I managed to build a cross gcc-7.2.0 with multilib support for several Cortex-M-Targets with and without hard and soft floating point according to the processor capabilities.
Now, after final install step (make install), I find only one libstc++.a in the installation directory. For the other C++ libraries I see the same problem.
I expected one in every multilib subdirectory, the same way as I can find libc, libm and the like. But there are no libstdc++.a in the multilib subdirectories.
I think, this is not right.
Linking my test project failes with
libstdc++.a(atexit_arm.o) uses VFP register arguments, ../target.elf does not.
This suggests problems with the multilib installation.
How can I fix this multilib problem in the build phase?
After I added some configuration options to the configure call of binutils and GCC, the multilib configuration is working like a charm now.
For binutils I added --enable-version-specific-runtime-libs.
For GCC I added --enable-multiarch --enable-version-specific-runtime-libs.
Don't know, if the multiarch option is really necessary for my problem, but I didn't investigate further and I leave the information here.
Related
My developing/producing environments are all CentOS-7.7.
In order to compile my program with gcc-8.3.0, I have installed "devtoolset-8" on my developing env, but it can not be used in the way same as gcc-4.8.5 that was shipped with CentOS7 oringinally.
Every time I need to compile a program, I must use "scl enable devtoolset-8 -- bash" to switch to gcc8 instead of gcc4.8.5.
When the program was deploying onto the producing-env, there is no gcc8, nor libstdc++.so.6.0.25, so it can not run.
I guess libstdc++.so.6.0.25 should be released with gcc8? I can neither install "devtoolset-8" on the producing-env, nor build gcc8 from source on the producing env.
The version of libstdc++ that can be installed from the official yum repo of CentOS, is libstdc++.so.6.0.19, hence my programs can not be loaded at the producing-env.
How to let such programs to run?
Thanks!
Pls forgive my Ugly English.
In order to not have to copy or ship a separate libstdc++.so but rather link statically (as suggested in a comment) against the C++ runtime, one can link C++ programs with -static-libstdc++ (also specifying -static-libgcc will also make sure that the program does not depend on a recent enough version of libgcc_s.so on the system - although that should rarely be a problem).
There can also be the issue of the target system having a version of glibc that is too old (relative to the build system). In that case, one could anyhow compile gcc of no matter how recent of a version on the older system, so that the resulting C++ executables as well as libstdc++ are linked against the older glibc. Linking C++ programs with -static-libstdc++ will again help to not depend on the program having to be able to find libstdc++.so at run-time.
Finally, the C++ program could also be linked with -static not depending on any dynamic libraries at all.
I'm a student doing research involving extending the TM capabilities of gcc. My goal is to make changes to gcc source, build gcc from the modified source, and, use the new executable the same way I'd use my distro's vanilla gcc.
I built and installed gcc in a different location (not /usr/bin/gcc), specifically because the modified gcc will be unstable, and because our project goal is to compare transactional programs compiled with the two different versions.
Our changes to gcc source impact both /gcc and /libitm. This means we are making a change to libitm.so, one of the shared libraries that get built.
My expectation:
when compiling myprogram.cpp with /usr/bin/g++, the version of libitm.so that will get linked should be the one that came with my distro;
when compiling it with ~/project/install-dir/bin/g++, the version of libitm.so that will get linked should be the one that just got built when I built my modified gcc.
But in reality it seems both native gcc and mine are using the same libitm, /usr/lib/x86_64-linux-gnu/libitm.so.1.
I only have a rough grasp of gcc internals as they apply to our project, but this is my understanding:
Our changes tell one compiler pass to conditionally insert our own "function builtin" instead of one it would normally use, and this is / becomes a "symbol" which needs to link to libitm.
When I use the new gcc to compile my program, that pass detects those conditions and successfully inserts the symbol, but then at runtime my program gives a "relocation error" indicating the symbol is not defined in the file it is searching in: ./test: relocation error: ./test: symbol _ITM_S1RU4, version LIBITM_1.0 not defined in file libitm.so.1 with link time reference
readelf shows me that /usr/lib/x86_64-linux-gnu/libitm.so.1 does not contain our new symbols while ~/project/install-dir/lib64/libitm.so.1 does; if I re-run my program after simply copying the latter libitm over the former (backing it up first, of course), it does not produce the relocation error anymore. But naturally this is not a permanent solution.
So I want the gcc I built to use the shared libs that were built along with it when linking. And I don't want to have to tell it where they are every time - my feeling is that it should know where to look for them since I deliberately built it somewhere else to behave differently.
This sounds like the kind of problem any amateur gcc developer would have when trying to make a dev environment and still be able to use both versions of gcc, but I had difficulty finding similar questions. I am thinking this is a matter of lacking certain config options when I configure gcc before building it. What is the right configuration to do this?
My small understanding of the instructions for building and installing gcc led me to do the following:
cd ~/project/
mkdir objdir
cd objdir
../source-dir/configure --enable-languages=c,c++ --prefix=/home/myusername/project/install-dir
make -j2
make install
I only have those config options because they seemed like the ones closest related to "only building the parts I need" and "not overwriting native gcc", but I could be wrong. After the initial config step I just re-run make -j2 and make install every time I change the code. All these steps do complete without errors, and they produce the ~/project/install-dir/bin/ folder, containing the gcc and g++ which behave as described.
I use ~/project/install-dir/bin/g++ -fgnu-tm -o myprogram myprogram.cpp to compile a transactional program, possibly with other options for programs with threads.
(I am using Xubuntu 16.04.3 (64 bit), within VirtualBox on Windows. The installed /usr/bin/gcc is version 5.4.0. Our source at ~/project/source-dir/ is a modified version of 5.3.0.)
You’re running into build- versus run-time linking differences. When you build with -fgnu-tm, the compiler knows where the library it needs is found, and it tells the linker where to find it; you can see this by adding -v to your g++ command. However when you run the resulting program, the dynamic linker doesn’t know it should look somewhere special for the ITM library, so it uses the default library in /usr/lib/x86_64-linux-gnu.
Things get even more confusing with ITM on Ubuntu because the library is installed system-wide, but the link script is installed in a GCC-private directory. This doesn’t happen with the default GCC build, so your own GCC build doesn’t do this, and you’ll see libitm.so in ~/project/install-dir/lib64.
To fix this at run-time, you need to tell the dynamic linker where to find the right library. You can do this either by setting LD_LIBRARY_PATH (to /home/.../project/install-dir/lib64), or by storing the path in the binary using -Wl,-rpath=/home/.../project/install-dir/lib64 when you build it.
I am trying to build a cross compiler for PowerPC e500mc with target powerpc-e500mc-eabi. As some websites mentioned, i built an bootstrap compiler first. and then tried to compile newlib with it. But i got some error like,
/bin/sh: powerpc-e500mc-eabi-cc: command not found
I want to know, can we directly compile GCC cross compiler without newLib. Also, can anyone tell me the exact pre-requisites for powerpc e500mc architecture. I have GMP, MPC, MPFR, BinUtils not sure whether newLib required or not.
You can build gcc without any C library, there is no need for newlib. Please find a list of dependencies from a crosstool-ng build below.
1. Alternatives
You can build everything manually, what you obviously attempted to do. This is possible but there are various constraints to keep in mind, for instance builds might fail if your filesystem is not case sensitive or if you build inside your source directory. Besides all the dependencies and their versions.
You let crosstool-ng build your cross toolchain. Crosstool-ng is mostly self-contained with few external dependencies. It will download and build all dependencies in the right versions for you, and it comes with various sample configurations. It will check for obstacles like a case insensitive filesystem. It lets you configure your cross toolchain in a similar way you configure a Linux kernel. I've built various cross toolchains by means of it for several years on several host systems without trouble, including Linux, OS X (homebrew) and Windows (cygwin). You find it here: http://crosstool-ng.org/.
I'm going to line out the steps that it takes to build a cross toolchain by means of crosstool-ng. I tested this setup on Windows (cygwin) today with the crosstool-ng from git.
2. Download crosstool-ng, build and install it.
Follow the steps in
docs/2 - Installing crosstool-NG.txt.
3. Configure your cross toolchain.
Follow the steps in
docs/3 - Configuring a toolchain.txt.
4. Example
mkdir powerpc-e500v2-eabi
cd powerpc-e500v2-eabi
ct-ng powerpc-e500v2-linux-gnuspe
The last step will create a configuration from a sample which I thought is similar enough to what you want. In the next step, we will adapt this configuration, the criteria are
no operating system
no C library
EABI
you want e500mc, the sample is e500v2. I leave it up to you to adapt the configuration to it.
4.1. Configuration
ct-ng menuconfig
Operating System
Target OS
Select bare-metal
C-library
C library
Select none
Target options
ABI
Select EABI
C-Compiler
gcc version
Select 5.1.0 (the sample configured a gcc 4.6.4
here)
Deselect C++ (since you do not want a C library
either and it'd pull extra dependencies)
Debug facilities
Deselect gdb (unless you want it, can pull extra dependencies)
4.2. Build
ct-ng build.4
Crosstool-ng will download and build the following dependencies.
gmp-6.0.0a
mpfr-3.1.2
isl-0.14
mpc-1.0.2
binutils-2.25
gcc-5.1.0
It will install the cross toolchain in
${HOME}/x-tools/powerpc-e500v2-eabi.
You can specify a different install prefix in the configuration.
This build issue can be fixed by creating a symlink powerpc-e500mc-eabi-cc pointing to powerpc-e500mc-eabi-gcc.
I'm trying to build gcc 4.7.2 under MinGw with cc,c++,fortran,objc and java.
When it reaches to the compiling libgcc at final linking level it gives an error.
cannot find dllcrt2.o
Followed by
cannot find -lmingwthrd,-lmingw32,-lmingwex,-lmoldname,-lmsvcrt,-ladvapi,-lshell32,-luser32,-lkernal32
I think this is because ld.exe couldn't locate /mingw/lib dir.Is there any solution to fix this?I tried googling but nothing worked.
What is wrong with the MinGw compilers they ship that you are trying to build your own set?
Configuring/building GCC is a complex task (even more so in such a hostile environment), a small slip up can lead to results like the ones you see. Recheck each step.
Grab a copy of the source for MinGw and study how they do it, check their documentation.
So I just downloaded mingw-w64-bin_i686-mingw_20110410.zip from here (GCC 4.7 apparently), and discovered it had a very recent version of the GCJ compiler.
I tried using it, but apparently gcj requires ecj1.exe, which is the Eclipse compiler for Java... so, where do I find a compatible version of the binaries of ECJ and the associated Java libraries that are needed (libgcj, etc.)?
Ideally this would be found on the MinGW-w64 project page, but it doesn't seem to exist.
(I've already tried copying them from a slightly older GCC version; it doesn't work.)
The cause for an openSUSE version of the gcc is basically this:
If the configure step of the compilation of gcc did not find the ecj.jar
file, ecj1 will be missing at the time when gcj, which has just been build,
is called.
ecj.jar can be taken from ftp://sourceware.org/pub/java/ecj-4.8.jar
for example.
The two options are:
i) Put ecj.jar in $HOME/share/java/ecj.jar, reconfigure gcc with
./configure .... --with-ecj-jar=$HOME/java/ecj.jar
and recompile gcc. Future compilations with that gcc will not require
ecj1 .
ii) Put ecj.jar in $HOME/share/java/ecj.jar and create ecj1(.exe)
through a compilation like
gcj -o$HOME/bin/ecj1(.exe) --main=org.eclipse.jdt.internal.compiler.batch.GCCMain $HOME/share/java/ecj.jar
assuming that the $HOME/bin is in the PATH for subsequent calls of gcj.
The thing that is actually "broken" here the fact that gcc 4.8.* is not shipped
by default with ecj.jar at some standard place.
That is a very old version of a MinGW-w64 toolchain.
I would suggest downloading one of my builds, I've had reports of gcj working (without libgcj, which does not work on Windows), although I can't seem to find a link to the discussion I had long ago with a user. The user's case had something to do with creating a JNI interface or something, which didn't require libgcj.
My old builds can be found here for 32-bit and here for 64-bit. I checked the 4.8 release build, and it contains the gcj compiler.
Would you be opposed to downloading the source and building it? I looked over the build doc in basic and advanced build docs. I didn't see anything about the GCJ compiler or ECJ, but you'll need gcc 4.5.1 in order to build it.