I am curious about using dlopen in Linux to call shared libraries.
Suppose I want to use a shared library in C whose name is fileName.so. I am working in a 64bit Ubuntu Linux and I include dlfcn.h and use dlopen function to access the shared library.
When I use dlopen(fileName.so, RTLD_LAZY), a NULL handle is returned and shared library is not opened. However, when I use dlopen("./fileName.so", RTLD_LAZY) the dlopen does its job and opens the shared library. It seems that the main point is in using ./ before file name.
It is appreciated if help me figure out why I should use ./ in my code. Thanks
POSIX says that dlopen() has to know where to look for the file and leaves the behaviour when the file name does not include a / implementation defined. On Linux, if you don't supply a pathname (a name with a / in it somewhere), then dlopen() only looks in 'standard places', specified by environment variables such as LD_LIBRARY_PATH or via /etc/ld.so.conf (or /etc/ld.so.cache; see also ldconfig(8)) or in standard places such as /lib and /usr/lib.
When you specify the relative name ./fileName.so, it knows to look in the current directory, which is not normally a place it looks.
Note that you can run into some interesting issues on systems that support both 32-bit and 64-bit executables, with various conventions being used for the locations of the different classes of library. Other variants of Unix use vaguely related systems — mostly using dlopen() et al these days (historically, it was not always thus), and using a wide variety of environment variables (DYLD_LIBRARY_PATH, LIBPATH, SHLIB_PATH, LD_RUN_PATH, LD_LIBRARY_PATH_32, LD_LIBRARY_PATH_64, ...).
./ is a relative path to the .so file. It means that the file is in the current directory.
In *nix, by default, when given a file name without an absolute or relative path, dlopen will search for the library in a set list of default locations.
The "main point" is in using double-quotes in your second example:
dlopen("./fileName.so", RTLD_LAZY)
If you want to include your own library/filename enclose it in double-quotes. You don't even need the ./ for that, provided that the file is in the current directory, as ./ suggests.
As per the dlopen manpage's example:
handle = dlopen("libm.so", RTLD_LAZY);
if (!handle) {
fprintf(stderr, "%s\n", dlerror());
exit(EXIT_FAILURE);
}
the filename is enclosed in quotes.
Though, as specified by a previous answer, dlopen will look in the "standard" places for "includes". Another way of including a library that's inside your working directory (though, obviously, not a shared system library) is to use the preprocessor directive with the filename enclosed within double quotes:
#include <stdio.h>
#include <stdlib.h>
#include "myCustomLibrary.h"
Related
I am currently learning the C programming language and I'm having some issues with importing modules I created.
I created a small module to read with fgets and flush the buffer from stdin perfectly and I don't want to keep writing the code every single time. I just want to import this small module like I used to do in Python. I didn't knew how because I'm not using an IDE. I'm just compiling with gcc in terminal and using a text editor. I tried to search with Google,but in vain.
You should create a header for your module that declares the functions in the module – and any other information that a consumer of the module needs. You might call that header weekly.h, a pun on your name, but you can choose any name you like within reason.
You should create a library (shared or static — that's up to you) that contains the functions (and any global variables, if you're so gauche as to have any) defined by your module. You might call it libweekly.so or libweekly.a — or using the extensions appropriate to your machine (.dylib and .a on macOS, for example). The source files might or might not be weekly.c — if there's more than one function, you'll probably have multiple source files, so they won't all be weekly.c. You should put this code (the header and the source files and their makefile) into a separate source directory.
You should install the header(s) and the library in a well-known location (e.g. $HOME/include for headers and $HOME/lib for the library — or maybe in the corresponding directories under /usr/local), and then ensure that the right options are used when compiling (-I$HOME/include for the headers) or linking (-L$HOME/lib and -lweekly).
Your source code using the module would contain:
#include "weekly.h"
and your code would be available. With shared libraries in $HOME/lib, you would have to ensure that the runtime system knows where to find the library. If you install it in /usr/local, that is done for you already. If you install it in $HOME/lib, you have to investigate things like /etc/ld.so.conf or the LD_LIBRARY_PATH or DYLIB_LIBRARY_PATH environment variables, etc.
You need to create a header file (.h) with your function declarations types and extern variables. Then in the program where you want to use those functions include this .h file and and add the compiled .o file (with your functions) to the object file list. And you are done.
I wants to get the names of execute files in some directory in Linux.
How can I do it?
I tried to use opendir like this:
dir = opendir(directoryName);
I need to get only the names of the execute files.
I programming in C.
thanks :)
You should define what you mean by executable files.
That could be any file with its execute bit (it is the owner, group, or other) set. Then test with access(2) & X_OK and/or use stat(2).
That could also be only ELF executables. See elf(5); then the issue might be to check that a file could indeed be executed, which might be difficult (what about missing library dependencies? or ill-formed ELF files?). Maybe use libelf (and/or libmagic to do the equivalent of file(1) command).
To scan recursively a file tree, use nftw(3); to scan just a directory use opendir(3) & readdir(3) (don't forget the closedir!), then you'll probably need to build the complete file path from each directory entry (perhaps using snprintf(3) or asprintf(3))
See also Advanced Linux Programming
In my c++ program, I used Tcl library and linked libtcl8.5.so, but the target hosts don't have tcl8.5, so I copied the libtcl8.5.so and tcl8.5 dir which contains init.tcl there, and set the environmet variable TCLLIBPATH to path/to/copied/tcl8.5, but when my program call Tcl_Init, it failed and said “package not known”.
It seems the copied tcl8.5/ cannot be init correctly.
How can I solve this problem?
If you change the location of the script library directory (tcl8.5/ in your case), you need to tell the shared library part of Tcl where it is. You do this using the TCL_LIBRARY environment variable, which if set should contain the absolute path that is the location of that directory (technically, the directory that contains init.tcl).
In a normal installation of Tcl, the correct location of that directory is baked directly into the shared library, but when you move things round (or when you are running Tcl's make test) the environment variable allows you to override.
You might wish to look into alternate packaging mechanisms; there have already been a few questions in the tcl tag on this matter (but the usual favorite — a starkit — is probably not suitable for your case given the fact that the program is mainly C++).
Compiling on the shared CentOS server is not allowed. Therefore, I compile my program in my Debian computer, linking it with Debian's system libraries such as libc, etc. Then I upload my program and the Debian system libraries and my program works. The only problem is that setlocale() does not work at CentOS. CentOS has "en_US.utf8" installed and works on all programs except mine. I suspect that I have to also upload Debian's locale files ? How could I link my program to the Debian locale files ? I tried to use LOCPATH but I am unsure of how it works exactly. Which files do I have to link to and how ?
C program:
setenv("LOCPATH", "/", 1);
if (setlocale(LC_ALL, "en_US.utf8") == NULL) {
puts("not set");
}
I used a hex editor to modify the path to /usr/lib/locale/locale-archive which apparently is the only file that setlocale() uses according to strace. This method is dirty but it worked.
According to man LOCPATH, this environment variable is non-standard, so its use is not recommended. No examples are given anywhere of how to use it nor what is meant exactly by a path to "locale's object files".
I guess the only real way of modifying the path is a glibc modification and recompilation.
Quote: LOCPATH is an environment variable that tells the setlocale() function the name of the directory from which to load locale object files. If LOCPATH is not defined, the default directory /usr/lib/nls/locale is searched. LOCPATH is similar to the PATH environment variable; it contains a list of z/OS UNIX directories separated by colons.
So just specifying / and hoping that it does a recursive search will not work.
You could also produce a static binary and upload that to the host.
I am running Ubuntu 10.10 on an IBM R51 machine. When I access list.h to read it(manually/humanly) I open /usr/src/linux-headers-2.6.35-22/include/linux .
But when coding a C program in terminal, I cant invoke any #include because it is not in the default /usr/include folders.
When I change the statement to reflect the path by typing #include "/usr/src/linux-headers-2.6.35-22/include/linux/list.h" it returns errors as list.h in turn calls other header files which are mentioned as located in "linux" folder
The header files are as you must be aware:
"linux/poison.h", "linux/prefetch.h" and "asm/system.h"
So if I have to copy each, I can but prefetch in turn calls other dependencies, which are not present in /usr/include directory. I hope you understand.
How can I solve this problem?
Are you sure these headers are really what you need ? The standard C headers should be under /usr/include
Anyhow you need to pass the header search path to the compiler via the '-I' flag.
Pass the path via -I
-I/usr/src/linux-headers-2.6.35-22/include/linux
Then in your C code
#include "list.h"
Link to GCC manual & preprocessor directives
The header files you are using are designed for internal use of the Linux kernel. They were not designed to be used by a userland program.
If you MUST use these headers (the Linux kernel list implementation is brilliant), copy the headers into your program source directory. Copy each file that is referenced, edit each one to remove whatever assumptions exist about being used in-kernel, and recurse until you're finished. I might suggest to make your own prefetch() macro that simply does nothing, rather than try to untangle <linux/prefetch.h>. Do the same for <linux/poison.h>, and untangle <linux/types> and <linux/stddef.h> (not too hard here :) as best you can.
And also be sure you license your project GPLv2 (and specifically GPLv2, the Linux kernel's COPYING file is quite strict that GPLv2 is the only license that applies; there is debate whether the GPL allows specifying only one version, but that is the license Linus chose ages ago, and the license that is valid on all files unless specified otherwise).
adding -I/usr/src/linux is a no-go, since unsanitized header files are not meant to be used from user programs
you could manually copy list.h to your own project and sanitize
or use a library that is specifically for userspace and provides the same functionality (since you already used libHX elsewhere, you might want to continue reading into the linked list chapter)