Hi Jakko > Well you gave me the general code of the proc filesystem; > alas, I want to know where is the code that _uses_ the proc code > in order to export proc directories of the process ids of the system > (i.w. /proc/1/, /proc/2/, ...) try to check proc_pid_readdir() in fs/proc/base.c. It will give you the idea how thos /proc/<pid>s are created. Short story, I got impression they are created on the fly (i.e when you read them), not when a process is created. Same thing with destruction. The trick is faking file read offset as pid selector, and from there, task's properties are fetched and showed as subdirectory under /proc/<pid>/ . Sounds like expensive operation? Yes it is, but IIRC dcache and icache will make the subsequent operation faster. However, the contents are volatile since they represent the current task's status. Realizing these, you will find out why sometimes using "top" with small update frequency is not preferred :) It traverses /proc/<pid>/ again and again, reopening file descriptors instead of creating and hold them as long as it can regards, Mulyadi -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
