[linux-2.6-block.git] / arch / powerpc / platforms / powernv / memtrace.c

/*
 * Copyright (C) IBM Corporation, 2014, 2017
 * Anton Blanchard, Rashmica Gupta.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 */

#define pr_fmt(fmt) "memtrace: " fmt

#include <linux/bitops.h>
#include <linux/string.h>
#include <linux/memblock.h>
#include <linux/init.h>
#include <linux/moduleparam.h>
#include <linux/fs.h>
#include <linux/debugfs.h>
#include <linux/slab.h>
#include <linux/memory.h>
#include <linux/memory_hotplug.h>
#include <asm/machdep.h>
#include <asm/debugfs.h>

/* This enables us to keep track of the memory removed from each node. */
struct memtrace_entry {
	void *mem;
	u64 start;
	u64 size;
	u32 nid;
	struct dentry *dir;
	char name[16];
};

static u64 memtrace_size;

static struct memtrace_entry *memtrace_array;
static unsigned int memtrace_array_nr;


static ssize_t memtrace_read(struct file *filp, char __user *ubuf,
			     size_t count, loff_t *ppos)
{
	struct memtrace_entry *ent = filp->private_data;

	return simple_read_from_buffer(ubuf, count, ppos, ent->mem, ent->size);
}

static bool valid_memtrace_range(struct memtrace_entry *dev,
				 unsigned long start, unsigned long size)
{
	if ((start >= dev->start) &&
	    ((start + size) <= (dev->start + dev->size)))
		return true;

	return false;
}

static int memtrace_mmap(struct file *filp, struct vm_area_struct *vma)
{
	unsigned long size = vma->vm_end - vma->vm_start;
	struct memtrace_entry *dev = filp->private_data;

	if (!valid_memtrace_range(dev, vma->vm_pgoff << PAGE_SHIFT, size))
		return -EINVAL;

	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);

	if (remap_pfn_range(vma, vma->vm_start,
			    vma->vm_pgoff + (dev->start >> PAGE_SHIFT),
			    size, vma->vm_page_prot))
		return -EAGAIN;

	return 0;
}

static const struct file_operations memtrace_fops = {
	.llseek = default_llseek,
	.read	= memtrace_read,
	.mmap	= memtrace_mmap,
	.open	= simple_open,
};

static int check_memblock_online(struct memory_block *mem, void *arg)
{
	if (mem->state != MEM_ONLINE)
		return -1;

	return 0;
}

static int change_memblock_state(struct memory_block *mem, void *arg)
{
	unsigned long state = (unsigned long)arg;

	mem->state = state;

	return 0;
}

static bool memtrace_offline_pages(u32 nid, u64 start_pfn, u64 nr_pages)
{
	u64 end_pfn = start_pfn + nr_pages - 1;

	if (walk_memory_range(start_pfn, end_pfn, NULL,
	    check_memblock_online))
		return false;

	walk_memory_range(start_pfn, end_pfn, (void *)MEM_GOING_OFFLINE,
			  change_memblock_state);

	if (offline_pages(start_pfn, nr_pages)) {
		walk_memory_range(start_pfn, end_pfn, (void *)MEM_ONLINE,
				  change_memblock_state);
		return false;
	}

	walk_memory_range(start_pfn, end_pfn, (void *)MEM_OFFLINE,
			  change_memblock_state);

	lock_device_hotplug();
	remove_memory(nid, start_pfn << PAGE_SHIFT, nr_pages << PAGE_SHIFT);
	unlock_device_hotplug();

	return true;
}

static u64 memtrace_alloc_node(u32 nid, u64 size)
{
	u64 start_pfn, end_pfn, nr_pages;
	u64 base_pfn;

	if (!NODE_DATA(nid) || !node_spanned_pages(nid))
		return 0;

	start_pfn = node_start_pfn(nid);
	end_pfn = node_end_pfn(nid);
	nr_pages = size >> PAGE_SHIFT;

	/* Trace memory needs to be aligned to the size */
	end_pfn = round_down(end_pfn - nr_pages, nr_pages);

	for (base_pfn = end_pfn; base_pfn > start_pfn; base_pfn -= nr_pages) {
		if (memtrace_offline_pages(nid, base_pfn, nr_pages) == true)
			return base_pfn << PAGE_SHIFT;
	}

	return 0;
}

static int memtrace_init_regions_runtime(u64 size)
{
	u32 nid;
	u64 m;

	memtrace_array = kcalloc(num_online_nodes(),
				sizeof(struct memtrace_entry), GFP_KERNEL);
	if (!memtrace_array) {
		pr_err("Failed to allocate memtrace_array\n");
		return -EINVAL;
	}

	for_each_online_node(nid) {
		m = memtrace_alloc_node(nid, size);

		/*
		 * A node might not have any local memory, so warn but
		 * continue on.
		 */
		if (!m) {
			pr_err("Failed to allocate trace memory on node %d\n", nid);
			continue;
		}

		pr_info("Allocated trace memory on node %d at 0x%016llx\n", nid, m);

		memtrace_array[memtrace_array_nr].start = m;
		memtrace_array[memtrace_array_nr].size = size;
		memtrace_array[memtrace_array_nr].nid = nid;
		memtrace_array_nr++;
	}

	return 0;
}

static struct dentry *memtrace_debugfs_dir;

static int memtrace_init_debugfs(void)
{
	int ret = 0;
	int i;

	for (i = 0; i < memtrace_array_nr; i++) {
		struct dentry *dir;
		struct memtrace_entry *ent = &memtrace_array[i];

		ent->mem = ioremap(ent->start, ent->size);
		/* Warn but continue on */
		if (!ent->mem) {
			pr_err("Failed to map trace memory at 0x%llx\n",
				 ent->start);
			ret = -1;
			continue;
		}

		snprintf(ent->name, 16, "%08x", ent->nid);
		dir = debugfs_create_dir(ent->name, memtrace_debugfs_dir);
		if (!dir)
			return -1;

		ent->dir = dir;
		debugfs_create_file("trace", 0400, dir, ent, &memtrace_fops);
		debugfs_create_x64("start", 0400, dir, &ent->start);
		debugfs_create_x64("size", 0400, dir, &ent->size);
	}

	return ret;
}

static int memtrace_enable_set(void *data, u64 val)
{
	if (memtrace_size)
		return -EINVAL;

	if (!val)
		return -EINVAL;

	/* Make sure size is aligned to a memory block */
	if (val & (memory_block_size_bytes() - 1))
		return -EINVAL;

	if (memtrace_init_regions_runtime(val))
		return -EINVAL;

	if (memtrace_init_debugfs())
		return -EINVAL;

	memtrace_size = val;

	return 0;
}

static int memtrace_enable_get(void *data, u64 *val)
{
	*val = memtrace_size;
	return 0;
}

DEFINE_SIMPLE_ATTRIBUTE(memtrace_init_fops, memtrace_enable_get,
					memtrace_enable_set, "0x%016llx\n");

static int memtrace_init(void)
{
	memtrace_debugfs_dir = debugfs_create_dir("memtrace",
						  powerpc_debugfs_root);
	if (!memtrace_debugfs_dir)
		return -1;

	debugfs_create_file("enable", 0600, memtrace_debugfs_dir,
			    NULL, &memtrace_init_fops);

	return 0;
}
machine_device_initcall(powernv, memtrace_init);
Commit	Line	Data
9d5171a8 RG	1	/*
	2	* Copyright (C) IBM Corporation, 2014, 2017
	3	* Anton Blanchard, Rashmica Gupta.
	4	*
	5	* This program is free software; you can redistribute it and/or modify
	6	* it under the terms of the GNU General Public License as published by
	7	* the Free Software Foundation; either version 2 of the License, or
	8	* (at your option) any later version.
	9	*/
	10
	11	#define pr_fmt(fmt) "memtrace: " fmt
	12
	13	#include <linux/bitops.h>
	14	#include <linux/string.h>
	15	#include <linux/memblock.h>
	16	#include <linux/init.h>
	17	#include <linux/moduleparam.h>
	18	#include <linux/fs.h>
	19	#include <linux/debugfs.h>
	20	#include <linux/slab.h>
	21	#include <linux/memory.h>
	22	#include <linux/memory_hotplug.h>
	23	#include <asm/machdep.h>
	24	#include <asm/debugfs.h>
	25
	26	/* This enables us to keep track of the memory removed from each node. */
	27	struct memtrace_entry {
	28	void *mem;
	29	u64 start;
	30	u64 size;
	31	u32 nid;
	32	struct dentry *dir;
	33	char name[16];
	34	};
	35
	36	static u64 memtrace_size;
	37
	38	static struct memtrace_entry *memtrace_array;
	39	static unsigned int memtrace_array_nr;
	40
	41
	42	static ssize_t memtrace_read(struct file filp, char __user ubuf,
	43	size_t count, loff_t *ppos)
	44	{
	45	struct memtrace_entry *ent = filp->private_data;
	46
	47	return simple_read_from_buffer(ubuf, count, ppos, ent->mem, ent->size);
	48	}
	49
	50	static bool valid_memtrace_range(struct memtrace_entry *dev,
	51	unsigned long start, unsigned long size)
	52	{
	53	if ((start >= dev->start) &&
	54	((start + size) <= (dev->start + dev->size)))
	55	return true;
	56
	57	return false;
	58	}
	59
	60	static int memtrace_mmap(struct file filp, struct vm_area_struct vma)
	61	{
	62	unsigned long size = vma->vm_end - vma->vm_start;
	63	struct memtrace_entry *dev = filp->private_data;
	64
65	if (!valid_memtrace_range(dev, vma->vm_pgoff << PAGE_SHIFT, size))
66	return -EINVAL;
67
68	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
69
70	if (remap_pfn_range(vma, vma->vm_start,
71	vma->vm_pgoff + (dev->start >> PAGE_SHIFT),
72	size, vma->vm_page_prot))
73	return -EAGAIN;
74
75	return 0;
76	}
77
78	static const struct file_operations memtrace_fops = {
79	.llseek = default_llseek,
80	.read = memtrace_read,
81	.mmap = memtrace_mmap,
82	.open = simple_open,
83	};
84
9d5171a8 RG	85	static int check_memblock_online(struct memory_block mem, void arg)
	86	{
	87	if (mem->state != MEM_ONLINE)
	88	return -1;
	89
	90	return 0;
	91	}
	92
	93	static int change_memblock_state(struct memory_block mem, void arg)
	94	{
	95	unsigned long state = (unsigned long)arg;
	96
	97	mem->state = state;
	98
	99	return 0;
	100	}
	101
	102	static bool memtrace_offline_pages(u32 nid, u64 start_pfn, u64 nr_pages)
	103	{
	104	u64 end_pfn = start_pfn + nr_pages - 1;
	105
	106	if (walk_memory_range(start_pfn, end_pfn, NULL,
	107	check_memblock_online))
	108	return false;
	109
	110	walk_memory_range(start_pfn, end_pfn, (void *)MEM_GOING_OFFLINE,
	111	change_memblock_state);
	112
	113	if (offline_pages(start_pfn, nr_pages)) {
	114	walk_memory_range(start_pfn, end_pfn, (void *)MEM_ONLINE,
	115	change_memblock_state);
	116	return false;
	117	}
	118
	119	walk_memory_range(start_pfn, end_pfn, (void *)MEM_OFFLINE,
	120	change_memblock_state);
	121
9d5171a8 RG	122	lock_device_hotplug();
	123	remove_memory(nid, start_pfn << PAGE_SHIFT, nr_pages << PAGE_SHIFT);
	124	unlock_device_hotplug();
	125
	126	return true;
	127	}
	128
	129	static u64 memtrace_alloc_node(u32 nid, u64 size)
	130	{
	131	u64 start_pfn, end_pfn, nr_pages;
	132	u64 base_pfn;
	133
	134	if (!NODE_DATA(nid) \|\| !node_spanned_pages(nid))
	135	return 0;
	136
	137	start_pfn = node_start_pfn(nid);
	138	end_pfn = node_end_pfn(nid);
	139	nr_pages = size >> PAGE_SHIFT;
	140
	141	/* Trace memory needs to be aligned to the size */
	142	end_pfn = round_down(end_pfn - nr_pages, nr_pages);
	143
	144	for (base_pfn = end_pfn; base_pfn > start_pfn; base_pfn -= nr_pages) {
	145	if (memtrace_offline_pages(nid, base_pfn, nr_pages) == true)
	146	return base_pfn << PAGE_SHIFT;
	147	}
	148
	149	return 0;
	150	}
	151
	152	static int memtrace_init_regions_runtime(u64 size)
	153	{
	154	u32 nid;
	155	u64 m;
	156
	157	memtrace_array = kcalloc(num_online_nodes(),
	158	sizeof(struct memtrace_entry), GFP_KERNEL);
	159	if (!memtrace_array) {
	160	pr_err("Failed to allocate memtrace_array\n");
	161	return -EINVAL;
	162	}
	163
	164	for_each_online_node(nid) {
	165	m = memtrace_alloc_node(nid, size);
	166
	167	/*
	168	* A node might not have any local memory, so warn but
	169	* continue on.
	170	*/
	171	if (!m) {
	172	pr_err("Failed to allocate trace memory on node %d\n", nid);
	173	continue;
	174	}
	175
	176	pr_info("Allocated trace memory on node %d at 0x%016llx\n", nid, m);
	177
	178	memtrace_array[memtrace_array_nr].start = m;
	179	memtrace_array[memtrace_array_nr].size = size;
	180	memtrace_array[memtrace_array_nr].nid = nid;
	181	memtrace_array_nr++;
	182	}
	183
	184	return 0;
	185	}
186
187	static struct dentry *memtrace_debugfs_dir;
188
189	static int memtrace_init_debugfs(void)
190	{
191	int ret = 0;
192	int i;
193
194	for (i = 0; i < memtrace_array_nr; i++) {
195	struct dentry *dir;
196	struct memtrace_entry *ent = &memtrace_array[i];
197
198	ent->mem = ioremap(ent->start, ent->size);
199	/* Warn but continue on */
200	if (!ent->mem) {
201	pr_err("Failed to map trace memory at 0x%llx\n",
202	ent->start);
203	ret = -1;
204	continue;
205	}
206
207	snprintf(ent->name, 16, "%08x", ent->nid);
208	dir = debugfs_create_dir(ent->name, memtrace_debugfs_dir);
209	if (!dir)
210	return -1;
211
212	ent->dir = dir;
213	debugfs_create_file("trace", 0400, dir, ent, &memtrace_fops);
214	debugfs_create_x64("start", 0400, dir, &ent->start);
215	debugfs_create_x64("size", 0400, dir, &ent->size);
216	}
217
218	return ret;
219	}
220
221	static int memtrace_enable_set(void *data, u64 val)
222	{
223	if (memtrace_size)
224	return -EINVAL;
225
226	if (!val)
227	return -EINVAL;
228
229	/* Make sure size is aligned to a memory block */
230	if (val & (memory_block_size_bytes() - 1))
231	return -EINVAL;
232
233	if (memtrace_init_regions_runtime(val))
234	return -EINVAL;
235
236	if (memtrace_init_debugfs())
237	return -EINVAL;
238
239	memtrace_size = val;
240
241	return 0;
242	}
243
244	static int memtrace_enable_get(void data, u64 val)
245	{
246	*val = memtrace_size;
247	return 0;
248	}
249
250	DEFINE_SIMPLE_ATTRIBUTE(memtrace_init_fops, memtrace_enable_get,
251	memtrace_enable_set, "0x%016llx\n");
252
253	static int memtrace_init(void)
254	{
255	memtrace_debugfs_dir = debugfs_create_dir("memtrace",
256	powerpc_debugfs_root);
257	if (!memtrace_debugfs_dir)
258	return -1;
259
260	debugfs_create_file("enable", 0600, memtrace_debugfs_dir,
261	NULL, &memtrace_init_fops);
262
263	return 0;
264	}
265	machine_device_initcall(powernv, memtrace_init);