Because of the trick of process-archive-undefs, all .mo objects, even
with --enable-modules, are dependencies of executables.

This breaks CFLAGS propogation because the compiling of module object
will happen too early before building for DSO.

With GCC 5, the linking would fail because .o doesn't have -fPIC. Also,
BUILD_DSO will be missed. (module-common.o will have it, so the stamp
symbol was still liked in .so).

Fix the problem by forcing the CFLAGS on individual .o-cflags during
unnest-vars.

Reported-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Fam Zheng <famz@redhat.com>
Cc: qemu-stable@nongnu.org # 2.3
Message-Id: <1430981715-31465-1-git-send-email-famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Module build is broken since commit c261d774 ( rules.mak: Fix DSO
build by pulling in archive symbols). That commit added .mo placeholders
of DSO to -y variables, in order to pull stub symbols to executable. But
the placeholders are unintentionally expanded in -y, rather than
filtered out while linking.

Fix it by moving the -objs expanding to before inserting .mo
placeholders.  Note that passing -cflags and -libs to member objects are
also moved to keep it happening before object expanding.

Reported-by: Bharata B Rao <bharata.rao@gmail.com>
Tested-by: Bharata B Rao <bharata.rao@gmail.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Expand %.mo-objs in -y nested objects, so that we can write combined
object -cflags rules like what will be done in the coming patch.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Commit c261d774 added one more binutils
tool: nm also needs a cross prefix.

Signed-off-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1411070108-8954-1-git-send-email-sw@weilnetz.de
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

This fixes an issue with module build system. block/iscsi.so is
currently broken:

    $ ~/build/last/qemu-img
    Failed to open module: /home/fam/build/master/block-iscsi.so:
    undefined symbol: qmp_query_uuid
    qemu-img: Not enough arguments
    Try 'qemu-img --help' for more information

To fix this, we should (at least) let qemu-img link qmp_query_uuid from
libqemustub.a. (There are a few other symbols missing, as well.)

This patch changes the linking rules to:

1) Build ".mo" with "ld -r -o $@ $^" for each ".so", and later build .so
   with it.

2) Always build all the .mo before linking the executables. This is
   achieved by adding those .mo files to the executables' "-y"
   variables.

3) When linking an executable, those .mo files in its "-y" variables are
   filtered out, and replaced by one or more -Wl,-u,$symbol flags. This
   is done in the added macro "process-archive-undefs".

   These "-Wl,-u,$symbol" flags will force ld to pull in the function
   definition from the archives when linking.

   Note that the .mo objects, that are actually meant to be linked in
   the executables, are already expanded in unnest-vars, before the
   linking command. So we are safe to simply filter out .mo for the
   purpose of pulling undefined symbols.

   process-archive-undefs works as this: For each ".mo", find all the
   undefined symbols in it, filter ones that are defined in the
   archives. For each of these symbols, generate a "-Wl,-u,$symbol" in
   the link command, and put them before archive names in the command
   line.

Suggested-by: H.J. Lu <hjl.tools@gmail.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Use common rule (macro) to install and strip binaries, and use
it in all places where we install binaries, instead of fixing
bugs like 1319493 in every place.
(This fixes https://bugs.launchpad.net/bugs/1319493

)

Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>

Duplicate removal was added to extract-libs in order to avoid including
the same library multiple times into the linking command line; this could
potentially happen when using "foo.mo-libs" (which adds the library to
all components, causing it to appear N times if the module is composed
of N objects).  However, sorting and removing duplicates causes problems
with static linking, and also with space-separated linker options as
found in some Mac OS X packaging systems.  Furthermore, the "optimization"
is really a non-problem since we do not expect .mo modules to be composed
of many files.

Reported-by: Sean Bruno <sbruno@ignoranthack.me>
Tested-by: Sean Bruno <sbruno@ignoranthack.me>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1402929805-16836-1-git-send-email-pbonzini@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

The macro unnest-vars is the most important, complicated but hard to
track magic in QEMU's build system.

Rewrite it in a (hopefully) clearer way, with more comments, to make it
easier to understand and maintain.

Remove DSO_CFLAGS and module-objs-m that are not used.

A bonus fix of this version is, per object variables are properly
protected in save-objs and load-objs, before including sub-dir
Makefile.objs, just as nested variables are. So the occasional same
object name from different directory levels won't step on each other's
foot.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

fix-obj-vars has the undesired side effect of breaking -cflags
-objs and -libs variables in the toplevel Makefile.objs.  The
variables in the toplevel Makefile.objs do not need any fix,
so fix-obj-vars need not do anything.

Since we are touching it, remove the now unnecessary $(if)
in the callers.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

This is needed in order to use per-object flags variables.

Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

The compiling is done in a subdir, so the extraction of per-object libs
and cflags are referencing objects with ../ prefixed. So prefix the
per-object variables "foo.o-cflags" and "foo.o-libs" to
"../foo.o-cflags" and "../foo.o-libs".

Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Don't sort the extracted options, sort the objects.

Reported-by: Christian Mahnke <cmahnke@googlemail.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

This patch adds loading, stamp checking and initialization of modules.

The init function of dynamic module is no longer directly called as
__attribute__((constructor)) in static linked version, it is called
only after passed the checking of presense of stamp symbol:

    qemu_stamp_$RELEASEHASH

where $RELEASEHASH is generated by hashing version strings and content
of configure script.

With this, modules built from a different tree/version/configure will
not be loaded.

The module loading code requires gmodule-2.0.

Modules are searched under
 - CONFIG_MODDIR
 - executable folder (to allow running qemu-{img,io} in the build
   directory)
 - ../ of executable folder (to allow running system emulator in the
   build directory)

Modules are linked under their subdir respectively, then copied to top
level of build directory for above convinience, e.g.:
    $(BUILD_DIR)/block/curl.so -> $(BUILD_DIR)/block-curl.so

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Add necessary rules and flags for shared object generation.
The new rules introduced here are:

1) %.o in $(common-obj-m) is compiled to %.o, then linked to %.so.

2) %.mo in $(common-obj-m) is the placeholder for %.so for pattern
matching in Makefile. It's linked to "-shared" with all its dependencies
(multiple *.o) as input. Which means the list of depended objects must
be specified in each sub-Makefile.objs:

    foo.mo-objs := bar.o baz.o qux.o

in the same style with foo.o-cflags and foo.o-libs. The objects here
will be prefixed with "$(obj)/" if it's a subdirectory Makefile.objs.

3) For all files ending up in %.so, the following is added automatically:

    foo.o-cflags += -fPIC -DBUILD_DSO

Also introduce --enable-modules in configure, the option will enable
support of shared object build. Otherwise objects are static linked to
executables.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Adds extract-libs in LINK to expand any "per object libs", the syntax to define
such a libs options is like:

        foo.o-libs := $(CURL_LIBS)

in block/Makefile.objs.

Similarly,

        foo.o-cflags := $(FOO_CFLAGS)

is also supported.

"foo.o" must be listed in a nested var (e.g. common-obj-y) to make the
option variables effective.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Makefile.target includes rule.mak and unnested common-obj-y, then prefix
them with '../', this will ignore object specific QEMU_CFLAGS in subdir
Makefile.objs:

    $(obj)/curl.o: QEMU_CFLAGS += $(CURL_CFLAGS)

Because $(obj) here is './block', instead of '../block'. This doesn't
hurt compiling because we basically build all .o from top Makefile,
before entering Makefile.target, but it will affact arriving per-object
libs support.

The starting point of $(obj) is passed in as argument of unnest-vars, as
well as nested variables, so that different Makefiles can pass in a
right value.

Signed-off-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

If we have a C++ compiler available, link with it, because we might be
linking some C++ files in. This allows us to include C++ object files
in the QEMU binary proper.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>

The A64 disassembler libvixl uses .cc as its suffix for
C++ source files, so add support for it (we already support
.cpp).

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>

Add new string testing functions which return a y/n result:
 eq : are two strings equal (ignoring leading/trailing space)?
 ne : are two strings unequal?
 isempty : is a string empty?
 notempty : is a string non-empty?

Based on an idea by Ákos Kovács <akoskovacs@gmx.com>.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Add new logical functions for handling y/n values like those we
use in CONFIG_FOO variables:
 lnot : logical NOT
 land : logical AND
 lor : logical OR
 lxor : logical XOR
 leqv : logical equality, inverse of lxor
 lif : like Make's $(if) but with an eq-like test

Based on an idea by Ákos Kovács <akoskovacs@gmx.com>.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Add configuration for C++ compiler in configure and Makefiles.
The C++ compiler is choosed as following:
 - ${CXX}, if it is specified.
 - ${cross_prefix}g++, if ${cross_prefix} is specified.
 - Otherwise, c++ is used.

Currently, usage of C++ language is only for access to Windows VSS
using COM+ services in qemu-guest-agent for Windows.

Signed-off-by: Tomoki Sekiyama <tomoki.sekiyama@hds.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Micael Roth <mdroth@linux.vnet.ibm.com>
Signed-off-by: Michael Roth <mdroth@linux.vnet.ibm.com>

While commit c02817e5 fixed compilation
without an installed libtool, moving the dependencies to rules.mak does
not work because the version-*-y variables are not defined yet.  Building
in a clean tree thus fails.

Revert the commit and remove the dummy /bin/false assignment to LIBTOOL.
This makes the build work, at the price of slightly worse errors when
there are Makefile bugs.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1367425815-15083-1-git-send-email-pbonzini@redhat.com
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

The following error occurs when building dtc module:

        CHK version_gen.h
         CC libfdt/fdt.o
cc1: error: dtc: No such file or directory [-Werror]
cc1: all warnings being treated as errors
make[1]: *** [libfdt/fdt.o] Error 1
make: *** [subdir-dtc] Error 2

In rules.mak, "-I$(<D) -I$(@D)" was expanded to "-Idtc -I." when
building submodule dct. Due to the using of "-Wmissing-include-dirs,
a warning would be rarsed. To avoid it, add "-I$(<D) -I$(@D)" to
QEMU_INCLUDES instead of QEMU_CFLAGS so that QEMU_CFLAGS does not
contain the "-Idtc".

Cc: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Cc: Blue Swirl <blauwirbel@gmail.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Dunrong Huang <riegamaths@gmail.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 1367247132-19622-1-git-send-email-riegamaths@gmail.com
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

This makes the test on $(LIBTOOL) work.  Otherwise, LIBTOOL
is /bin/false by the time the test is done.

Fixes Win32 compilation without a working cross-libtool.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Compile and link with version.lo

Signed-off-by: Marc-André Lureau <mlureau@redhat.com>
Reviewed-by: Alon Levy <alevy@redhat.com>

It is needed to give that flag to the linker as well, but latest
libtool 2.4.2 still swallows that argument, so let's pass it with
libtool -Wc argument.

qemu-1.4.0/stubs/arch-query-cpu-def.c:6: undefined reference to `__stack_chk_guard'

Signed-off-by: Marc-André Lureau <mlureau@redhat.com>
Reviewed-by: Alon Levy <alevy@redhat.com>

Using a global pattern makes it easier to clean out
old generated files.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

This addresses two issues with config generation
1. rule generating timestamp has side effect.
Thus cleanup on error does not work.
2. rule for handling timestamp is too generic.
It can create any missing .h file.
As a result when .h file is removed, build
might try to create it using this rule which
results in build errors.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

There is no reason for it to be in block-obj-y, in particular.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Do not fail at build time, instead just disable the library if libtool
is not present.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

This patch moves the complication of using libtool to the generic
rules.mak file.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Build option ROM .S files with separate preprocessor and
assembler steps because the C compiler could be unsuitable.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Make it correct for nested directories, and move the static part
from Makefile to configure.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Weak symbols were a nice idea, but they turned out not to be a good one.
Toolchain support is just too sparse, in particular llvm-gcc is totally
broken.

This patch uses a surprisingly low-tech approach: a static library.
Symbols in a static library are always overridden by symbols in an
object file.  Furthermore, if you place each function in a separate
source file, object files for unused functions will not be taken in.
This means that each function can use all the dependencies that it needs
(especially QAPI stuff such as error_setg).

Thus, all stubs are placed in separate object files and put together in
a static library.  The library then is linked to all programs.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Tested-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>

MacOSX 10.8 ("Mountain Lion") requires us to compile our one
Objective-C source file with clang even if the rest of QEMU
requires a real gcc, because the system headers we use make
use of Apple's "Blocks" extension to C/ObjC, and mainline
gcc doesn't support that. Since we only need to use a true
gcc for the parts of QEMU that use the fixed-register
env variable, we can simply use clang to build the ObjC
file: it will link to the gcc-built objects with no problems.

Add the necessary support for an OBJCC variable in the
makefile and configure machinery; we default to clang
if we have it, otherwise whatever CC is (since gcc
might be the Apple gcc which does support Blocks).

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>

After commit dcff25f2, Dependency file
are taken from the directories that have a Makefile.objs file.  This is
not enough, since files can be included from other directories.
So, pick them from directories that have an object file in them.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>