Slackware :: Cross Compiling For Arm With Buildroot ?
Jan 27, 2011
I'm trying to build some boot and kernel images for my WD Mybook world edition, which uses an arm processor on my x86_64 host box. Ultimately i'd like to install slackarm. Lilo doesn't work on arm so you have to compile your own kernel and have the bootloader statically linked to kernel inorder to boot the OS. So i'm using my slackware install to cross compile the boot loaders (u-boot) and kernel image (uImage).
I've successfully compiled the buildroot environment which has generated the arm-linux-* binaries, eg, gcc 3.4.2 for arm.
Code:
These are in a build folder (shortened in text below), which i add to my PATH
Code:
I then move to the directory where i want to build the stage1 boot loader, which is supplied in the GPL code from WD.
Code:
However the build fails with a segfault, which looks to be caused by the incorrect libraries and compiler being used:
Code:
It seems to be using my x86_64 libs and gcc 4.4.4 which is the host compiler not the target.
I've tried setting ARCH=arm but got the same results.
Do i need to export some additional variables or unset some to prevent the host files being used instead of the targets?
I am trying to cross-compile my Qt application for armhf on Jessie, so what I did was install qt5-default:armhf, but when I tried running qmake I found out that it contains also an armhf version of qmake so I was not able to run it.Then I installed the amd64 version to have at least a runnable version of qmake, but it was only to find out that qmake contains only hardcoded paths, so I could not use it for the armhf libraries, well I could but I had to run qmake (amd64 version) and afterwards open the makefile and change all paths from amd64 to armhf.
But as this is not so convenient I was wondering if there is a normal way to do this, I know you can compile the source yourself but I had quite a lot of problems with dependencies there. Is it possible with the standard repository binaries or should I just go back to compiling the source myself? Maybe I can compile only the building tools without building all the libraries I already got from the repository.
I try to cross compile udev-160 for arm.I use the buildroot toolchain.I work on a PC i386 with the Debian system.I started by running "configure" script to generate the makefile:$./configure --host=arm-linux
but I obtain the following error :checking for acl_init in -lacl... no configure: error: libacl not found
I don't know how to install the libacl for ARM. I can't found the libacl when I configure the buildroot toolchain.
I have synology nas box DS710+ that has intel atom processor inside. i've installed ubuntu under virtual box, downloaded needed toolchain for x86 processors, and decided to cross-compile openvpn.
I 've started from compiling lzo library, but first question that appeared in my head - how i have to move all compiled binaries to my synology?
1. configure - okay 2. make - okay 3. make install - this is not necessary?
Will --prefix parameter helps me? or do i have to use chroot? cause there is a plenty amount of files that i need to move to my synology.
I developed a test program(my first under linux) some thing like g++ -I$(BOOST_INCLUDE) -L$(BOOST_LIBS) threading.cpp -lpthread -lboostThread -lOtherLib -o threadingTest.out...This thing works in GCC4.5, but to compile it under windows visual studio 2008, do I have to create a new project file and then include all libs? In short, is there any IDE or utility, where I just compile it either on linux or windows without going and modifying project options again and again ?
i am trying to compile too the sqlite for arm9 - mini2440.could you please tell me what was the steps you make? im read that book but im still having some trouble while i compile i have been trying to do the compilation.
i am trying to cross compile sqlite3 with following options(i got these options from previous threads in these forum)./configure --host=arm-linux CC=/usr/local/arm-linux/bin/arm-linux-gcc AR=arm-linux-ar STRIP=arm-linux-strip RANLIB=arm-linux-ranlib while executing make command i am getting following error
I have been installing gentoo on my desktop for the past few weeks. (Not a whole lot of time to get to it) and I must say that I wished I had made the move a whole lot earlier. Anyway, my next canididate for gentoo would be my ibook g3 which has a slow processor. I have heard that for these systems cross compiling is a lot better of an option than compiling on the system itself. I have read the cross-compiling guide and am left a bit confused. Could I use my thumbdrive to transfer compiled binaries from my desktop to my laptop and install the packages that way? Also, what would I do when updates come out for my laptop. Is there any way of monitoring the package statuses?
I have written a program that I would like to cross compile for x86 and x86_64 architectures. I have tried google and the search function here to no success, most information I find is too specific (instructions for a specific program), or dealing with cross compiling for windows on linux.Does anyone know of a tutorial dealing with straight making a 32 bit binary on a 64 bit processor (both are intel)?
I wanted to cross compile sms server tools(url)for arm-linux with arm-linux-gcc compiler but I can not this is all information that I have from it's error:
[root@localhost smstools3]# cat Makefile # Makefile # If you change destination of executables, remember to change # startup script (init.d/sms3) too. #BINDIR=/usr/local/bin BINDIR=/mnt/sd/Utility/sms/ ///I changed here for cross compilation code....
I just finished building a cross-compiler for i386-elf. But when I try to use it, the terminal gives me this error:
Code: Select all/home/isaac/Cross-compiler/lib/gcc/i386-elf/4.8.2/../../../../i386-elf/bin/ld: cannot find crt0.o: No such file or directory /home/isaac/Cross-compiler/lib/gcc/i386-elf/4.8.2/../../../../i386-elf/bin/ld: cannot find -lc collect2: error: ld returned 1 exit status
I am cross compiling a linux kernel for the mips platform. I have installed the tool chain and everything. When i try to build the kernel i get the following error message.
HOSTCC -static scripts/basic/fixdep scripts/basic/fixdep.c: In function traps scripts/basic/fixdep.c:377:2: warning: dereferencing type-punned pointer will break strict-aliasing rules scripts/basic/fixdep.c:379:4: warning: dereferencing type-punned pointer will break strict-aliasing rules [Code].....
If i set up the toolchain on a fedora 13 machine with the exact same procedure, the cross compilation works fine. I can't figure out how this can be fixed. I am currently considering downgrading to fedora 13 due to this.
I'm a newbie when it comes to embedded linux / cross compiling. Never had to have anything to do with it b4 I bought my NAS. The Nas only has a 500MHZ ARM9 so compiling anything is painfully slow. Also only has 128MB ram. I managed to cross compile the kernel but am not sure how would you cross compile something like HPLIP for example.
I wonder if there is an easy way to accomplish the following: I made a tool and part of it needs to be compiled. It would be easy if I can automate compilation for multiple platforms (x86 / x86_64 / RHEL / SLES / Ubuntu). Is there a way to do this automated on a single host? Or should I set up an os for each distro? Also is there a way to generate RPM and deb files also on the same host without having to edit the meta files each time a version changes?
We have a device here at work running an Arm processor, I have a cross-compiler for the device, it was distributed to our company with the linux distribution we are using on the device (montavista linux). I would love to be able to cross compile some software for this device, however, I really suck at this kind of thing.
Whenever I try to cross compile software, I usually get an error like "unrecognized host platform" (I'm paraphrasing obviously), but the toolchain is not that much different from a generic Arm toolchain, just with some extra libraries installed and stuff. I tried using other arm toolchains, but those are a crapshoot, they compile fine, but they don't always run correctly on the device (for instance, busybox).
The reason I get the unrecognized platform error is because the toolchain has a weird prefix on it (arm_v5t_le-) which the configure script can't handle, but if it would just try to cross compile it like it does with other arm targets, it would work fine. How do I make it work?
I wanted to know how can I cross compile SMS SERVER TOOL for an embedded computer and make just one binary file for it or how can I change all of its default files places like its demon and object file and gather all of them to one directory to execute and use and run.let me explain it better for you : I have an embedded computer with Linux OS that its file system is read only and I can not add any file to /usr /lib and ..... and I can just mount a SD memory card to it and copy all of my programs to it and run them from there as you understand I have two choices to choose, first make one big binary file for each program that I am doing it now and it is not a suitable solution and the second is finding the way to change default place of shared object file of my program.now you tell me what can I do to solving this problem.
I have a need to recompile mesa. The --enable-32-bit option falls over every header and basic 32 bit lib, and won't compile. But if I try to install 32 bit packages, I could slap a 32 bit binary straight on top of my 64 bit one. What's the accepted way of doing this? I can't seem to find a package in Slackware 13.0
i try to compile freej on slackware 13.1 the problem is with ffmpeg >= 0.6 i have this error
Quote:
In file included from /usr/include/libavutil/avutil.h:81, from /usr/include/ffmpeg/avcodec.h:30, from ../src/include/video_layer.h:22, from ../src/include/impl_layers.h:48,
I'm running 13.1 64bit with alienbob's multilib extension and want to add openmotif. I used Eric's Slackbuild from [URL] applying it to openmotif-2.3.3 and succeeded in compiling the 64bit-version.Next I tried to compile in 32bit-mode, following alienbob instructions:
move lib64 to lib in openmotif.Slackbuild export ARCH=x86_64 . /etc/profile.d/32dev.sh ./openmotif.Slackbuild
linking stops with error-message
Quote:
/usr/lib64/gcc/x86_64-slackware-linux/4.4.4/../../../../lib/crt1.o: In function `_start': /glibc-tmp-88b9835d8519e3638ef4d116a6932cb2/glibc-2.11.1/csu/../sysdeps/i386/elf/start.S:115: undefined reference to `main' collect2: ld returned 1 exit status
Specifically, where does /usr/lib64/gcc/ come from?
Having trouble compiling the regular par2cmdline-0.4, I stumbled upon [URL] modified version of par2cmdline-0.4 with tbb which allows for concurrent processing of files. I have tried compiling this with no luck on my system and was wondering if anyone has had any luck. The ./configure goes through fine, and tbb is detected and used however it craps out just moments after issuing the make command. I have tried both tbb 2.2 and 3.0, but to no avail.
