Hi:
May someone share knowledge about 'signals'?
I have read some documentation about it but I still don't find how exactly is a 'signal' routine executed. Does this routine gets executed at 'owner' process context? At the beginning or end or in the middle of 'owner' porcess? Or at another 'process' context?
Please help.

Welcome to LQ.

Signals are a form of interprocess communications. The fact that a signal is sent comprises the total communication. In code you can set up a signal handler, and when the given signal occurs the signal handler gets called, the example link below does exactly what I'm describing. Confusion may be sometimes because a lot of handlers are automatically established. For instance if you create a child process using fork(), the parent process will receive a signal if/when the child process terminates. But in many cases, new coders may not understand that they need to look for that terminate signal from the child, or that it even is part of that programming mode, and then also not understand the method by which that signal may arrive.

Why not review the example code here and try it out on your own.

Signals interrupt the running process and run the signal handler in the process context. When they return the process resumes.

Thanx for your answer. I made the mistake to use the 'owner' word instead of 'target' (process).
I have read your explanations and links, however I still have these questions:
1) So, a 'signal' handler execution is done when the 'target' process receives its execution token from the Scheduler?
2) Or a 'signal' handler execution takes place in whatever process 'context' that happens to be executing when the 'signal' was sent? (The same as a hardware ISR).

Signals are sent to specific processes, not to the process that happens to be executing when the signal is sent.

In the interactive shell, you can send signals using the kill command. It requires a process ID. In a program, you would use the kill(2) system call, which also takes a process ID as a parameter. The process with this ID will receive the signal.

Signals can also be sent to groups of processes; see the man pages for the kill(1) command and the kill(2) system call.

Signals can also be sent from the kernel without the kill system call. For example, if a process accesses protected memory, or executes a non-existing or privileged instruction, or divides by zero.

Thanx for your replies.
From your answers and some links provided above, in summary:

1) A signal handler is executed when the target process gets its execution token form the Scheduler. This can be at the very beginning if there are corresponding 'signal pending' bits enabled.

2) Once the target process is running, its 'signal pending' bits are evaluated whenever a system call is performed. This way the signal handler can be executed in between the execution slice of the target process.