Santosh Sivaraj

• articles

### Tags

• process

As you would have noticed the beginning of the document, it was told the explanation will be based on the program we had written earlier in this series. The program was compiled using the command

$cc sample-source.c -lm  As you have probably noticed, we are linking the program with the math library. This was done intentionally to explain the dynamic loader. Any file can be identified using the file command. Running file on the a.out executable gives us some information on the type of executable. $ file a.out
a.out: ELF 64-bit LSB executable, x86-64, version 1
(SYSV), dynamically linked (uses shared libs), for
GNU/Linux 2.6.32, not stripped


The description shows that the executable compiled is a ELF (Executable and Linkable Format) type executable for a 64 bit machine and used shared libraries. To find out what are all the shared libraries that went into making the ELF executable can be found using the ldd program.

\$ ldd a.out
linux-vdso.so.1 =>  (0x00007fff1d368000)
libm.so.6 => /lib64/libm.so.6 (0x0000003cede00000)
libc.so.6 => /lib64/libc.so.6 (0x0000003cece00000)
/lib64/ld-linux-x86-64.so.2 (0x0000003ceca00000)


The output shows that our tiny executable uses the services of four different shared libraries. {The linux-vdso.so library will be explained elsewhere in this series). The libm.so is the math library that we linked in while compiling out program code. The next is the standard C library. The last library is the dynamic loader which will help during the start of the program execution to use other library functions that we are linked with. This section will describe the functions of the dynamic loader in detail.

Before we can proceed with the dynamic loader in detail, there are some background information that needs to be known about the compilation process.