Good day
Currently, I'm working on an embedded device based on arm-linux. I want to build GCC for my target architecture with Glibc. GCC builds successful, but I have trouble with Glibc build.
I use the latest version of Glibc (ftp.gnu.org/gnu/glibc/glibc-2.12.1.tar.gz) and port for them (ftp.gnu.org/gnu/glibc/glibc-ports-2.12.1.tar.gz)
my configuration line:
../../glibc-2.12.1/configure --host=arm-none-linux-gnueabi --prefix=/home/anatoly/Desktop/ARM/build/glibc-build --enable-add-ons --with-binutils=/home/anatoly/Desctop/ARM/toolchain/arm/bin/
configuration script work fine, but i get some compile error:
...
/home/anatoly/Desktop/ARM/src/glibc-2.12.1/malloc/libmemusage_pic.a(memusage.os): In function me':
/home/anatoly/Desktop/ARM/src/glibc-2.12.1/malloc/lmemusage.c:253: undefined reference to__eabi+read_tp'
...
I also tried using the old version (2.11, 2.10) but have the same error.
Does anybody know the solution for this problem?
Use a precompiled toolchain, like those provided by code sourcery.
If you want to make your own, optimised (premature optimization is the root of all evil), use crosstool-NG, which is a tool dedicated to cross-compilation toolchain building.
If you are not convinced, and want to do everything with your own hands, ask your question on the crosstool-NG mailing list.
Try substituting arm-linux-gnueabi for arm-none-linux-gnueabi. Check that a compiler, loader etc. with the prefix you used for "host" exist on your path.
Related
I'm attempting to cross-compile from Linux (NixOS) to Windows and encountering some frustrations.
There seem to be two parts that together are breaking the build:
Code in my Rust project requires multithreading, and as such requires a version of libpthread for Windows.
To build properly, I need a build.rs file. For some reason, Rust requires a version of libpthread for Linux for that.
What's the problem? Well, the build.rs has to be built with regular GCC and not MinGW because it needs to execute on my system. But for some reason, GCC is attempting to link to the Windows libpthread library instead of the system one, and as such is failing with an error about not supporting the library format.
(Failed) Alternatives
If I remove the build.rs, the project builds fine. Unfortunately, I need it for full functionality.
If I remove the Windows version of libpthread the build.rs builds and runs correctly, but MinGW fails with a missing library error when building the rest of the project.
Solution Paths?
Either I have to figure out why GCC's linking to the wrong version of libpthread, or I have to disable -lpthread entirely for the build.rs. I have no idea why it would need pthreads, considering for testing I stripped it down to only fn main() {}.
I have no idea where to start on either of these, and I've already spent a couple of days getting the problem down to this. I'd appreciate some help!
I am using Buildroot (2017.02.5) to build a custom cross compilation toolchain. I have two buildroot configurations; one to build the RFS and one purely to build a toolchain. I have things configured this way because I don't want the toolchain to be rebuilt unless I intentionally rebuild it- the configuration which builds the RFS references this toolchain as an external toolchain.
Generally, the built toolchain works fine, but I have some existing applications (Linux userspace) which #include's <openssl/md5.h>. When I try to compile this, I get a "<openssl/md5.h>: No such file or directory" error, which is expected because the sysroot dir of the generated toolchain does not contain an openssl directory.
How can I make buildroot include openssl in the toolchain? All searches I have done seem to point to cross compiling openssl for my embedded target, which is not an issue. The issue is that I need to include it in the toolchain.
I have Target packages --> Libraries --> Crypto --> openssl set to y, but I don't think this makes any difference in this scenario since I believe it relates only to the RFS (and the defconfig in question does not build an RFS, only a toolchain).
I could compile OpenSSL outside of the buildroot tree and install it to the sysroot dir, but this doesn't seem correct as it would pollute sysroot.
I'm sure I'm missing something simple here- any help would be appreciated.
After some further reading of the buildroot documentation (which is very good), I figured that packages selected under Target packages do in fact get pushed into the sysroot of the toolchain (or are supposed to at least) which would make sense. The reason this didn't appear to be working was because I was doing a make toolchain as opposed to make all (or just a simple make). The packages didn't get built with the former, so they weren't in the sysroot of the toolchain.
I'm trying to build llvm 3.4.2 on CentOS 7 with GCC 4.8.5. I can successfully build all of the libs, except libLLVM-3.4.so doesn't want to build - everything else gets created except that one. What's worse is I've managed it previously in the past, but can't do it again now.
I've tried both with CMake and autotools, with the same result. Is there an option I'm missing? Or some error in configuration possibly?
Found it - it was the --enable-shared flag to configure which causes the main shared library to be built. Confusingly this does something completely different from the CMake BUILD_SHARED_LIBS variable, which causes each of the individual libraries to be built as a shared library instead of static, but doesn't build the main shared library.
You can use
LLVM_BUILD_LLVM_DYLIB
when building from source.
The official docs say
BUILD_SHARED_LIBS is only recommended for use by LLVM developers. If you want to build LLVM as a shared library, you should use the LLVM_BUILD_LLVM_DYLIB option.
source: https://llvm.org/docs/CMake.html
I wonder if anyone did managed to build the fresh SDL2 with the toolchain of the Android NDK(r8d).
SDL2 seems to be very close to the release (since yesterday it isn't "UNDER CONSTROCTION anymore: http://hg.libsdl.org/SDL/rev/0a3d2ec7af6d). It comes with an Android.mk and just compiles fine following the instructions in the bundled README.android file. My question is whether there's really no working automake based build is available or will be available to compile it on Android, or something's wrong with my toolchain setup?
I have installed the NDK toolchain following the instructions of the documentation located at $NDK/doc/STANDALONE-TOOLCHAIN.html. I'm using gcc 4.6. Here's one environment i use:
#!/bin/sh
export TOOLCHAIN=$HOME/Android/android-14-arm
export PATH=$TOOLCHAIN/bin:$PATH
export SYSROOT=$TOOLCHAIN/sysroot
export CROSS_COMPILE="arm-linux-androideabi"
export CC=$CROSS_COMPILE-gcc
export CXX=$CROSS_COMPILE-g++
export CPP=$CROSS_COMPILE-cpp
export CFLAGS="-march=armv7-a -mfloat-abi=softfp -mfpu=neon"
export LDFLAGS="-march=armv7-a -Wl,--fix-cortex-a8"
echo "Compiler set up for ARM 14"
The configure params:
./configure --host=arm-linux-androideabi --prefix=$SYSROOT/usr/local
With the same configuration i successfully built libjpeg-turbo v8 and SDL_image.
The configure script recognizes the cross-compiler, and builds the makefile, however, it finds X11 support, can't see the OpenGL ES... The make fails:
In file included from /usr/include/features.h:378:0,
from /usr/include/sys/types.h:27,
from ./include/SDL_stdinc.h:35,
...
I checked the configure log, i have no idea where the "/usr/include" comes from.
But in fact, the generated makefile adds that line in the EXTRA_CFLAGS to the compiler.
The NDK doc refers the --with-sysroot=$SYSROOT as optional, i've included it to see if it solves the problem, but that didn't help.
As a last effort i manually edited the Makefile, fixing that reference, and now the compiler complained about X11.h.
AFAIK Android has nothing to do with X11, so i guess the whole build-tree completely inappropriate to use with NDK.
I have also tried a different configuration, found in an older thread here.
Neither defining -DANDROID -mandroid -fomit-frame-pointer nor changing back to -march=armv7-a -mfloat-abi=softfp -mfpu=vfp -mthumb" solved the problem.
On previous projects, i had to refresh config.guess, and config.sub in order to get my compiler recognized. SDL doesn't seem to use those. Furthermore no Makefile.ac or Makefile.am comes with SDL to work with, and no templates for other platform could be used for a good starting point to create my own makefile. Additionally, i've never had to deal with makefiles, i really have no chance to sort out these problems. Even if it succeeds, i will probably need a configure tool as well, since i have no idea how ndk-build manages to install SDL2 without configure scripts.
Compiling the SDL sources with the project together is the only working - but ugly solution. I would like to deploy the necessary lib and header files by make install.
I hope the solution is something really easy and obvious thing that i just didn't think about...
This issue has been fixed at http://hg.libsdl.org/SDL/rev/4e57cfd9fca8 and expected for the 2.0.4 release. Note there are newer revisions with some related fixes about defines.
I got a little confused when it comes to QT and cross compiling
appliations for my arm-linux:
So far I have a linux running on my AT91SAM9263-EK and an appropriate
filessystem including QT libs build via buildroot.
Also I have build QT-4.8 on my ubuntu.
Now I want to build an example application:
I created a makefile in an examples folder in QT on my ubuntu using
qmake; I used the given qmake.conf in mkspecs/qws/linux-arm-g++.
when executing make I get an error because it includes qatomic_i386.h
and the message "error: impossible constraint in 'asm'".
this header file does obviously not match to my arm toolchain.
my question:
how to configure Qt on my ubuntu to build Qt binaries for my embedded linux
on arm? Do I need to include any libs build by the arm toolchain?
any help is appreciated!
regards
EDIT:
I use the -spec flag and pass the path to "mkspecs/qws/arm-linux-g++" where a "qmake.conf" is located. I did not change anything in here because I dont know wich qmake variable are relevant to link to my arm related libs.
So the right compiler is used, which I could verify when the make process starts. But I observed that in a config file called qconfig.h there is an ARCH flag which is set to i386 but I didnt figure out how one can configure this. I dont think I should manually edit this file.
EDIT2:
someone knows whats behind the file qconfig.h?? should I adjust it manually?
I will solve it by myself :)
After struggling a while and scanning the web I got a little deeper involved how everything works together. I did not understand how to generate an executable for my ARM target device. I figured out two things:
do not add your QT path for X11 at the beginning in your PATH variable. this might mess up your cross compilation.
edit the qmake.conf correspondingly. add your libs build for the target device which in my case are located within buildroot. Add theses lines to your qmake.conf file:
QMAKE_CFLAGS += -O3 -march=armv5te
QMAKE_CXXFLAGS += -O3 -march=armv5te
QMAKE_INCDIR_QT = /home/user/arm/toolchain/buildroot-2010.11/output/staging/usr/include
QMAKE_LIBDIR_QT = /home/user/arm/toolchain/buildroot-2010.11/output/staging/usr/lib
I got it running now. thanks to everyone!
Yes, either you provide the Qt libraries in your toolchain or you tell qmake where to find them. Also, I suspect you're calling qmake without the -spec parameters. If you are using the qmake you find in your distribution, it will use the default spec, which is not arm I guess. Add the -spec parameter and point it to the arm mkspec. Also, make sure the generated g++ commands link to the correct Qt libs compiled for arm.
You shall install QtSDK for embedded linux befor you use it to build your application. I'm afraid you just have QtSDK for x86 right now. After QtSDK for embedded linux installed, it has qws/linux-arm-g++ as the default mkspace typically. If you don't have QtSDK for embedded linux, you can build it from source. Then run qmake to create Makefile for you application.
$QTDIR_FOR_ARM/qmake
Reference:
Installing Qt for Embedded Linux and Cross-Compiling Qt for Embedded Linux Applications