// memory/vm related tapset
// Copyright (C) 2005, 2006 IBM Corp.
// Copyright (C) 2006 Intel Corporation.
//
// 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 memory-related events. 
// </tapsetdescription>
%{
#include <linux/mm.h>
%}

global VM_FAULT_OOM=0, VM_FAULT_SIGBUS=1, VM_FAULT_MINOR=2, VM_FAULT_MAJOR=3
global VM_FAULT_NOPAGE=4, VM_FAULT_LOCKED=5, VM_FAULT_ERROR=6
global FAULT_FLAG_WRITE=1

/**
 * sfunction vm_fault_contains - Test return value for page fault reason
 * @value: The fault_type returned by vm.page_fault.return
 * @test: The type of fault to test for (VM_FAULT_OOM or similar)
 */
function vm_fault_contains:long (value:long, test:long)
%{
	int res;
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,23)
	switch (THIS->test){
	case 0: res = THIS->value == VM_FAULT_OOM; break;
	case 1: res = THIS->value == VM_FAULT_SIGBUS; break;
	case 2: res = THIS->value == VM_FAULT_MINOR; break;
	case 3: res = THIS->value == VM_FAULT_MAJOR; break;
	default:
		res = 0; break;
	}
#else
	switch (THIS->test){
	case 0: res = THIS->value & VM_FAULT_OOM; break;
	case 1: res = THIS->value & VM_FAULT_SIGBUS; break;
	case 2: /* VM_FAULT_MINOR infered by that flags off */
		res = !((VM_FAULT_OOM | VM_FAULT_SIGBUS | VM_FAULT_MAJOR) & 
				 THIS->value);
		 break;
	case 3: res = THIS->value & VM_FAULT_MAJOR; break;
	case 4: res = THIS->value & VM_FAULT_NOPAGE; break;
	case 5: res = THIS->value & VM_FAULT_LOCKED; break;
	case 6: res = THIS->value & VM_FAULT_ERROR; break;
	default:
		res = 0;
	}
#endif
	THIS->__retvalue = (res != 0);
	return;
%}

/**
 * probe vm.pagefault - Records that a page fault occurred.
 * @address: The address of the faulting memory access; i.e. the address that caused the page fault.
 * @write_access: Indicates whether this was a write or read access; <command>1</command> indicates a write, 
 * while <command>0</command> indicates a read.
 *
 * Context: The process which triggered the fault
 */
probe vm.pagefault = kernel.function("__handle_mm_fault@mm/memory.c") ?,
                     kernel.function("handle_mm_fault@mm/memory.c") ?
{
%( kernel_v >= "2.6.31" %?
	write_access = $flags & FAULT_FLAG_WRITE
%:
	write_access = $write_access
%)
	address =  $address
}

/**
 * probe vm.pagefault.return - Indicates what type of fault occurred.
 * @fault_type: Returns either 
 * <command>0</command> (VM_FAULT_OOM) for out of memory faults, 
 * <command>2</command> (VM_FAULT_MINOR) for minor faults, <command>3</command> (VM_FAULT_MAJOR) for 
 * major faults, or <command>1</command> (VM_FAULT_SIGBUS) if the fault was neither OOM, minor fault, 
 * nor major fault.
 */
probe vm.pagefault.return = kernel.function("__handle_mm_fault@mm/memory.c").return ?,
                            kernel.function("handle_mm_fault@mm/memory.c").return ?
{
	fault_type = $return
}

/**
 * sfunction addr_to_node - Returns which node a given address belongs to within a NUMA system.
 * @addr: The address of the faulting memory access.
 *
 */
function addr_to_node:long(addr:long) %{ /* pure */ 
	int nid;
	int pfn = __pa(THIS->addr) >> PAGE_SHIFT;
	for_each_online_node(nid)
		if ( NODE_DATA(nid)->node_start_pfn <= pfn &&
			pfn < (NODE_DATA(nid)->node_start_pfn +
			NODE_DATA(nid)->node_spanned_pages) )
		{
			THIS->__retvalue = nid;
			break;
		}
%}

// Return whether a page to be copied is a zero page.
function _IS_ZERO_PAGE:long(from:long, vaddr:long) %{ /* pure */
    THIS->__retvalue = (THIS->from == (long) ZERO_PAGE(THIS->vaddr));
%}


/**
 * probe vm.write_shared - Attempts at writing to a shared page.
 * @address: The address of the shared write.
 *
 * Context:
 *  The context is the process attempting the write.
 *
 *  Fires when a process attempts to write to a shared page. 
 *  If a copy is necessary, this will be followed by a 
 *  <command>vm.write_shared_copy</command>.
 */
probe vm.write_shared = kernel.function("do_wp_page") {
    address = $address
}

/**
 * probe vm.write_shared_copy - Page copy for shared page write.
 * @address: The address of the shared write.
 * @zero: Boolean indicating whether it is a zero page
 *         (can do a clear instead of a copy).
 *
 * Context:
 *  The process attempting the write.
 *
 *  Fires when a write to a shared page requires a page copy.  This is
 *  always preceded by a <command>vm.shared_write</command>.
 */
probe vm.write_shared_copy = kernel.function("copy_cow_page")? {
    address = $address
    zero = _IS_ZERO_PAGE($from, address);
}


/**
 * probe vm.mmap - Fires when an <command>mmap</command> is requested.
 * @address: The requested address
 * @length: The length of the memory segment 
 *
 * Context:
 *  The process calling <command>mmap</command>.
 */
probe vm.mmap = kernel.function("do_mmap"), kernel.function("do_mmap2")? {
    address = $addr
    length = $len
}


/**
 * probe vm.munmap - Fires when an <command>munmap</command> is requested.
 * @address: The requested address
 * @length: The length of the memory segment 
 *
 * Context:
 *  The process calling <command>munmap</command>.
 */
probe vm.munmap = kernel.function("do_munmap") {
    address = $start
    length = $len
}

/**
 * probe vm.brk - Fires when a <command>brk</command> is requested (i.e. the heap will be resized).
 * @address: The requested address
 * @length: The length of the memory segment 
 *
 * Context:
 *  The process calling <command>brk</command>.
 */
probe vm.brk = kernel.function("do_brk") {
    address = $addr
    length = $len
}

/**
 * probe vm.oom_kill - Fires when a thread is selected for termination by the OOM killer.
 * @task: The task being killed
 *
 * Context:
 *  The process that tried to consume excessive memory, and thus
 *  triggered the OOM. <remark>(is this correct?)</remark>
 */
probe vm.oom_kill = kernel.function("__oom_kill_task") {
    task = $p
}