DEBUG_RUNTIME Makefile variable to pass -g to gcc when building LLVM's objects.
Without this, it's very hard to debug crashes that happen in Release-Asserts
mode but not Debug mode.
llvm-svn: 82775
try to use i686-darwin to build for arm-eabi, you'll quickly run into
several false assumptions that the target OS must be the same as the
host OS. These patches split $(OS) into $(HOST_OS) and $(TARGET_OS) to
help builds like "make check" and the test-suite able to cross
compile. Along the way a target of *-unknown-eabi is defined as
"Freestanding" so that TARGET_OS checks have something to work with.
Patch by Sandeep Patel!
llvm-svn: 79296
Controls whether libCompilerDriver should be loaded dynamically. By default this
is needed only on Win32, to make dynamic plugins work.
llvm-svn: 74759
Implemented by making lib/CompilerDriver a shared library that holds all the
global static data (CommandLine options, plugin registry) that we unfortunately
have to live with.
llvm-svn: 74417
Chris recently broke llvmc with his Makefile changes (r75379). That patch made
the global change .o -> .a, which caused built-in llvmc plugins to stop working
since plugin initialization in llvmc is based on static variables not referenced
from the main executable. This patch implements auto-generated forced references
to the plugin libraries.
llvm-svn: 74000
place it with -pedantic. Remove -Wunused since it
is implied by -Wall. Group -Wno-unused-parameter
with -Wall -W since it is the combination of these
two that turns on -Wunused-parameter.
llvm-svn: 73765
object files. Now we always build LLVM libraries into archives (.a files).
This makes the 'make' build work more like the cmake build, among other
things. Doing this exposed some latent circular library dependencies, so
I think that llvm-config wasn't quite right for .o files anyway.
llvm-svn: 73579
gcc rather than directly to the linker: gcc will
then modify the linker options it generates (it
will not use --relax for example, incompatible
with -r, as it otherwise would). This fixes the
sparc build.
llvm-svn: 72943
pthreads, otherwise ar complains of unresolved references to pthread_mutex_*.
LTO doesn't actually use pthreads, but ManagedStatic does.
Fix this by linking in LIBS (that contains pthreads) for LTO and gold.
For now this links in more libs than needed (libffi for example), we can figure
out later how to link in those libs per-tool later.
llvm-svn: 72430
- This matches the normal dependency generation code.
- This also fixes the problem that when building a normal and bitcode
archive from the same source, the dependency files would overwrite
one another. Which was bad.
llvm-svn: 71542
we're not testing out-of-date code. This also makes "make check" and
"make unittests" work out-of-the box right after the configure step, without
requiring the user to run "make tools-only" or "make libs-only".
Tested on Linux/x86_64 and Darwin/x86.
llvm-svn: 68708
- Required some extra makefile tweaks to introduce a new flag var
which only goes to compile/link tools but not the relink step,
otherwise we get a copy of libgcov in the relinked .o files.
- No configure magic for this.
llvm-svn: 66945
From the code: "There are "Proj" libs (defined by the user's project) and "LLVM"
libs (defined by the LLVM project)." LINK_COMPONENTS are clearly defined by the
LLVM project.
Additionally, this fixes an issue with llvmc's build process:-)
llvm-svn: 65940
target directories themselves. This also means that VMCore no longer
needs to know about every target's list of intrinsics. Future work
will include converting the PowerPC target to this interface as an
example implementation.
llvm-svn: 63765
essential problem was that the DAG can contain
random unused nodes which were never analyzed.
When remapping a value of a node being processed,
such a node may become used and need to be analyzed;
however due to operands being transformed during
analysis the node may morph into a different one.
Users of the morphing node need to be updated, and
this wasn't happening. While there I added a bunch
of documentation and sanity checks, so I (or some
other poor soul) won't have to scratch their head
over this stuff so long trying to remember how it
was all supposed to work next time some obscure
problem pops up! The extra sanity checking exposed
a few places where invariants weren't being preserved,
so those are fixed too. Since some of the sanity
checking is expensive, I added a flag to turn it
on. It is also turned on when building with
ENABLE_EXPENSIVE_CHECKS=1.
llvm-svn: 60797
use a SUB instruction instead of an ADD, because -128 can be
encoded in an 8-bit signed immediate field, while +128 can't be.
This avoids the need for a 32-bit immediate field in this case.
A similar optimization applies to 64-bit adds with 0x80000000,
with the 32-bit signed immediate field.
To support this, teach tablegen how to handle 64-bit constants.
llvm-svn: 57663
runnable on the build machine.
There are a few bits that need built for the build environment (TableGen).
This patch builds those bits, and the associated libraries, for the build
environment as well as the (usual) host environment.
Thanks to Eric C. and Devang P. for pre-commit review.
llvm-svn: 56975
Don't include system headers in the .d files.
Don't use $@ in the makefile rules, as there are two possible targets it could resolve to: use the one that we need explicitly.
llvm-svn: 45473
put into the CPP.Defines variable. Seems the convention was corrupted with
various changes made. It is important to get command line parameters into the
right variable because things like llvm-config and sub-makefiles depend on it.
llvm-svn: 38486
--enable-expensive-checks allows the developer to enable runtime
checking that can greatly increase compile time. Currently it only
turns on _GLIBCXX_DEBUG. Other expensive debugging checks added later
should be controlled by this configure option.
This patch also updates llvm-config with a --cppflags option to inform
llvm-gcc how to build itself so that it is compatible with an llvm that
was built with _GLIBCXX_DEBUG.
llvm-svn: 37777
Make the sys::Path::getFileStatus function more efficient by having it
return a pointer to the FileStatus structure rather than copy it. Adjust
uses of the function accordingly. Also, fix some memory issues in sys::Path.
llvm-svn: 35476
"clean" target item. This gets around having the built sources disappear
when only one build mode wants to be cleaned. With the "clean-all" target,
all build modes are cleaned so it also makes sense to clean out the built
sources at that point.
llvm-svn: 34003
a directory's Makefile will prevent the build products from that directory
from being installed. This is useful for tools and libraries that are
only useful as part of the build process.
llvm-svn: 33968
libraries linked with. This permits a project to still use USEDLIBS to
specify its own libraries in conjunction with LINK_COMPONENTS. llvm-stacker
needs this after libLLVMTransforms.a went away.
llvm-svn: 33886
depend on the compiler. This works around problems in the Stacker runtime
when the CFE changes in such a way that the assembly file needs to be
updated.
llvm-svn: 32773
premature, these libraries will be going away for the 2.0 release. Other
arrangements for profiling, gc, etc. should be made in the next few months.
llvm-svn: 31807
Turn on -Wunused and -Wno-unused-parameter. Clean up most of the resulting
fall out by removing unused variables. Remaining warnings have to do with
unused functions (I didn't want to delete code without review) and unused
variables in generated code. Maintainers should clean up the remaining
issues when they see them. All changes pass DejaGnu tests and Olden.
llvm-svn: 31380
will compute a locally wrong numbering for the intrinsics. This fixes a
nasty issue where the x86 backend started miscompiling stuff in a 'cvs up'd
build after the altivec intrinsics were added.
llvm-svn: 31172
Close out this long standing bug by removing the remaining overloaded
virtual functions in LLVM. The -Woverloaded-virtual option is now turned on.
llvm-svn: 29934
Final commit for this bug. This removes the last EH holdouts in LLVM
and turns off exception support by using the -fno-exceptions option. This
leads to the following reduction in library and executable sizes:
DEBUG BUILD RELEASE BUILD
before after delta before after delta
lib 162,328K 157,616K 4,712 17,864K 16,416K 1,448K
bin 571,444K 557,156K 14,288 63,296K 56,996K 6,300K
Debug Improvement: 19,000K (2.59%)
Release Improvement: 7,748K (9.55%)
llvm-svn: 29882
and a handler, which would produce errors like:
terminate called after throwing an instance of 'std::string'
we must comment out setting -fno-exceptions until PR797 is completely
fixed. Once libraries like lib/System and lib/Support are exception free,
we can turn it back on.
llvm-svn: 29768
1. Actually turn on -fno-exceptions in libraries that do not have the
REQUIRES_EH option in their Makefile. The following library file size
savings were made (DEBUG):
libLLVMDataStructure.a 525K
libLLVMCore.a 380K
libLLVMCodeGen.a 350K
libLLVMTransformUtils.a 305K
libLLVMScalarOpts.a 270K
libLLVMAnalysis.a 247K
libLLVMSelectionDAG.a 233K
libLLVMipo.a 175K
LLVMX86.o 123K
LLVMPPC.o 81K
libLLVMipa.a 17K
TOTAL 2,706K
Note that the savings is actually a little larger than this because
I didn't count any of the libraries that had small changes.
2. Remove REQUIRES_EH from the AsmParser library as it is now exception
free. This resulted in a nearly 78K drop in the size of the debug
library for AsmParser.
llvm-svn: 29767
1. Change the usage of LOADABLE_MODULE so that it implies all the things
necessary to make a loadable module. This reduces the user's burdern to
get a loadable module correctly built.
2. Document the usage of LOADABLE_MODULE in the MakefileGuide
3. Adjust the makefile for lib/Transforms/Hello to use the new specification
for building loadable modules
4. Adjust the sample project to not attempt to build a shared library for
its little library. This was just wasteful and not instructive at all.
llvm-svn: 29551
Turn -pedantic and -Wno-long-long compile flags on by default. In a few
places, avoid the warnings by removing these options in the local makefile.
One notable exception: lib/Target/CBackend/Writer.cpp. These warnings are
left on as a reminder to developers to clean them up.
llvm-svn: 28614
1. Remove the LLVM_DO_NOT_BUILD feature (not needed any more)
2. Ensure that lib/VMCore gets built first. This needs to be done because
VMCore now uses tblgen to generate the Intrinsics header which are
needed in other libraries. In parallel builds, this can cause problems.
llvm-svn: 28374
with which source is compiled are used when linking. This matters when a
project is using the LLVM makefiles and overrides CXXFLAGS to specify new
flags to use.
llvm-svn: 28322
libraries. This ensures that the project's libraries (which most likely
depend on LLVM libraries) come first on the command line and can thus be
resolved by the LLVM libraries that appear later.
llvm-svn: 28316
a -L option to gccld whenever we're building a bytecode module or archive.
This gets around the "Cannot find library 'crtend'" warning messages.
llvm-svn: 27621
1. Don't force debug builds to have assertion checking turned on always.
Let the default (on) be taken, or overridden by the command line
2. Create two new BuildModes based on assertion checking: Release+Assert
and Debug-Assert.
3. Ensure that when building a distribution we get a release build with
assertions enabled, regardless of the tree's configuration.
4. (unrelated) Fix library name generation for llvm-config usage.
llvm-svn: 27488
This facility allows LLVMLIBS to be specified with something like:
LLVMLIBS = config --libs jit
instead of:
LLVMLIBS = JIT
with the same effect. However, the llvm-config utility is much more versatile
than the single keyword approach. Note that "config" is the keyword after
which any arguments to llvm-config are allowed. When llvm-config is tested
and working well, we'll start using this and drop support for the JIT
keyword.
llvm-svn: 27057
is a handy tool for users of LLVM who want to be able to quickly get
information about LLVM's configuration. It is intended to be used in the
command line of other tools. Documentation will be forthcoming in a
subsequent patch.
llvm-svn: 26952
want to copy the files when the .cpp file changes, we want to copy them
to the .cvs versions when the .l/.y file change (like the comments even say).
This avoids having bogus changes show up in diffs.
llvm-svn: 26229
* Make runtimes and projects build with the new front-end by not relying on
'llvm-gcc -c' to build a .bc file. Instead, use llvm-gcc -S -emit-llvm,
then an explicit invocation of gccas. Also, don't use llvm-gcc to link
.bc files together, use gccld directly.
llvm-svn: 25707
1. When srcdir == objdir have "spotless" say that it isn't supported in
that mode rather than just let make say "no such target"
2. Minor doc cleanups
3. Fix the double rebuild problem with yacc files. A missing dependency
caused parallel builds to skip building the .cpp file after the .cpp
file was regenerated by bison.
llvm-svn: 24924
* Add --enable-debug-runtime option, defaults to disabled
* Pass the new config var, DEBUG_RUNTIME, to Makefiles
* Don't use -Wa,-strip-debug if debug-runtime is enabled
llvm-svn: 24891
into the LLVMAnalysis library.
This allows LLVMTranform and LLVMTransformUtils to be archives and linked
with LLVMAnalysis.a, which provides any missing definitions.
llvm-svn: 24036
SparcV9 JIT.
2. Make LLVMTransformUtils a relinked object file and always link it before
LLVMAnalysis.a. These two libraries have circular dependencies on each
other which creates problem when building the SparcV9 JIT. This change
fixes the dependency on all platforms problems with a minimum of fuss.
llvm-svn: 24023
pointer marking the end of the list, the zero *must* be cast to the pointer
type. An un-cast zero is a 32-bit int, and at least on x86_64, gcc will
not extend the zero to 64 bits, thus allowing the upper 32 bits to be
random junk.
The new END_WITH_NULL macro may be used to annotate a such a function
so that GCC (version 4 or newer) will detect the use of un-casted zero
at compile time.
llvm-svn: 23888
These changes modify the makefiles so that the output of flex and bison are
placed in the SRC directory, not the OBJ directory. It is intended that they
be checked in as any other LLVM source so that platforms without convenient
access to flex/bison can be compiled. From now on, if you change a .y or
.l file you *must* also commit the generated .cpp and .h files.
llvm-svn: 23115
Move the implementation of the fix from Makefile.rules to Makefile. This
ensures that it is only checked on a top-level rebuild, and not in every
single subdirectory. This removes some annoying messages from the build and
numerous executions of config.status if the .in file changes but not
substantively enough to cause the .h file to be modified by config.status.
llvm-svn: 23039
Make any header files that are automatically generated be preconditions of
the compilation. This ensures that if a *.h.in file is changed then its
corresponding *.h file gets updated on the next rebuild. Note that this can
lead to confusing (but correct) results if the *.h.in file changed
unsubstantially so that autoheader doesn't update the *.h file. In that case,
manually touch the *.h file in question to restore order. Moral of the story,
if you're going to "touch" a *.in file then modify it substantially.
llvm-svn: 23006
Jeffrey Yasskin