Re: About limited of process memory!

On Nov 27, 6:22 pm, Joe Pfeiffer <pfeif...@xxxxxxxxxxx> wrote:

In general (threads and various odd uses of the clone() call to create
new processes are special cases), every process in Linux has its own
address space.

There are no exceptions, every process in Linux has its own address
space. The 'clone' call does not always create a new process. It can
also create a new thread in an existing process.

Do not confuse the special legacy usage of "process" in the specific
context of Linux kernel development with the definition of a "process"
used by applications and every other context. A "process" is the set
of contexts of execution that share an address space.

Because Linux once had a one-to-one correspondence between threads and
processes, the term 'process' is sometimes used in kernel source code
and when talking about the kernel to refer to something that is
actually neither a process nor a thread, but is in fact a KSE (kernel
scheduling entity). This legacy usage in the kernel internals should
*not* be allowed to spill over into application context, where
'process' has a precise definition.

DS
.

Relevant Pages