fpArgC - MAX_FPRARGS) * sizeof(double);
/*-----------------------------------------------------------------*/
/* The stack space allocated needs to be a multiple of 16 bytes. */
/*-----------------------------------------------------------------*/
bytes = (bytes + 15) & ~0xF;
/*-----------------------------------------------------------------*/
/* Add in the space for the copied structures. */
/*-----------------------------------------------------------------*/
bytes += structCopySize;
cif->flags = flags;
cif->bytes = bytes;
return FFI_OK;
}
/*======================== End of Routine ============================*/
/*====================================================================*/
/* */
/* Name - ffi_call. */
/* */
/* Function - Call the FFI routine. */
/* */
/*====================================================================*/
void
ffi_call(ffi_cif *cif,
void (*fn)(),
void *rvalue,
void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/*-----------------------------------------------------------------*/
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
/*-----------------------------------------------------------------*/
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
ecif.rvalue = alloca(cif->rtype->size);
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
ffi_call_SYSV(ffi_prep_args,
&ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
break;
default:
FFI_ASSERT(0);
break;
}
}
/*======================== End of Routine ============================*/
+161
View File
@@ -0,0 +1,161 @@
/* -----------------------------------------------------------------------
sysv.S - Copyright (c) 2000 Software AG
S390 Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#ifdef HAVE_MACHINE_ASM_H
#include <machine/asm.h>
#endif
.text
# r2: ffi_prep_args
# r3: &ecif
# r4: cif->bytes
# r5: fig->flags
# r6: ecif.rvalue
# sp+0: fn
# This assumes we are using gas.
.globl ffi_call_SYSV
.type ffi_call_SYSV,%function
ffi_call_SYSV:
# Save registers
stm %r7,%r15,28(%r15)
l %r7,96(%r15) # Get A(fn)
lr %r0,%r15
ahi %r15,-128 # Make room for my args
st %r0,0(%r15) # Set backchain
lr %r11,%r15 # Establish my stack register
sr %r15,%r4 # Make room for fn args
ahi %r15,-96 # Make room for new frame
lr %r10,%r15 # Establish stack build area
ahi %r15,-96 # Stack for next call
lr %r1,%r7
stm %r2,%r7,96(%r11) # Save args on my stack
#------------------------------------------------------------------
# move first 3 parameters in registers
#------------------------------------------------------------------
lr %r9,%r2 # r9: &ffi_prep_args
lr %r2,%r10 # Parm 1: &stack Parm 2: &ecif
basr %r14,%r9 # call ffi_prep_args
#------------------------------------------------------------------
# load first 5 parameter registers
#------------------------------------------------------------------
lm %r2,%r6,24(%r10)
#------------------------------------------------------------------
# load fp parameter registers
#------------------------------------------------------------------
ld %f0,48(%r10)
ld %f2,56(%r10)
#------------------------------------------------------------------
# call function
#------------------------------------------------------------------
lr %r15,%r10 # Set new stack
l %r9,116(%r11) # Get &fn
basr %r14,%r9 # Call function
#------------------------------------------------------------------
# On return:
# r2: Return value (r3: Return value + 4 for long long)
#------------------------------------------------------------------
#------------------------------------------------------------------
# If the return value pointer is NULL, assume no return value.
#------------------------------------------------------------------
icm %r6,15,112(%r11)
jz .Lepilogue
l %r5,108(%r11) # Get return type
#------------------------------------------------------------------
# return INT
#------------------------------------------------------------------
chi %r5,FFI_TYPE_INT
jne .Lchk64
st %r2,0(%r6)
j .Lepilogue
.Lchk64:
#------------------------------------------------------------------
# return LONG LONG (signed/unsigned)
#------------------------------------------------------------------
chi %r5,FFI_TYPE_UINT64
je .LdoLongLong
chi %r5,FFI_TYPE_SINT64
jne .LchkFloat
.LdoLongLong:
stm %r2,%r3,0(%r6)
j .Lepilogue
.LchkFloat:
#------------------------------------------------------------------
# return FLOAT
#------------------------------------------------------------------
chi %r5,FFI_TYPE_FLOAT
jne .LchkDouble
std %f0,0(%r6)
j .Lepilogue
.LchkDouble:
#------------------------------------------------------------------
# return DOUBLE or LONGDOUBLE
#------------------------------------------------------------------
chi %r5,FFI_TYPE_DOUBLE
jne .LchkStruct
std %f0,0(%r6)
std %f2,8(%r6)
j .Lepilogue
.LchkStruct:
#------------------------------------------------------------------
# Structure - rvalue already set as sent as 1st parm to routine
#------------------------------------------------------------------
chi %r5,FFI_TYPE_STRUCT
je .Lepilogue
.Ldefault:
#------------------------------------------------------------------
# return a pointer
#------------------------------------------------------------------
st %r2,0(%r6)
j .Lepilogue
.Lepilogue:
l %r15,0(%r11)
l %r4,56(%r15)
lm %r7,%r15,28(%r15)
br %r4
.ffi_call_SYSV_end:
.size ffi_call_SYSV,.ffi_call_SYSV_end-ffi_call_SYSV
+422
View File
@@ -0,0 +1,422 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1996 Cygnus Solutions
Sparc Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments */
void ffi_prep_args_v8(char *stack, extended_cif *ecif)
{
int i;
int tmp;
int avn;
void **p_argv;
char *argp;
ffi_type **p_arg;
tmp = 0;
/* Skip 16 words for the window save area */
argp = stack + 16*sizeof(int);
/* This should only really be done when we are returning a structure,
however, it's faster just to do it all the time...
if ( ecif->cif->rtype->type == FFI_TYPE_STRUCT ) */
*(int *) argp = (long)ecif->rvalue;
/* And 1 word for the structure return value. */
argp += sizeof(int);
#ifdef USING_PURIFY
/* Purify will probably complain in our assembly routine, unless we
zero out this memory. */
((int*)argp)[0] = 0;
((int*)argp)[1] = 0;
((int*)argp)[2] = 0;
((int*)argp)[3] = 0;
((int*)argp)[4] = 0;
((int*)argp)[5] = 0;
#endif
avn = ecif->cif->nargs;
p_argv = ecif->avalue;
for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types;
i && avn;
i--, p_arg++)
{
size_t z;
if (avn)
{
avn--;
if ((*p_arg)->type == FFI_TYPE_STRUCT
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
|| (*p_arg)->type == FFI_TYPE_LONGDOUBLE
#endif
)
{
*(unsigned int *) argp = (unsigned long)(* p_argv);
z = sizeof(int);
}
else
{
z = (*p_arg)->size;
if (z < sizeof(int))
{
z = sizeof(int);
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(signed int *) argp = *(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
*(unsigned int *) argp = *(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
*(signed int *) argp = *(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
*(unsigned int *) argp = *(UINT16 *)(* p_argv);
break;
default:
FFI_ASSERT(0);
}
}
else
{
memcpy(argp, *p_argv, z);
}
}
p_argv++;
argp += z;
}
}
return;
}
int ffi_prep_args_v9(char *stack, extended_cif *ecif)
{
int i, ret = 0;
int tmp;
void **p_argv;
char *argp;
ffi_type **p_arg;
tmp = 0;
/* Skip 16 words for the window save area */
argp = stack + 16*sizeof(long long);
#ifdef USING_PURIFY
/* Purify will probably complain in our assembly routine, unless we
zero out this memory. */
((long long*)argp)[0] = 0;
((long long*)argp)[1] = 0;
((long long*)argp)[2] = 0;
((long long*)argp)[3] = 0;
((long long*)argp)[4] = 0;
((long long*)argp)[5] = 0;
#endif
p_argv = ecif->avalue;
if (ecif->cif->rtype->type == FFI_TYPE_STRUCT &&
ecif->cif->rtype->size > 32)
{
*(unsigned long long *) argp = (unsigned long)ecif->rvalue;
tmp = 1;
}
for (i = 0, p_arg = ecif->cif->arg_types; i < ecif->cif->nargs;
i++, p_arg++)
{
size_t z;
z = (*p_arg)->size;
switch ((*p_arg)->type)
{
case FFI_TYPE_STRUCT:
if (z > 16)
{
/* For structures larger than 16 bytes we pass reference. */
*(unsigned long long *) argp = (unsigned long)* p_argv;
argp += sizeof(long long);
tmp++;
p_argv++;
continue;
}
/* FALLTHROUGH */
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
ret = 1; /* We should promote into FP regs as well as integer. */
break;
}
if (z < sizeof(long long))
{
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(signed long long *) argp = *(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
*(unsigned long long *) argp = *(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
*(signed long long *) argp = *(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
*(unsigned long long *) argp = *(UINT16 *)(* p_argv);
break;
case FFI_TYPE_SINT32:
*(signed long long *) argp = *(SINT32 *)(* p_argv);
break;
case FFI_TYPE_UINT32:
*(unsigned long long *) argp = *(UINT32 *)(* p_argv);
break;
case FFI_TYPE_FLOAT:
*(float *) (argp + 4) = *(FLOAT32 *)(* p_argv); /* Right justify */
break;
case FFI_TYPE_STRUCT:
memcpy(argp, *p_argv, z);
break;
default:
FFI_ASSERT(0);
}
z = sizeof(long long);
tmp++;
}
else if (z == sizeof(long long))
{
memcpy(argp, *p_argv, z);
z = sizeof(long long);
tmp++;
}
else
{
if ((tmp & 1) && (*p_arg)->alignment > 8)
{
tmp++;
argp += sizeof(long long);
}
memcpy(argp, *p_argv, z);
z = 2 * sizeof(long long);
tmp += 2;
}
p_argv++;
argp += z;
}
return ret;
}
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
{
int wordsize;
if (cif->abi != FFI_V9)
{
wordsize = 4;
/* If we are returning a struct, this will already have been added.
Otherwise we need to add it because it's always got to be there! */
if (cif->rtype->type != FFI_TYPE_STRUCT)
cif->bytes += wordsize;
/* sparc call frames require that space is allocated for 6 args,
even if they aren't used. Make that space if necessary. */
if (cif->bytes < 4*6+4)
cif->bytes = 4*6+4;
}
else
{
wordsize = 8;
/* sparc call frames require that space is allocated for 6 args,
even if they aren't used. Make that space if necessary. */
if (cif->bytes < 8*6)
cif->bytes = 8*6;
}
/* Adjust cif->bytes. to include 16 words for the window save area,
and maybe the struct/union return pointer area, */
cif->bytes += 16 * wordsize;
/* The stack must be 2 word aligned, so round bytes up
appropriately. */
cif->bytes = ALIGN(cif->bytes, 2 * wordsize);
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
cif->flags = cif->rtype->type;
break;
case FFI_TYPE_STRUCT:
if (cif->abi == FFI_V9 && cif->rtype->size > 32)
cif->flags = FFI_TYPE_VOID;
else
cif->flags = FFI_TYPE_STRUCT;
break;
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
if (cif->abi != FFI_V9)
{
cif->flags = FFI_TYPE_SINT64;
break;
}
/* FALLTHROUGH */
default:
cif->flags = FFI_TYPE_INT;
break;
}
return FFI_OK;
}
int ffi_V9_return_struct(ffi_type *arg, int off, char *ret, char *intg, char *flt)
{
ffi_type **ptr = &arg->elements[0];
while (*ptr != NULL)
{
if (off & ((*ptr)->alignment - 1))
off = ALIGN(off, (*ptr)->alignment);
switch ((*ptr)->type)
{
case FFI_TYPE_STRUCT:
off = ffi_V9_return_struct(*ptr, off, ret, intg, flt);
break;
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
#if FFI_TYPE_LONGDOUBLE != FFI_TYPE_DOUBLE
case FFI_TYPE_LONGDOUBLE:
#endif
memcpy(ret + off, flt + off, (*ptr)->size);
off += (*ptr)->size;
break;
default:
memcpy(ret + off, intg + off, (*ptr)->size);
off += (*ptr)->size;
break;
}
ptr++;
}
return off;
}
extern int ffi_call_V8(void *, extended_cif *, unsigned,
unsigned, unsigned *, void (*fn)());
extern int ffi_call_V9(void *, extended_cif *, unsigned,
unsigned, unsigned *, void (*fn)());
void ffi_call(ffi_cif *cif, void (*fn)(), void *rvalue, void **avalue)
{
extended_cif ecif;
void *rval = rvalue;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
ecif.rvalue = rvalue;
if (cif->rtype->type == FFI_TYPE_STRUCT)
{
if (cif->rtype->size <= 32)
rval = alloca(64);
else
{
rval = NULL;
if (rvalue == NULL)
ecif.rvalue = alloca(cif->rtype->size);
}
}
switch (cif->abi)
{
case FFI_V8:
#ifdef SPARC64
/* We don't yet support calling 32bit code from 64bit */
FFI_ASSERT(0);
#else
ffi_call_V8(ffi_prep_args_v8, &ecif, cif->bytes,
cif->flags, rvalue, fn);
#endif
break;
case FFI_V9:
#ifdef SPARC64
ffi_call_V9(ffi_prep_args_v9, &ecif, cif->bytes,
cif->flags, rval, fn);
if (rvalue && rval && cif->rtype->type == FFI_TYPE_STRUCT)
ffi_V9_return_struct(cif->rtype, 0, (char *)rvalue, (char *)rval, ((char *)rval)+32);
#else
/* And vice versa */
FFI_ASSERT(0);
#endif
break;
default:
FFI_ASSERT(0);
break;
}
}
+94
View File
@@ -0,0 +1,94 @@
/* -----------------------------------------------------------------------
v8.S - Copyright (c) 1996, 1997 Cygnus Solutions
Sparc Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
#define STACKFRAME 96 /* Minimum stack framesize for SPARC */
#define ARGS (64+4) /* Offset of register area in frame */
.text
.align 8
.globl ffi_call_V8
.globl _ffi_call_V8
ffi_call_V8:
_ffi_call_V8:
save %sp, -STACKFRAME, %sp
sub %sp, %i2, %sp ! alloca() space in stack for frame to set up
add %sp, STACKFRAME, %l0 ! %l0 has start of
! frame to set up
mov %l0, %o0 ! call routine to set up frame
call %i0
mov %i1, %o1 ! (delay)
ld [%l0+ARGS], %o0 ! call foreign function
ld [%l0+ARGS+4], %o1
ld [%l0+ARGS+8], %o2
ld [%l0+ARGS+12], %o3
ld [%l0+ARGS+16], %o4
ld [%l0+ARGS+20], %o5
call %i5
mov %l0, %sp ! (delay) switch to frame
nop ! STRUCT returning functions skip 12 instead of 8 bytes
! If the return value pointer is NULL, assume no return value.
tst %i4
bz done
nop
cmp %i3, FFI_TYPE_INT
be,a done
st %o0, [%i4] ! (delay)
cmp %i3, FFI_TYPE_FLOAT
be,a done
st %f0, [%i4+0] ! (delay)
cmp %i3, FFI_TYPE_SINT64
be longlong
cmp %i3, FFI_TYPE_DOUBLE
bne done
nop
st %f0, [%i4+0]
st %f1, [%i4+4]
done:
ret
restore
longlong:
st %o0, [%i4+0]
st %o1, [%i4+4]
ret
restore
.ffi_call_V8_end:
.size ffi_call_V8,.ffi_call_V8_end-ffi_call_V8
+126
View File
@@ -0,0 +1,126 @@
/* -----------------------------------------------------------------------
v9.S - Copyright (c) 2000 Cygnus Solutions
Sparc 64bit Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
#if defined(__arch64__) || defined(__sparcv9)
/* Only compile this in for 64bit builds, because otherwise the object file
will have inproper architecture due to used instructions. */
#define STACKFRAME 128 /* Minimum stack framesize for SPARC */
#define STACK_BIAS 2047
#define ARGS (128) /* Offset of register area in frame */
.text
.align 8
.globl ffi_call_V9
.globl _ffi_call_V9
ffi_call_V9:
_ffi_call_V9:
save %sp, -STACKFRAME, %sp
sub %sp, %i2, %sp ! alloca() space in stack for frame to set up
add %sp, STACKFRAME+STACK_BIAS, %l0 ! %l0 has start of
! frame to set up
mov %l0, %o0 ! call routine to set up frame
call %i0
mov %i1, %o1 ! (delay)
brz,pt %o0, 1f
ldx [%l0+ARGS], %o0 ! call foreign function
ldd [%l0+ARGS], %f0
ldd [%l0+ARGS+8], %f2
ldd [%l0+ARGS+16], %f4
ldd [%l0+ARGS+24], %f6
ldd [%l0+ARGS+32], %f8
ldd [%l0+ARGS+40], %f10
ldd [%l0+ARGS+48], %f12
ldd [%l0+ARGS+56], %f14
ldd [%l0+ARGS+64], %f16
ldd [%l0+ARGS+72], %f18
ldd [%l0+ARGS+80], %f20
ldd [%l0+ARGS+88], %f22
ldd [%l0+ARGS+96], %f24
ldd [%l0+ARGS+104], %f26
ldd [%l0+ARGS+112], %f28
ldd [%l0+ARGS+120], %f30
1: ldx [%l0+ARGS+8], %o1
ldx [%l0+ARGS+16], %o2
ldx [%l0+ARGS+24], %o3
ldx [%l0+ARGS+32], %o4
ldx [%l0+ARGS+40], %o5
call %i5
sub %l0, STACK_BIAS, %sp ! (delay) switch to frame
! If the return value pointer is NULL, assume no return value.
brz,pn %i4, done
nop
cmp %i3, FFI_TYPE_INT
be,a,pt %icc, done
stx %o0, [%i4] ! (delay)
cmp %i3, FFI_TYPE_FLOAT
be,a,pn %icc, done
st %f0, [%i4+0] ! (delay)
cmp %i3, FFI_TYPE_DOUBLE
be,a,pn %icc, done
std %f0, [%i4+0] ! (delay)
cmp %i3, FFI_TYPE_STRUCT
be,pn %icc, dostruct
cmp %i3, FFI_TYPE_LONGDOUBLE
bne,pt %icc, done
nop
std %f0, [%i4+0]
std %f2, [%i4+8]
done: ret
restore
dostruct:
/* This will not work correctly for unions. */
stx %o0, [%i4+0]
stx %o1, [%i4+8]
stx %o2, [%i4+16]
stx %o3, [%i4+24]
std %f0, [%i4+32]
std %f2, [%i4+40]
std %f4, [%i4+48]
std %f6, [%i4+56]
ret
restore
.ffi_call_V9_end:
.size ffi_call_V9,.ffi_call_V9_end-ffi_call_V9
#endif
File renamed without changes.
+15
View File
@@ -0,0 +1,15 @@
## Process this file with automake to produce Makefile.in.
AUTOMAKE_OPTIONS = foreign dejagnu
# Setup the testing framework, if you have one
EXPECT = `if [ -f $(top_builddir)/../expect/expect ] ; then \
echo $(top_builddir)/../expect/expect ; \
else echo expect ; fi`
RUNTEST = `if [ -f $(top_srcdir)/../dejagnu/runtest ] ; then \
echo $(top_srcdir)/../dejagnu/runtest ; \
else echo runtest; fi`
RUNTESTFLAGS = @AM_RUNTESTFLAGS@
+1
View File
@@ -0,0 +1 @@
load_lib "standard.exp"
+44
View File
@@ -0,0 +1,44 @@
# Copyright (C) 2001 Red Hat, Inc.
load_lib "libgloss.exp"
global tmpdir
if ![info exists tmpdir] {
set tmpdir "/tmp"
}
proc test_libffi {src} {
global tmpdir srcdir
set executable $tmpdir/[file tail [file rootname $src].x]
regsub "^$srcdir/?" $src "" testcase
# If we couldn't rip $srcdir out of `src' then just do the best we can.
# The point is to reduce the unnecessary noise in the logs. Don't strip
# out too much because different testcases with the same name can confuse
# `test-tool'.
if [string match "/*" $testcase] {
set testcase "[file tail [file dirname $src]]/[file tail $src]"
}
remote_file build delete $executable;
verbose "Testing $testcase" 1
set comp_output [target_compile $src $executable executable "additional_flags=-g additional_flags=-I$srcdir/../include libs=../.libs/libffi.a"]
set result [libffi_load "$executable" "" ""]
set status [lindex $result 0];
set output [lindex $result 1];
$status "$testcase"
# if { $status == "pass" } {
# remote_file build delete $executable;
# }
}
# Local Variables:
# tcl-indent-level:4
# End:
+13
View File
@@ -0,0 +1,13 @@
global srcdir subdir
catch "glob -nocomplain ${srcdir}/${subdir}/*.c" srcfiles
verbose "srcfiles are $srcfiles"
set prefix ""
foreach x $srcfiles {
test_libffi $x
}
# Local Variables:
# tcl-indent-level:4
# End:
+4
View File
@@ -0,0 +1,4 @@
#include <ffi.h>
#define CHECK(x) !(x) ? abort() : 0
+55
View File
@@ -0,0 +1,55 @@
#include "ffitest.h"
static int floating(int a, float b, double c, long double d, int e)
{
int i;
i = (int) ((float)a/b + ((float)c/(float)d));
return i;
}
int
main ()
{
ffi_cif cif;
ffi_type *args[5];
void *values[5];
int si1, si2;
float f;
double d;
long double ld;
int rint __attribute__((aligned(8)));
args[0] = &ffi_type_sint;
values[0] = &si1;
args[1] = &ffi_type_float;
values[1] = &f;
args[2] = &ffi_type_double;
values[2] = &d;
args[3] = &ffi_type_longdouble;
values[3] = &ld;
args[4] = &ffi_type_sint;
values[4] = &si2;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 5,
&ffi_type_sint, args) == FFI_OK);
si1 = 6;
f = 3.14159;
d = (double)1.0/(double)3.0;
ld = 2.71828182846L;
si2 = 10;
floating (si1, f, d, ld, si2);
ffi_call(&cif, FFI_FN(floating), &rint, values);
printf ("%d vs %d\n", rint, floating (si1, f, d, ld, si2));
CHECK(rint == floating(si1, f, d, ld, si2));
exit (0);
}
+63
View File
@@ -0,0 +1,63 @@
#include "ffitest.h"
#include <float.h>
static float many(float f1,
float f2,
float f3,
float f4,
float f5,
float f6,
float f7,
float f8,
float f9,
float f10,
float f11,
float f12,
float f13)
{
#if 1
printf("%f %f %f %f %f %f %f %f %f %f %f %f %f\n",
(double) f1, (double) f2, (double) f3, (double) f4, (double) f5,
(double) f6, (double) f7, (double) f8, (double) f9, (double) f10,
(double) f11, (double) f12, (double) f13);
#endif
return ((f1/f2+f3/f4+f5/f6+f7/f8+f9/f10+f11/f12) * f13);
}
int
main ()
{
ffi_cif cif;
ffi_type *args[13];
void *values[13];
float fa[13];
float f, ff;
int i;
for (i = 0; i < 13; i++)
{
args[i] = &ffi_type_float;
values[i] = &fa[i];
fa[i] = (float) i;
}
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 13,
&ffi_type_float, args) == FFI_OK);
ffi_call(&cif, FFI_FN(many), &f, values);
ff = many(fa[0], fa[1],
fa[2], fa[3],
fa[4], fa[5],
fa[6], fa[7],
fa[8], fa[9],
fa[10],fa[11],fa[12]);
if (f - ff < FLT_EPSILON)
exit(0);
else
abort();
}
+38
View File
@@ -0,0 +1,38 @@
#include "ffitest.h"
static size_t my_strlen(char *s)
{
return (strlen(s));
}
int
main ()
{
ffi_cif cif;
ffi_type *args[1];
void *values[1];
int rint __attribute__((aligned(8)));
char *s;
args[0] = &ffi_type_pointer;
values[0] = (void*) &s;
/* Initialize the cif */
CHECK(ffi_prep_cif(&cif, FFI_DEFAULT_ABI, 1,
&ffi_type_sint, args) == FFI_OK);
s = "a";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 1);
s = "1234567";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 7);
s = "1234567890123456789012345";
ffi_call(&cif, FFI_FN(my_strlen), &rint, values);
CHECK(rint == 25);
exit (0);
}
+153
View File
@@ -0,0 +1,153 @@
/* -----------------------------------------------------------------------
types.c - Copyright (c) 1996, 1998 Cygnus Solutions
Predefined ffi_types needed by libffi.
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 RED HAT 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 <ffi.h>
#include <ffi_private.h>
/* Type definitions */
#define FFI_INTEGRAL_TYPEDEF(n, s, a, t) ffi_type ffi_type_##n = { s, a, t, NULL }
#define FFI_AGGREGATE_TYPEDEF(n, e) ffi_type ffi_type_##n = { 0, 0, FFI_TYPE_STRUCT, e }
/* Size and alignment are fake here. They must not be 0. */
FFI_INTEGRAL_TYPEDEF(void, 1, 1, FFI_TYPE_VOID);
FFI_INTEGRAL_TYPEDEF(uint8, 1, 1, FFI_TYPE_UINT8);
FFI_INTEGRAL_TYPEDEF(sint8, 1, 1, FFI_TYPE_SINT8);
FFI_INTEGRAL_TYPEDEF(uint16, 2, 2, FFI_TYPE_UINT16);
FFI_INTEGRAL_TYPEDEF(sint16, 2, 2, FFI_TYPE_SINT16);
FFI_INTEGRAL_TYPEDEF(uint32, 4, 4, FFI_TYPE_UINT32);
FFI_INTEGRAL_TYPEDEF(sint32, 4, 4, FFI_TYPE_SINT32);
FFI_INTEGRAL_TYPEDEF(float, 4, 4, FFI_TYPE_FLOAT);
#if defined __alpha__ || defined SPARC64
FFI_INTEGRAL_TYPEDEF(pointer, 8, 8, FFI_TYPE_POINTER);
#else
FFI_INTEGRAL_TYPEDEF(pointer, 4, 4, FFI_TYPE_POINTER);
#endif
#ifdef __i386__
FFI_INTEGRAL_TYPEDEF(uint64, 8, 4, FFI_TYPE_UINT64);
FFI_INTEGRAL_TYPEDEF(sint64, 8, 4, FFI_TYPE_SINT64);
#elif defined X86_WIN32
FFI_INTEGRAL_TYPEDEF(uint64, 8, 4, FFI_TYPE_UINT64);
FFI_INTEGRAL_TYPEDEF(sint64, 8, 4, FFI_TYPE_SINT64);
#elif defined ARM
FFI_INTEGRAL_TYPEDEF(uint64, 8, 4, FFI_TYPE_UINT64);
FFI_INTEGRAL_TYPEDEF(sint64, 8, 4, FFI_TYPE_SINT64);
#elif defined M68K
FFI_INTEGRAL_TYPEDEF(uint64, 8, 4, FFI_TYPE_UINT64);
FFI_INTEGRAL_TYPEDEF(sint64, 8, 4, FFI_TYPE_SINT64);
#else
FFI_INTEGRAL_TYPEDEF(uint64, 8, 8, FFI_TYPE_UINT64);
FFI_INTEGRAL_TYPEDEF(sint64, 8, 8, FFI_TYPE_SINT64);
#endif
#ifdef __i386__
FFI_INTEGRAL_TYPEDEF(double, 8, 4, FFI_TYPE_DOUBLE);
FFI_INTEGRAL_TYPEDEF(longdouble, 12, 4, FFI_TYPE_LONGDOUBLE);
#elif defined X86_WIN32
FFI_INTEGRAL_TYPEDEF(double, 8, 4, FFI_TYPE_DOUBLE);
FFI_INTEGRAL_TYPEDEF(longdouble, 12, 4, FFI_TYPE_LONGDOUBLE);
#elif defined ARM
FFI_INTEGRAL_TYPEDEF(double, 8, 4, FFI_TYPE_DOUBLE);
FFI_INTEGRAL_TYPEDEF(longdouble, 8, 4, FFI_TYPE_LONGDOUBLE);
#elif defined M68K
FFI_INTEGRAL_TYPEDEF(double, 8, 4, FFI_TYPE_DOUBLE);
FFI_INTEGRAL_TYPEDEF(longdouble, 12, 4, FFI_TYPE_LONGDOUBLE);
#elif defined SPARC
FFI_INTEGRAL_TYPEDEF(double, 8, 8, FFI_TYPE_DOUBLE);
#ifdef SPARC64
FFI_INTEGRAL_TYPEDEF(longdouble, 16, 16, FFI_TYPE_LONGDOUBLE);
#else
FFI_INTEGRAL_TYPEDEF(longdouble, 16, 8, FFI_TYPE_LONGDOUBLE);
#endif
#else
FFI_INTEGRAL_TYPEDEF(double, 8, 8, FFI_TYPE_DOUBLE);
FFI_INTEGRAL_TYPEDEF(longdouble, 8, 8, FFI_TYPE_LONGDOUBLE);
#endif
#if SIZEOF_SHORT==1
extern ffi_type ffi_type_sshort __attribute__((alias("ffi_type_sint8")));
extern ffi_type ffi_type_ushort __attribute__((alias("ffi_type_uint8")));
#elif SIZEOF_SHORT==2
extern ffi_type ffi_type_sshort __attribute__((alias("ffi_type_sint16")));
extern ffi_type ffi_type_ushort __attribute__((alias("ffi_type_uint16")));
#elif SIZEOF_SHORT==4
extern ffi_type ffi_type_sshort __attribute__((alias("ffi_type_sint32")));
extern ffi_type ffi_type_ushort __attribute__((alias("ffi_type_uint32")));
#endif
#if SIZEOF_INT==2
extern ffi_type ffi_type_sint __attribute__((alias("ffi_type_sint16")));
extern ffi_type ffi_type_uint __attribute__((alias("ffi_type_uint16")));
#elif SIZEOF_INT==4
extern ffi_type ffi_type_sint __attribute__((alias("ffi_type_sint32")));
extern ffi_type ffi_type_uint __attribute__((alias("ffi_type_uint32")));
#elif SIZEOF_INT==8
extern ffi_type ffi_type_sint __attribute__((alias("ffi_type_sint64")));
extern ffi_type ffi_type_uint __attribute__((alias("ffi_type_uint64")));
#endif
#if SIZEOF_LONG==4
extern ffi_type ffi_type_slong __attribute__((alias("ffi_type_sint32")));
extern ffi_type ffi_type_ulong __attribute__((alias("ffi_type_uint32")));
#elif SIZEOF_INT==8
extern ffi_type ffi_type_slong __attribute__((alias("ffi_type_sint64")));
extern ffi_type ffi_type_ulong __attribute__((alias("ffi_type_uint64")));
#endif
+508
View File
@@ -0,0 +1,508 @@
/* -----------------------------------------------------------------------
ffi.c - Copyright (c) 1996, 1998, 1999 Cygnus Solutions
x86 Foreign Function Interface
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 CYGNUS SOLUTIONS 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 <ffi.h>
#include <ffi_private.h>
#include <stdlib.h>
/* ffi_prep_args is called by the assembly routine once stack space
has been allocated for the function's arguments */
/*@-exportheader@*/
void ffi_prep_args(char *stack, extended_cif *ecif)
/*@=exportheader@*/
{
register unsigned int i;
register int tmp;
register unsigned int avn;
register void **p_argv;
register char *argp;
register ffi_type **p_arg;
tmp = 0;
argp = stack;
if ( ecif->cif->rtype->type == FFI_TYPE_STRUCT ) {
*(void **) argp = ecif->rvalue;
argp += 4;
}
avn = ecif->cif->nargs;
p_argv = ecif->avalue;
for (i = ecif->cif->nargs, p_arg = ecif->cif->arg_types;
(i != 0) && (avn != 0);
i--, p_arg++)
{
size_t z;
/* Align if necessary */
if (((*p_arg)->alignment - 1) & (unsigned) argp) {
argp = (char *) ALIGN(argp, (*p_arg)->alignment);
}
if (avn != 0)
{
avn--;
z = (*p_arg)->size;
if (z < sizeof(int))
{
z = sizeof(int);
switch ((*p_arg)->type)
{
case FFI_TYPE_SINT8:
*(signed int *) argp = (signed int)*(SINT8 *)(* p_argv);
break;
case FFI_TYPE_UINT8:
*(unsigned int *) argp = (unsigned int)*(UINT8 *)(* p_argv);
break;
case FFI_TYPE_SINT16:
*(signed int *) argp = (signed int)*(SINT16 *)(* p_argv);
break;
case FFI_TYPE_UINT16:
*(unsigned int *) argp = (unsigned int)*(UINT16 *)(* p_argv);
break;
case FFI_TYPE_SINT32:
*(signed int *) argp = (signed int)*(SINT32 *)(* p_argv);
break;
case FFI_TYPE_UINT32:
*(unsigned int *) argp = (unsigned int)*(UINT32 *)(* p_argv);
break;
case FFI_TYPE_STRUCT:
*(unsigned int *) argp = (unsigned int)*(UINT32 *)(* p_argv);
break;
default:
FFI_ASSERT(0);
}
}
else
{
memcpy(argp, *p_argv, z);
}
p_argv++;
argp += z;
}
}
return;
}
/* Perform machine dependent cif processing */
ffi_status ffi_prep_cif_machdep(ffi_cif *cif)
{
/* Set the return type flag */
switch (cif->rtype->type)
{
case FFI_TYPE_VOID:
case FFI_TYPE_STRUCT:
case FFI_TYPE_SINT64:
case FFI_TYPE_FLOAT:
case FFI_TYPE_DOUBLE:
case FFI_TYPE_LONGDOUBLE:
cif->flags = (unsigned) cif->rtype->type;
break;
case FFI_TYPE_UINT64:
cif->flags = FFI_TYPE_SINT64;
break;
default:
cif->flags = FFI_TYPE_INT;
break;
}
return FFI_OK;
}
/*@-declundef@*/
/*@-exportheader@*/
extern void ffi_call_SYSV(void (*)(char *, extended_cif *),
/*@out@*/ extended_cif *,
unsigned, unsigned,
/*@out@*/ unsigned *,
void (*fn)());
/*@=declundef@*/
/*@=exportheader@*/
void ffi_call(/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ void **avalue)
{
extended_cif ecif;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
{
/*@-sysunrecog@*/
ecif.rvalue = alloca(cif->rtype->size);
/*@=sysunrecog@*/
}
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
/*@-usedef@*/
ffi_call_SYSV(ffi_prep_args, &ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
/*@=usedef@*/
break;
default:
FFI_ASSERT(0);
break;
}
}
/** private members **/
static void ffi_prep_incoming_args_SYSV (char *stack, void **ret,
void** args, ffi_cif* cif);
static void ffi_closure_SYSV ();
static void ffi_closure_raw_SYSV ();
/* This function is jumped to by the trampoline, on entry, %ecx (a
* caller-save register) holds the address of the closure.
* Clearly, this requires __GNUC__, so perhaps we should translate this
* into an assembly file if this is to be distributed with ffi.
*/
static void
ffi_closure_SYSV ()
{
// this is our return value storage
long double res;
// our various things...
void *args;
ffi_cif *cif;
void **arg_area;
ffi_closure *closure;
unsigned short rtype;
void *resp = (void*)&res;
/* grab the trampoline context pointer */
asm ("movl %%ecx,%0" : "=r" (closure));
cif = closure->cif;
arg_area = (void**) alloca (cif->nargs * sizeof (void*));
asm ("leal 8(%%ebp),%0" : "=q" (args));
/* this call will initialize ARG_AREA, such that each
* element in that array points to the corresponding
* value on the stack; and if the function returns
* a structure, it will re-set RESP to point to the
* structure return address. */
ffi_prep_incoming_args_SYSV(args, (void**)&resp, arg_area, cif);
(closure->fun) (cif, resp, arg_area, closure->user_data);
rtype = cif->flags;
/* now, do a generic return based on the value of rtype */
if (rtype == FFI_TYPE_INT)
{
asm ("movl (%0),%%eax" : : "r" (resp) : "eax");
}
else if (rtype == FFI_TYPE_FLOAT)
{
asm ("flds (%0)" : : "r" (resp) : "st" );
}
else if (rtype == FFI_TYPE_DOUBLE)
{
asm ("fldl (%0)" : : "r" (resp) : "st", "st(1)" );
}
else if (rtype == FFI_TYPE_LONGDOUBLE)
{
asm ("fldt (%0)" : : "r" (resp) : "st", "st(1)" );
}
else if (rtype == FFI_TYPE_SINT64)
{
asm ("movl 0(%0),%%eax;"
"movl 4(%0),%%edx"
: : "r"(resp)
: "eax", "edx");
}
}
/*@-exportheader@*/
static void
ffi_prep_incoming_args_SYSV(char *stack, void **rvalue,
void **avalue, ffi_cif *cif)
/*@=exportheader@*/
{
register unsigned int i;
register int tmp;
register unsigned int avn;
register void **p_argv;
register char *argp;
register ffi_type **p_arg;
tmp = 0;
argp = stack;
if ( cif->rtype->type == FFI_TYPE_STRUCT ) {
*rvalue = *(void **) argp;
argp += 4;
}
avn = cif->nargs;
p_argv = avalue;
for (i = cif->nargs, p_arg = cif->arg_types;
(i != 0) && (avn != 0);
i--, p_arg++)
{
size_t z;
/* Align if necessary */
if (((*p_arg)->alignment - 1) & (unsigned) argp) {
argp = (char *) ALIGN(argp, (*p_arg)->alignment);
}
if (avn != 0)
{
avn--;
z = (*p_arg)->size;
/* because we're little endian, this is
what it turns into. */
*p_argv = (void*) argp;
p_argv++;
argp += z;
}
}
return;
}
/* How to make a trampoline. Derived from gcc/config/i386/i386.c. */
#define FFI_INIT_TRAMPOLINE(TRAMP,FUN,CTX) \
({ unsigned char *__tramp = (unsigned char*)(TRAMP); \
unsigned int __fun = (unsigned int)(FUN); \
unsigned int __ctx = (unsigned int)(CTX); \
unsigned int __dis = __fun - ((unsigned int) __tramp + 10); \
*(unsigned char*) &__tramp[0] = 0xb9; \
*(unsigned int*) &__tramp[1] = __ctx; \
*(unsigned char*) &__tramp[5] = 0xe9; \
*(unsigned int*) &__tramp[6] = __dis; \
})
/* the cif must already be prep'ed */
ffi_status
ffi_prep_closure (ffi_closure* closure,
ffi_cif* cif,
void (*fun)(ffi_cif*,void*,void**,void*),
void *user_data)
{
FFI_ASSERT (cif->abi == FFI_SYSV);
FFI_INIT_TRAMPOLINE (&closure->tramp[0], \
&ffi_closure_SYSV, \
(void*)closure);
closure->cif = cif;
closure->user_data = user_data;
closure->fun = fun;
return FFI_OK;
}
/* ------- Native raw API support -------------------------------- */
#if !FFI_NO_RAW_API
static void
ffi_closure_raw_SYSV ()
{
// this is our return value storage
long double res;
// our various things...
void *args;
ffi_raw *raw_args;
ffi_cif *cif;
ffi_raw_closure *closure;
unsigned short rtype;
void *resp = (void*)&res;
/* grab the trampoline context pointer */
asm ("movl %%ecx,%0" : "=r" (closure));
/* take the argument pointer */
asm ("leal 8(%%ebp),%0" : "=q" (args));
/* get the cif */
cif = closure->cif;
/* the SYSV/X86 abi matches the RAW API exactly, well.. almost */
raw_args = (ffi_raw*) args;
(closure->fun) (cif, resp, raw_args, closure->user_data);
rtype = cif->flags;
/* now, do a generic return based on the value of rtype */
if (rtype == FFI_TYPE_INT)
{
asm ("movl (%0),%%eax" : : "r" (resp) : "eax");
}
else if (rtype == FFI_TYPE_FLOAT)
{
asm ("flds (%0)" : : "r" (resp) : "st" );
}
else if (rtype == FFI_TYPE_DOUBLE)
{
asm ("fldl (%0)" : : "r" (resp) : "st", "st(1)" );
}
else if (rtype == FFI_TYPE_LONGDOUBLE)
{
asm ("fldt (%0)" : : "r" (resp) : "st", "st(1)" );
}
else if (rtype == FFI_TYPE_SINT64)
{
asm ("movl 0(%0),%%eax; movl 4(%0),%%edx"
: : "r"(resp)
: "eax", "edx");
}
}
ffi_status
ffi_prep_raw_closure (ffi_raw_closure* closure,
ffi_cif* cif,
void (*fun)(ffi_cif*,void*,ffi_raw*,void*),
void *user_data)
{
int i;
FFI_ASSERT (cif->abi == FFI_SYSV);
// we currently don't support certain kinds of arguments for raw
// closures. This should be implemented by a seperate assembly language
// routine, since it would require argument processing, something we
// don't do now for performance.
for (i = cif->nargs-1; i >= 0; i--)
{
FFI_ASSERT (cif->arg_types[i]->type != FFI_TYPE_STRUCT);
FFI_ASSERT (cif->arg_types[i]->type != FFI_TYPE_LONGDOUBLE);
}
FFI_INIT_TRAMPOLINE (&closure->tramp[0], &ffi_closure_raw_SYSV,
(void*)closure);
closure->cif = cif;
closure->user_data = user_data;
closure->fun = fun;
return FFI_OK;
}
static void
ffi_prep_args_raw(char *stack, extended_cif *ecif)
{
memcpy (stack, ecif->avalue, ecif->cif->bytes);
}
/* we borrow this routine from libffi (it must be changed, though, to
* actually call the function passed in the first argument. as of
* libffi-1.20, this is not the case.)
*/
extern void
ffi_call_SYSV(void (*)(char *, extended_cif *),
/*@out@*/ extended_cif *,
unsigned, unsigned,
/*@out@*/ unsigned *,
void (*fn)());
void
ffi_raw_call(/*@dependent@*/ ffi_cif *cif,
void (*fn)(),
/*@out@*/ void *rvalue,
/*@dependent@*/ ffi_raw *fake_avalue)
{
extended_cif ecif;
void **avalue = (void **)fake_avalue;
ecif.cif = cif;
ecif.avalue = avalue;
/* If the return value is a struct and we don't have a return */
/* value address then we need to make one */
if ((rvalue == NULL) &&
(cif->rtype->type == FFI_TYPE_STRUCT))
{
/*@-sysunrecog@*/
ecif.rvalue = alloca(cif->rtype->size);
/*@=sysunrecog@*/
}
else
ecif.rvalue = rvalue;
switch (cif->abi)
{
case FFI_SYSV:
/*@-usedef@*/
ffi_call_SYSV(ffi_prep_args_raw, &ecif, cif->bytes,
cif->flags, ecif.rvalue, fn);
/*@=usedef@*/
break;
default:
FFI_ASSERT(0);
break;
}
}
#endif
+167
View File
@@ -0,0 +1,167 @@
/* -----------------------------------------------------------------------
sysv.S - Copyright (c) 1996, 1998 Cygnus Solutions
X86 Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
#include <ffi_private.h>
.text
.globl ffi_prep_args
# This assumes we are using gas.
.balign 16
.globl ffi_call_SYSV
.type ffi_call_SYSV,@function
ffi_call_SYSV:
.LFB1:
pushl %ebp
.LCFI0:
movl %esp,%ebp
.LCFI1:
# Make room for all of the new args.
movl 16(%ebp),%ecx
subl %ecx,%esp
movl %esp,%eax
# Place all of the ffi_prep_args in position
pushl 12(%ebp)
pushl %eax
call *8(%ebp)
# Return stack to previous state and call the function
addl $8,%esp
call *28(%ebp)
# Remove the space we pushed for the args
movl 16(%ebp),%ecx
addl %ecx,%esp
# Load %ecx with the return type code
movl 20(%ebp),%ecx
# If the return value pointer is NULL, assume no return value.
cmpl $0,24(%ebp)
jne retint
# Even if there is no space for the return value, we are
# obliged to handle floating-point values.
cmpl $FFI_TYPE_FLOAT,%ecx
jne noretval
fstp %st(0)
jmp epilogue
retint:
cmpl $FFI_TYPE_INT,%ecx
jne retfloat
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
movl %eax,0(%ecx)
jmp epilogue
retfloat:
cmpl $FFI_TYPE_FLOAT,%ecx
jne retdouble
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstps (%ecx)
jmp epilogue
retdouble:
cmpl $FFI_TYPE_DOUBLE,%ecx
jne retlongdouble
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstpl (%ecx)
jmp epilogue
retlongdouble:
cmpl $FFI_TYPE_LONGDOUBLE,%ecx
jne retint64
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstpt (%ecx)
jmp epilogue
retint64:
cmpl $FFI_TYPE_SINT64,%ecx
jne retstruct
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
movl %eax,0(%ecx)
movl %edx,4(%ecx)
retstruct:
# Nothing to do!
noretval:
epilogue:
movl %ebp,%esp
popl %ebp
ret
.LFE1:
.ffi_call_SYSV_end:
.size ffi_call_SYSV,.ffi_call_SYSV_end-ffi_call_SYSV
.section .eh_frame,"aw",@progbits
__FRAME_BEGIN__:
.4byte .LLCIE1
.LSCIE1:
.4byte 0x0
.byte 0x1
.byte 0x0
.byte 0x1
.byte 0x7c
.byte 0x8
.byte 0xc
.byte 0x4
.byte 0x4
.byte 0x88
.byte 0x1
.align 4
.LECIE1:
.set .LLCIE1,.LECIE1-.LSCIE1
.4byte .LLFDE1
.LSFDE1:
.4byte .LSFDE1-__FRAME_BEGIN__
.4byte .LFB1
.4byte .LFE1-.LFB1
.byte 0x4
.4byte .LCFI0-.LFB1
.byte 0xe
.byte 0x8
.byte 0x85
.byte 0x2
.byte 0x4
.4byte .LCFI1-.LCFI0
.byte 0xd
.byte 0x5
.align 4
.LEFDE1:
.set .LLFDE1,.LEFDE1-.LSFDE1
+125
View File
@@ -0,0 +1,125 @@
/* -----------------------------------------------------------------------
win32.S - Copyright (c) 1996, 1998, 2001 Cygnus Solutions
X86 Foreign Function Interface
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 CYGNUS SOLUTIONS 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.
----------------------------------------------------------------------- */
#define LIBFFI_ASM
#include <ffi.h>
.text
.globl ffi_prep_args
# This assumes we are using gas.
.balign 16
.globl _ffi_call_SYSV
_ffi_call_SYSV:
pushl %ebp
movl %esp,%ebp
# Make room for all of the new args.
movl 16(%ebp),%ecx
subl %ecx,%esp
movl %esp,%eax
# Place all of the ffi_prep_args in position
pushl 12(%ebp)
pushl %eax
call *8(%ebp)
# Return stack to previous state and call the function
addl $8,%esp
call *28(%ebp)
# Remove the space we pushed for the args
movl 16(%ebp),%ecx
addl %ecx,%esp
# Load %ecx with the return type code
movl 20(%ebp),%ecx
# If the return value pointer is NULL, assume no return value.
cmpl $0,24(%ebp)
jne retint
# Even if there is no space for the return value, we are
# obliged to handle floating-point values.
cmpl $FFI_TYPE_FLOAT,%ecx
jne noretval
fstp %st(0)
jmp epilogue
retint:
cmpl $FFI_TYPE_INT,%ecx
jne retfloat
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
movl %eax,0(%ecx)
jmp epilogue
retfloat:
cmpl $FFI_TYPE_FLOAT,%ecx
jne retdouble
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstps (%ecx)
jmp epilogue
retdouble:
cmpl $FFI_TYPE_DOUBLE,%ecx
jne retlongdouble
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstpl (%ecx)
jmp epilogue
retlongdouble:
cmpl $FFI_TYPE_LONGDOUBLE,%ecx
jne retint64
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
fstpt (%ecx)
jmp epilogue
retint64:
cmpl $FFI_TYPE_SINT64,%ecx
jne retstruct
# Load %ecx with the pointer to storage for the return value
movl 24(%ebp),%ecx
movl %eax,0(%ecx)
movl %edx,4(%ecx)
retstruct:
# Nothing to do!
noretval:
epilogue:
movl %ebp,%esp
popl %ebp
ret
.ffi_call_SYSV_end:
-4
View File
@@ -1,4 +0,0 @@
man_MANS = mcs.1
EXTRA_DIST = $(man_MANS)
-118
View File
@@ -1,118 +0,0 @@
.TH mcs 1 "6 January 2001"
.SH NAME
mcs \- Mono Compiler Suite.
.SH SYNOPSIS
.B mcs
[option] [source-files]
.SH DESCRIPTION
mcs is the Mono C# compiler, an implementation of the ECMA-334
specification. You can pass one or more options to drive the
compiler, and a set of source files. Extra options or arguments can
be provided in a response file. Response files are referenced by
prepending the @ symbol to the response file name.
.PP
C# source files must end with a ".cs" extension. Compilation of C#
source code requires all the files that make up a library, module or
executable to be provided on the command line. There is no support
for partial compilation. To achieve the benefits of partial
compilation, you should compile programs into their own assemblies,
and later reference them with the "-r" flag.
.PP
The Mono C# compiler generates images (.exe files) that contain CIL
byte code that can be executed by any system that implements a Common
Language Infrastructure virtual machine such as the Microsoft .NET
runtime engine on Windows or the Mono runtime engine on Unix systems.
Executables are not bound to a specific CPU or operating system.
.PP
.SH OPTIONS
.TP
.I \-\-about
Displays information about the Mono C# compiler
.TP
.I \-\-checked
Sets the default compilation mode to `checked'. This makes all
the math operations checked (the default is unchecked).
.TP
.I \-\-define SYMBOL
Defines a symbol named SYMBOL. This can be tested in the source code
by the pre-processor.
.TP
.I \-\-fatal
This is used for debugging the compiler. This makes the error emission
generate an exception that can be caught by a debugger.
.TP
.I \-\-stacktrace
Generates a stack trace at the time the error is reported, useful for
debugging the compiler.
.TP
.I \-L PATH
Directs the compiler to look for libraries in the specified path.
Multiple paths can be provided.
.TP
.I \-\-nostdlib
Use this flag if you want to compile the core library. This makes the
compiler load its internal types from the assembly being compiled.
.TP
.I \-\-nowarn XXX
Makes the compiler ignore warning XXX.
.TP
.I \-o FNAME
Names the output file to be generated.
.TP
.I \-\-optimize
Turns on optimizations in the compiler.
.TP
.I \-\-parse
Used for benchmarking. The compiler will only parse its input files.
.TP
.I \-\-probe X L
Probes for the code to generate an error named `X' in line `L'. This
is only used by the test suite.
.TP
.I \-\-target KIND
Used to specify the desired target. The possible values are: exe,
winexe, library and module.
.TP
.I \-\-timestamp
Another debugging flag. Used to display the times at various points
in the compilation process.
.TP
.I \-\-unsafe
Enables compilation of unsafe code.
.TP
.I \-\-werror
Treat warnings as errors.
.TP
.I \-\-wlevel LEVEL
Sets the warning level. 0 is the lowest warning level, and 4 is the
highest. The default is 2.
.TP
.I \-r ASSEMBLY
Reference the named assembly. Use this to use classes from the named
assembly in your program.
.TP
.I \-v
Debugging. Turns on verbose yacc parsing.
.PP
.SH AUTHORS
The Mono C# Compiler was written by Miguel de Icaza and Ravi Pratap at
Ximian.
.PP
.SH LICENSE
The Mono Compiler Suite is released under the terms of the GNU GPL.
Please read the accompanying `COPYING' file for details. Alternative
licenses are available from Ximian.
.PP
.SH BUGS
To report bugs in the compiler, you can use `bug-buddy', or you can
file bug reports in our bug tracking system:
http://bugzilla.ximian.com.
.SH MAILING LIST
The Mono Mailing List is available at: [email protected]
.SH MORE INFORMATION
The Mono C# compiler is developed by Ximian, Inc
(http://www.ximian.com) (http://www.ximian.com) and is based on the
ECMA C# language standard available here:
http://www.ecma.ch/ecma1/STAND/ecma-334.htm
-1
View File
@@ -2,7 +2,6 @@ C# Compiler:
Miguel de Icaza ([email protected])
Class Libraries:
Gaurav Vaish <[email protected]>
Jeff Stedfast ([email protected])
Joe Shaw ([email protected])
Miguel de Icaza ([email protected])
-8
View File
@@ -1,8 +0,0 @@
2002-01-20 Nick Drochak <[email protected]>
* on the 'test' target, make sure NUnit is built first before building
and running tests
2002-01-20 Nick Drochak <[email protected]>
* add nunit directory to the list of dirs to build in
+1 -3
View File
@@ -18,6 +18,4 @@ Layout:
tests/
Sample tests
Thanks a lot to Sergey Chaban for his help during the development of
the C# compiler.
Loaded 100 of 1859 files, more files were not shown because too many files have changed in this diff. Show more