I statically compiled and linked a program in an up-to-date Linux machine, and ran it in another Linux which is 9 years old. It gave me an error "FATAL: kernel too old" and quit. Specifically, the new one is Fedora 18 (gcc 4.7.2, glibc 2.16, kernel 3.7.2) and the old one is RHEL4.8 (glibc 2.3.4, kernel 2.6.9). Since it's static linking, glibc version shouldn't matter. I guess the problem here is that the program calls system calls that's not in the old kernel.
If development on the old system is not an option, how can I build the program in the new system and run in the older (or even better, both)? I was looking for a way to run gcc in a compatible mode, which only calls old system calls. No luck yet.
The easiest option is to always build on the older system.
Alternatively, copy the glibc headers and static libraries from the old system to the new and link against those.
If that doesn't work, you'll have to rebuild glibc with --enable-kernel=2.6.9 or something like that.
Related
I'm cross-compiling an application for aarch64 on my x86 Ubuntu Bionic system, and I have problems with glibc version mismatch. My cross-compile toolchain was using v2.27, while the system that is to run the application has v2.24. I thought that it might be due to my toolchain having a too high version, so I decided to downgrade.
After removing all previous cross-compilation installs, I installed gcc-4.8-aarch64-linux-gnu (as I had successfully cross-compiled the application with this version on a different host system), thinking that it would install an older aarch64 version of glibc to /usr/aarch64-linux-gnu/lib/. However, again, v2.27 was installed (I verified that this directory didn't exist before installing the new cross-compilation toolchain).
So my question is twofold:
What determines which aarch64 version of glibc is installed on my system when installing gcc-4.8-aarch64-linux-gnu? Is it directly tied to my own system's x86 version of glibc?
Is there a correct way to install the aarch64 version of glibc v2.24 (or lower) on my system?
I concur with your hypothesis. After battling similar symptoms for 40 hours straight, I've discovered this confirmation:
https://packages.ubuntu.com/impish/gcc-10-aarch64-linux-gnu
https://packages.debian.org/bullseye/gcc-aarch64-linux-gnu
Note that Ubuntu 21.10 (Impish) and Debian 11 (Bullseye) have packages for a gcc 10 cross compiler. Be wary of the very confusing fact the Ubuntu's default package is actually gcc 11, but Debian 11's default is gcc 10. The similar version numbers of Debian and gcc are a coincidence. Also ignore for now the fact that Ubuntu's package is gcc 10.3.0 and Debian's is gcc 10.2.1.
Focus instead on the recommendations and dependencies of each package. Ultimately the Ubuntu package calls up libc >= 2.34, while the Debian package calls up libc >= 2.28.
Sure enough, when I cross-compile from Impish on x86 for Bullseye on aarch64 (despite having a complete SYSROOT for the target), I get this at runtime:
/lib/aarch64-linux-gnu/libc.so.6: version 'GLIBC_2.34' not found
But your question remains, is there any tie between the host libc and that used by the cross-compiler? The answer is a definite maybe.
See this excellent answer and links for an overview of a cross-compiler. The take-away:
You don't just cross-compile glibc, you need to cross-compile an entire toolchain. Toolchain components are ALWAYS: ld + gcc + libc + gdb.
So the C library is an integral part of the cross-compiler.
What shenanigans then, are going on when you install gcc-aarch64-linux-gnu? It's just a compiler - only one of the four parts of a toolchain.
Well apparently there's some flexibility. Technically, a cross-compiler can be naked. That's typically only useful when you're compiling an operating system, rather than an executable that runs on an operating system. So you can construct special toolchains for special purposes.
But for the standard purpose (cross compiling for Linux on another architecture) you want a typical toolchain. Which is where the package's dependencies and recommendations come in. A gcc is always in want of an ld which is always in want of a libc, and the ménage à trois is intimate. In fact, gcc is built with libc using ld in a complex do-si-do. See this example from a great guide by Preshing on Programming:
It's possible to force separation and link to other libraries, but it's not easy.
For example, the linker you use has a set of default search directories that are baked in. From the fine manual:
The default set of paths searched (without being specified with -L) depends on which emulation mode ld is using, and in some cases also on how it was configured.
And it gets more intwined. By default, gcc will call on a dynamic linker whose location is hard-coded. For a cross-compiler, it might be something like /lib/ld-linux-aarch64.so.1. Not only that, the executable may also end up with the hardcoded path, as its program interpreter.
Again, if you're careful you can tear apart the toolchain and override things. But not only is it tricky to enforce, particularly if you have a complex build, the multitude of combinations of options and paths means there are also often bugs. So your host environment can easily leak into your cross-compiling toolchain.
So in summary, cross-compiling requires a toolchain. While pulling a cross-compiler from a package manager seems like an easy and legitimate thing to do, it comes with a lot of implicit baggage. You can either carefully follow the package dependencies to check what version you're getting, or use one of the many dedicated toolchain environments, such as crosstool-NG.
I'm trying to download glibc 2.23 sources and build them on my Ubuntu system.
I need to build that specific version from sources for getting modified version of glibc customized for my research, and it will be used only within my research apps using the loader environment variables (e.g., LD_PREDLOAD or LD_LIBRARY_PATH).
But, when building it as following, I got a huge file as an output (libc.so weights about 11MB):
download the sources to some local dir (let's say /tmp/glibc/)
create new directory for build results (/tmp/glibc/build)
run configure from build dir:
< build-dir >$ ../configure --prefix=< build-dir >
As a result, the build process will produce libc.so file under build-dir with a size of 11MB.
Is there anyway to reduce the size of the built libc.so?
p.s.
Here are my system details:
Linux version 4.4.0-93-generic (buildd#lgw01-03) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.4) ) #116-Ubuntu SMP Fri Aug 11 21:17:51 UTC 2017
Thanks :)
Building glibc from source could be a bad idea. See this and some comments there. Its current version is GNU libc 2.26... Consider instead upgrading your entire Ubuntu distribution (Ubuntu 17.10 should be released in a few weeks, end of October 2017)
../configure --prefix= build-dir
is a misunderstanding of the role of --prefix in autoconf-ed software. It relates to where the software is installed, not to its build directory.
(and I don't know exactly what should be your --prefix since libc is so essential to your system, perhaps it should be --prefix=/ but you should check carefully)
Is there any way to reduce the size of the built libc.so?
You might use (very carefully) strip(1), but you risk breaking your system.
And you might not care about reducing the size of libc since it is used (and shared) by almost every software on your Linux system!
BTW, consider also musl-libc. It can cohabit nicely with GNU glibc, and in practice is used only by programs built with musl-gcc (provided by it).
If you are doing some research, it would be reasonable to work in a chroot(2)-ed environment. See also schroot. You could install with the help of make install DESTDIR=/tmp/instmylibc then copy that /tmp/instmylibc appropriately. Read more about autoconf
PS. Be sure to at least back up your important data before such dangerous experimentations. I don't think that the size of your libc.so should be a significant concern. But you need to use chroot, perhaps with the help of debootstrap during installation of the chrooted environment.
I want to build a application which supports eBPF on CentOS 7 (the kernel version is 3.10.0):
if(setsockopt(sock, SOL_SOCKET, SO_ATTACH_BPF, prog_fd, sizeof(prog_f)) {
......
}
So I download a 4.0.5 version, make the following configurations on:
CONFIG_BPF=y
CONFIG_BPF_SYSCALL=y
Then follow this link to build and install a 4.0.5 kernel.
After executing make modules_install install, I find there is still no SO_ATTACH_BPF in <asm-generic/socket.h>, so the above code can't be compiled successfully.
How to build Linux kernel to support SO_ATTACH_BPF socket option?
In my setup, which is based on Fedora 21, I use very similar steps to those you linked to compile and install the latest kernel. As an additional step, I will do the following from the kernel build tree to install the kernel header files into /usr/local/include:
sudo make INSTALL_HDR_PATH=/usr/local headers_install
This will cause both the stock kernel header files to remain installed in /usr/include/{linux,asm,asm-generic,...}, and the new kernel header files to be installed in /usr/local/include/{linux,asm,asm-generic,...}. During your test program compile, depending on which build system you use, you may need to prefix gcc/clang with -I/usr/local/include.
Your newly installed kernel supports SO_ATTACH_BPF, but your current libc package doesn't now about that (as you mention, distro's native 3.10.0 kernel lacks of given option support).
You need to update libc package as well for use new kernel's features in user space programs.
Glibc 2.10(or any >2.10) with compile flag PER_THREAD and ATOMIC_FASTBINS behaves totally different then glibc 2.10 without those flags.
If my Linux is using glibc 2.10 I still don't know the exact version because it doesn't say anything about compilation flags. Ubuntu may use those flags in theirs glibc and Debian not?
How to list used compilation parameters, having glibc shared library file?
You won't find this information in /lib/libc.so.6. Though, if you're running Debian or Ubuntu you can still grab the source package (apt-get source libc6) and have a look at debian/rules file.
You can also write a quick test that checks glibc behavior and conclude if it has been compiled with these flags or not.
I built gcc 4.4.6 (to use CUDA) on a fast server, it takes about 10 min. However, on my own desktop, it takes kinda for ever to compile.
So both machines are 64 bit Linux, although 1 is Ubuntu while the other is Arch Linux. Arch Linux has new kernel version.
So on the server, I installed the built gcc-4.4.6 to /opt. And I just copy /opt/gcc-4.4.6 to my PC's /opt/gcc-4.4.6.
em, seems like it doesn't quite work, when I tried
./x86_64-unknown-linux-gnu-gcc ~/Development/c/hello/hello.c
it shows
x86_64-unknown-linux-gnu-gcc: error trying to exec 'cc1': execvp: No such file or directory
So what can I do now?
Thanks,
Alfred
If the systems are similar enough, you could compile GCC on the big machine (don't forget that GCC needs to be configured and built in a directory outside of its source tree), then run make -j3 all and then make install DESTDIR=/tmp/gccinst/ and copy that /tmp/gccinst directory to your small machine, and finally copy it into the root filesystem (on the small machine).
However, GCC 4.4.6 is quite old today, if you are compiling GCC try to compile GCC 4.6.2 (or 4.6.1 at least).
And (shameless plug for my work) if you compile a GCC 4.6, please enable plugins on it, then you might try the GCC MELT [meta-] plugin (MELT is a high level domain specific language to ease the development of GCC extensions).