// kernel process tapset
// Copyright (C) 2006 Intel Corporation.
// Copyright (C) 2014-2017 Red Hat Inc.
//
// This file is part of systemtap, and is free software.  You can
// redistribute it and/or modify it under the terms of the GNU General
// Public License (GPL); either version 2, or (at your option) any
// later version.
// <tapsetdescription>
//  This family of probe points is used to probe process-related activities.
// </tapsetdescription>

@__private30 function _IS_ERR:long(ptr:long) %{ /* pure */
    STAP_RETVALUE = IS_ERR((const void *)(uintptr_t)STAP_ARG_ptr);
%}


/**
 * probe kprocess.create - Fires whenever a new process or thread is successfully created
 * @new_pid: The PID of the newly created process
 * @new_tid: The TID of the newly created task
 *
 * Context:
 *  Parent of the created process.
 *
 *  Fires whenever a new process is successfully created, either as a result of
 *  fork (or one of its syscall variants), or a new kernel thread.
 */
probe kprocess.create = kernel.function("copy_process").return {
    task = $return
    if (_IS_ERR(task)) next
    new_pid = task_pid(task)
    new_tid = task_tid(task)
}


/**
 * probe kprocess.start - Starting new process
 *
 * Context:
 * Newly created process.
 *
 * Fires immediately before a new process begins execution.
 *
 */
probe kprocess.start = kernel.function("schedule_tail") { }


/**
 * probe kprocess.exec - Attempt to exec to a new program
 *
 * @filename: The path to the new executable
 * @name: Name of the system call ("execve") (SystemTap v2.5+)
 * @args: The arguments to pass to the new executable, including
 * the 0th arg (SystemTap v2.5+)
 * @argstr: A string containing the filename followed by the
 * arguments to pass, excluding 0th arg (SystemTap v2.5+)
 *
 * Context:
 *  The caller of exec.
 *
 *  Fires whenever a process attempts to exec to a new program. Aliased
 *  to the syscall.execve probe in SystemTap v2.5+.
 */
%(systemtap_v <= "2.4" %?
probe kprocess.exec = 
    kernel.function("do_execve"),
    kernel.function("compat_do_execve") ?
{
    filename = kernel_string($filename)
}
%:
probe kprocess.exec = syscall.execve
{
   /*
   name = "execve"
   filename = user_string_quoted(@choose_defined($filename, $name))
   # kernel 3.0 changed the pointer's name to __argv
   __argv = @choose_defined($__argv, $argv)
   args = __get_argv(__argv, 0)
   argstr = sprintf("%s %s", filename, __get_argv(__argv, 1))
   */
}
%)


/**
 * probe kprocess.exec_complete - Return from exec to a new program
 * @errno: The error number resulting from the exec
 * @success: A boolean indicating whether the exec was successful
 * @name: Name of the system call ("execve") (SystemTap v2.5+)
 * @retstr: A string representation of errno (SystemTap v2.5+)
 *
 * Context:
 *  On success, the context of the new executable.
 *  On failure, remains in the context of the caller.
 *
 *  Fires at the completion of an exec call. Aliased to the
 *  syscall.execve.return probe in SystemTap v2.5+.
 */
%(systemtap_v <= "2.4" %?
probe kprocess.exec_complete =
    kernel.function("do_execve").return,
    kernel.function("compat_do_execve").return ?
%:
probe kprocess.exec_complete = syscall.execve.return
%)
{
    errno = $return
    success = (errno >= 0)
    /*
    name = "execve"
    retstr = return_str(1, $return)
    */
}


/**
 * probe kprocess.exit - Exit from process
 * @code: The exit code of the process
 *
 * Context:
 *  The process which is terminating.
 *
 *  Fires when a process terminates.  This will always be followed by a
 *  kprocess.release, though the latter may be delayed if the process waits in a
 *  zombie state.
 */
probe kprocess.exit = kernel.function("do_exit") {
    code = $code
}


/**
 * probe kprocess.release - Process released
 * @task: A task handle to the process being released
 * @released_pid: PID of the process being released
 * @released_tid: TID of the task being released
 * @pid: Same as @released_pid for compatibility (deprecated)
 *
 * Context:
 *  The context of the parent, if it wanted notification of this process'
 *  termination, else the context of the process itself.
 *
 *  Fires when a process is released from the kernel.  This always follows a
 *  kprocess.exit, though it may be delayed somewhat if the process waits in a
 *  zombie state.
 */
probe kprocess.release = kernel.function("release_task") {
    task = $p
%(systemtap_v <= "1.7" %?
    pid = task_pid($p)
%)
    released_pid = task_pid($p)
    released_tid = task_tid($p)
}