path: root/drivers/staging/android/lowmemorykiller.c

/* drivers/misc/lowmemorykiller.c
 *
 * The lowmemorykiller driver lets user-space specify a set of memory thresholds
 * where processes with a range of oom_score_adj values will get killed. Specify
 * the minimum oom_score_adj values in
 * /sys/module/lowmemorykiller/parameters/adj and the number of free pages in
 * /sys/module/lowmemorykiller/parameters/minfree. Both files take a comma
 * separated list of numbers in ascending order.
 *
 * For example, write "0,8" to /sys/module/lowmemorykiller/parameters/adj and
 * "1024,4096" to /sys/module/lowmemorykiller/parameters/minfree to kill
 * processes with a oom_score_adj value of 8 or higher when the free memory
 * drops below 4096 pages and kill processes with a oom_score_adj value of 0 or
 * higher when the free memory drops below 1024 pages.
 *
 * The driver considers memory used for caches to be free, but if a large
 * percentage of the cached memory is locked this can be very inaccurate
 * and processes may not get killed until the normal oom killer is triggered.
 *
 * Copyright (C) 2007-2008 Google, Inc.
 *
 * This software is licensed under the terms of the GNU General Public
 * License version 2, as published by the Free Software Foundation, and
 * may be copied, distributed, and modified under those terms.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/oom.h>
#include <linux/sched.h>
#include <linux/swap.h>
#include <linux/rcupdate.h>
#include <linux/notifier.h>
#define ENHANCED_LMK_ROUTINE

#ifdef ENHANCED_LMK_ROUTINE
#define LOWMEM_DEATHPENDING_DEPTH 3
#endif

static uint32_t lowmem_debug_level = 1;
static int lowmem_adj[6] = {
	0,
	1,
	6,
	12,
};
static int lowmem_adj_size = 4;
static int lowmem_minfree[6] = {
	3 * 512,	/* 6MB */
	2 * 1024,	/* 8MB */
	4 * 1024,	/* 16MB */
	16 * 1024,	/* 64MB */
};
static int lowmem_minfree_size = 4;

#ifdef ENHANCED_LMK_ROUTINE
static struct task_struct *lowmem_deathpending[LOWMEM_DEATHPENDING_DEPTH] = {NULL,};
#else
static struct task_struct *lowmem_deathpending;
#endif
static unsigned long lowmem_deathpending_timeout;

#define lowmem_print(level, x...)			\
	do {						\
		if (lowmem_debug_level >= (level))	\
			pr_info(x);			\
	} while (0)

static int
task_notify_func(struct notifier_block *self, unsigned long val, void *data);

static struct notifier_block task_nb = {
	.notifier_call	= task_notify_func,
};

static int
task_notify_func(struct notifier_block *self, unsigned long val, void *data)
{
	struct task_struct *task = data;

#ifdef ENHANCED_LMK_ROUTINE
	int i = 0;
	for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++)
		if (task == lowmem_deathpending[i]) {
			lowmem_deathpending[i] = NULL;
		break;
	}
#else
	if (task == lowmem_deathpending)
		lowmem_deathpending = NULL;
#endif
	return NOTIFY_OK;
}

#ifdef CONFIG_ANDROID_LMK_ADJ_RBTREE
static struct task_struct *pick_next_from_adj_tree(struct task_struct *task);
static struct task_struct *pick_first_task(void);
static struct task_struct *pick_last_task(void);
#endif

static int lowmem_shrink(struct shrinker *s, struct shrink_control *sc)
{
	struct task_struct *tsk;
#ifdef ENHANCED_LMK_ROUTINE
	struct task_struct *selected[LOWMEM_DEATHPENDING_DEPTH] = {NULL,};
#else
	struct task_struct *selected = NULL;
#endif
	int rem = 0;
	int tasksize;
	int i;
	int min_score_adj = OOM_SCORE_ADJ_MAX + 1;
	int minfree = 0;
#ifdef ENHANCED_LMK_ROUTINE
	int selected_tasksize[LOWMEM_DEATHPENDING_DEPTH] = {0,};
	int selected_oom_adj[LOWMEM_DEATHPENDING_DEPTH] = {OOM_ADJUST_MAX,};
	int all_selected_oom = 0;
	int max_selected_oom_idx = 0;
#else
	int selected_tasksize = 0;
	int selected_oom_score_adj;
#endif
	int array_size = ARRAY_SIZE(lowmem_adj);
#ifndef CONFIG_DMA_CMA
	int other_free = global_page_state(NR_FREE_PAGES) - totalreserve_pages;
#else
	int other_free = global_page_state(NR_FREE_PAGES) -
					global_page_state(NR_FREE_CMA_PAGES);
#endif
	int other_file = global_page_state(NR_FILE_PAGES) -
						global_page_state(NR_SHMEM);

	/*
	 * If we already have a death outstanding, then
	 * bail out right away; indicating to vmscan
	 * that we have nothing further to offer on
	 * this pass.
	 *
	 */
#ifdef ENHANCED_LMK_ROUTINE
	for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++) {
		if (lowmem_deathpending[i] &&
			time_before_eq(jiffies, lowmem_deathpending_timeout))
			return 0;
	}
#else
	if (lowmem_deathpending &&
	    time_before_eq(jiffies, lowmem_deathpending_timeout))
		return 0;
#endif

	if (lowmem_adj_size < array_size)
		array_size = lowmem_adj_size;
	if (lowmem_minfree_size < array_size)
		array_size = lowmem_minfree_size;
	for (i = 0; i < array_size; i++) {
		minfree = lowmem_minfree[i];
		if (other_free < minfree && other_file < minfree) {
			min_score_adj = lowmem_adj[i];
			break;
		}
	}
	if (sc->nr_to_scan > 0)
		lowmem_print(3, "lowmem_shrink %lu, %x, ofree %d %d, ma %d\n",
				sc->nr_to_scan, sc->gfp_mask, other_free,
				other_file, min_score_adj);
	rem = global_page_state(NR_ACTIVE_ANON) +
		global_page_state(NR_ACTIVE_FILE) +
		global_page_state(NR_INACTIVE_ANON) +
		global_page_state(NR_INACTIVE_FILE);
	if (sc->nr_to_scan <= 0 || min_score_adj == OOM_SCORE_ADJ_MAX + 1) {
		lowmem_print(5, "lowmem_shrink %lu, %x, return %d\n",
			     sc->nr_to_scan, sc->gfp_mask, rem);
		return rem;
	}

#ifdef ENHANCED_LMK_ROUTINE
	for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++)
		selected_oom_adj[i] = min_adj;
#else
	selected_oom_score_adj = min_score_adj;
#endif
	rcu_read_lock();

#ifdef CONFIG_ANDROID_LMK_ADJ_RBTREE
	for (tsk = pick_first_task();
		tsk != pick_last_task() && tsk != NULL;
		tsk = pick_next_from_adj_tree(tsk)) {
#else
	for_each_process(tsk) {
#endif
		struct task_struct *p;
		int oom_score_adj;
#ifdef ENHANCED_LMK_ROUTINE
		int is_exist_oom_task = 0;
#endif
		if (tsk->flags & PF_KTHREAD)
			continue;

		p = find_lock_task_mm(tsk);
		if (!p)
			continue;

		oom_score_adj = p->signal->oom_score_adj;
		if (oom_score_adj < min_score_adj) {
			task_unlock(p);
#ifdef CONFIG_ANDROID_LMK_ADJ_RBTREE
			break;
#else
			continue;
#endif
		}
		if (fatal_signal_pending(p)) {
			lowmem_print(2, "skip slow dying process %d\n", p->pid);
			task_unlock(p);
			continue;
		}
		tasksize = get_mm_rss(p->mm);
		task_unlock(p);
		if (tasksize <= 0)
			continue;

#ifdef ENHANCED_LMK_ROUTINE
		if (all_selected_oom < LOWMEM_DEATHPENDING_DEPTH) {
			for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++) {
				if (!selected[i]) {
					is_exist_oom_task = 1;
					max_selected_oom_idx = i;
					break;
				}
			}
		} else if (selected_oom_adj[max_selected_oom_idx] < oom_adj ||
			(selected_oom_adj[max_selected_oom_idx] == oom_adj &&
			selected_tasksize[max_selected_oom_idx] < tasksize)) {
			is_exist_oom_task = 1;
		}

		if (is_exist_oom_task) {
			selected[max_selected_oom_idx] = p;
			selected_tasksize[max_selected_oom_idx] = tasksize;
			selected_oom_adj[max_selected_oom_idx] = oom_adj;

			if (all_selected_oom < LOWMEM_DEATHPENDING_DEPTH)
				all_selected_oom++;

			if (all_selected_oom == LOWMEM_DEATHPENDING_DEPTH) {
				for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++) {
					if (selected_oom_adj[i] < selected_oom_adj[max_selected_oom_idx])
						max_selected_oom_idx = i;
					else if (selected_oom_adj[i] == selected_oom_adj[max_selected_oom_idx] &&
						selected_tasksize[i] < selected_tasksize[max_selected_oom_idx])
						max_selected_oom_idx = i;
				}
			}

			lowmem_print(2, "select %d (%s), adj %d, size %d, to kill\n",
				p->pid, p->comm, oom_adj, tasksize);
		}
#else
		if (selected) {
			if (oom_score_adj < selected_oom_score_adj)
#ifdef CONFIG_ANDROID_LMK_ADJ_RBTREE
				break;
#else
				continue;
#endif
			if (oom_score_adj == selected_oom_score_adj &&
			    tasksize <= selected_tasksize)
				continue;
		}
		selected = p;
		selected_tasksize = tasksize;
		selected_oom_score_adj = oom_score_adj;
		lowmem_print(2, "select '%s' (%d), adj %d, size %d, to kill\n",
			     p->comm, p->pid, oom_score_adj, tasksize);
#endif
	}
#ifdef ENHANCED_LMK_ROUTINE
	for (i = 0; i < LOWMEM_DEATHPENDING_DEPTH; i++) {
		if (selected[i]) {
			lowmem_print(1, "send sigkill to %d (%s), adj %d, size %d\n",
				selected[i]->pid, selected[i]->comm,
				selected_oom_adj[i], selected_tasksize[i]);
			lowmem_deathpending[i] = selected[i];
			lowmem_deathpending_timeout = jiffies + HZ;
			force_sig(SIGKILL, selected[i]);
			rem -= selected_tasksize[i];
		}
	}
#else
	if (selected) {
		lowmem_print(1, "Killing '%s' (%d), adj %d,\n" \
				"   to free %ldkB on behalf of '%s' (%d) because\n" \
				"   cache %ldkB is below limit %ldkB for oom_score_adj %d\n" \
				"   Free memory is %ldkB above reserved\n",
			     selected->comm, selected->pid,
			     selected_oom_score_adj,
			     selected_tasksize * (long)(PAGE_SIZE / 1024),
			     current->comm, current->pid,
			     other_file * (long)(PAGE_SIZE / 1024),
			     minfree * (long)(PAGE_SIZE / 1024),
			     min_score_adj,
			     other_free * (long)(PAGE_SIZE / 1024));
		lowmem_deathpending = selected;
		lowmem_deathpending_timeout = jiffies + HZ;
		send_sig(SIGKILL, selected, 0);
		rem -= selected_tasksize;
	}
#endif
	lowmem_print(4, "lowmem_shrink %lu, %x, return %d\n",
		     sc->nr_to_scan, sc->gfp_mask, rem);
	rcu_read_unlock();
	return rem;
}

static struct shrinker lowmem_shrinker = {
	.shrink = lowmem_shrink,
	.seeks = DEFAULT_SEEKS * 16
};

static int __init lowmem_init(void)
{
	task_free_register(&task_nb);
	register_shrinker(&lowmem_shrinker);
	return 0;
}

static void __exit lowmem_exit(void)
{
	unregister_shrinker(&lowmem_shrinker);
	task_free_unregister(&task_nb);
}

#ifdef CONFIG_ANDROID_LOW_MEMORY_KILLER_AUTODETECT_OOM_ADJ_VALUES
static int lowmem_oom_adj_to_oom_score_adj(int oom_adj)
{
	if (oom_adj == OOM_ADJUST_MAX)
		return OOM_SCORE_ADJ_MAX;
	else
		return (oom_adj * OOM_SCORE_ADJ_MAX) / -OOM_DISABLE;
}

static void lowmem_autodetect_oom_adj_values(void)
{
	int i;
	int oom_adj;
	int oom_score_adj;
	int array_size = ARRAY_SIZE(lowmem_adj);

	if (lowmem_adj_size < array_size)
		array_size = lowmem_adj_size;

	if (array_size <= 0)
		return;

	oom_adj = lowmem_adj[array_size - 1];
	if (oom_adj > OOM_ADJUST_MAX)
		return;

	oom_score_adj = lowmem_oom_adj_to_oom_score_adj(oom_adj);
	if (oom_score_adj <= OOM_ADJUST_MAX)
		return;

	lowmem_print(1, "lowmem_shrink: convert oom_adj to oom_score_adj:\n");
	for (i = 0; i < array_size; i++) {
		oom_adj = lowmem_adj[i];
		oom_score_adj = lowmem_oom_adj_to_oom_score_adj(oom_adj);
		lowmem_adj[i] = oom_score_adj;
		lowmem_print(1, "oom_adj %d => oom_score_adj %d\n",
			     oom_adj, oom_score_adj);
	}
}

static int lowmem_adj_array_set(const char *val, const struct kernel_param *kp)
{
	int ret;

	ret = param_array_ops.set(val, kp);

	/* HACK: Autodetect oom_adj values in lowmem_adj array */
	lowmem_autodetect_oom_adj_values();

	return ret;
}

static int lowmem_adj_array_get(char *buffer, const struct kernel_param *kp)
{
	return param_array_ops.get(buffer, kp);
}

static void lowmem_adj_array_free(void *arg)
{
	param_array_ops.free(arg);
}

static struct kernel_param_ops lowmem_adj_array_ops = {
	.set = lowmem_adj_array_set,
	.get = lowmem_adj_array_get,
	.free = lowmem_adj_array_free,
};

static const struct kparam_array __param_arr_adj = {
	.max = ARRAY_SIZE(lowmem_adj),
	.num = &lowmem_adj_size,
	.ops = &param_ops_int,
	.elemsize = sizeof(lowmem_adj[0]),
	.elem = lowmem_adj,
};
#endif

#ifdef CONFIG_ANDROID_LMK_ADJ_RBTREE
DEFINE_SPINLOCK(lmk_lock);
struct rb_root tasks_scoreadj = RB_ROOT;
/*
 * Makesure to invoke the function with holding sighand->siglock
 */
void add_2_adj_tree(struct task_struct *task)
{
	struct rb_node **link;
	struct rb_node *parent = NULL;
	struct signal_struct *sig_entry;
	s64 key = task->signal->oom_score_adj;

	/*
	 * Find the right place in the rbtree:
	 */
	spin_lock(&lmk_lock);
	link =  &tasks_scoreadj.rb_node;
	while (*link) {
		parent = *link;
		sig_entry = rb_entry(parent, struct signal_struct, adj_node);

		if (key < sig_entry->oom_score_adj)
			link = &parent->rb_right;
		else
			link = &parent->rb_left;
	}

	rb_link_node(&task->signal->adj_node, parent, link);
	rb_insert_color(&task->signal->adj_node, &tasks_scoreadj);
	spin_unlock(&lmk_lock);
}

/*
 * Makesure to invoke the function with holding sighand->siglock
 */
void delete_from_adj_tree(struct task_struct *task)
{
	spin_lock(&lmk_lock);
	if (!RB_EMPTY_NODE(&task->signal->adj_node)) {
		rb_erase(&task->signal->adj_node, &tasks_scoreadj);
		RB_CLEAR_NODE(&task->signal->adj_node);
	}
	spin_unlock(&lmk_lock);
}


static struct task_struct *pick_next_from_adj_tree(struct task_struct *task)
{
	struct rb_node *next;
	struct signal_struct *next_tsk_sig;

	spin_lock(&lmk_lock);
	next = rb_next(&task->signal->adj_node);
	spin_unlock(&lmk_lock);

	if (!next)
		return NULL;

	next_tsk_sig = rb_entry(next, struct signal_struct, adj_node);
	return next_tsk_sig->curr_target->group_leader;
}

static struct task_struct *pick_first_task(void)
{
	struct rb_node *left;
	struct signal_struct *first_tsk_sig;

	spin_lock(&lmk_lock);
	left = rb_first(&tasks_scoreadj);
	spin_unlock(&lmk_lock);

	if (!left)
		return NULL;

	first_tsk_sig = rb_entry(left, struct signal_struct, adj_node);
	return first_tsk_sig->curr_target->group_leader;
}

static struct task_struct *pick_last_task(void)
{
	struct rb_node *right;
	struct signal_struct *last_tsk_sig;

	spin_lock(&lmk_lock);
	right = rb_last(&tasks_scoreadj);
	spin_unlock(&lmk_lock);

	if (!right)
		return NULL;

	last_tsk_sig = rb_entry(right, struct signal_struct, adj_node);
	return last_tsk_sig->curr_target->group_leader;
}
#endif

module_param_named(cost, lowmem_shrinker.seeks, int, S_IRUGO | S_IWUSR);
#ifdef CONFIG_ANDROID_LOW_MEMORY_KILLER_AUTODETECT_OOM_ADJ_VALUES
__module_param_call(MODULE_PARAM_PREFIX, adj,
		    &lowmem_adj_array_ops,
		    .arr = &__param_arr_adj,
		    S_IRUGO | S_IWUSR, 0664);
__MODULE_PARM_TYPE(adj, "array of int");
#else
module_param_array_named(adj, lowmem_adj, int, &lowmem_adj_size,
			 S_IRUGO | S_IWUSR);
#endif
module_param_array_named(minfree, lowmem_minfree, uint, &lowmem_minfree_size,
			 S_IRUGO | S_IWUSR);
module_param_named(debug_level, lowmem_debug_level, uint, S_IRUGO | S_IWUSR);

module_init(lowmem_init);
module_exit(lowmem_exit);

MODULE_LICENSE("GPL");