[deliverable/linux.git] / drivers / nvdimm / pfn_devs.c

/*
 * Copyright(c) 2013-2015 Intel Corporation. All rights reserved.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of version 2 of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * 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.
 */
#include <linux/blkdev.h>
#include <linux/device.h>
#include <linux/genhd.h>
#include <linux/sizes.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include "nd-core.h"
#include "pfn.h"
#include "nd.h"

static void nd_pfn_release(struct device *dev)
{
	struct nd_region *nd_region = to_nd_region(dev->parent);
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);

	dev_dbg(dev, "%s\n", __func__);
	nd_detach_ndns(&nd_pfn->dev, &nd_pfn->ndns);
	ida_simple_remove(&nd_region->pfn_ida, nd_pfn->id);
	kfree(nd_pfn->uuid);
	kfree(nd_pfn);
}

static struct device_type nd_pfn_device_type = {
	.name = "nd_pfn",
	.release = nd_pfn_release,
};

bool is_nd_pfn(struct device *dev)
{
	return dev ? dev->type == &nd_pfn_device_type : false;
}
EXPORT_SYMBOL(is_nd_pfn);

struct nd_pfn *to_nd_pfn(struct device *dev)
{
	struct nd_pfn *nd_pfn = container_of(dev, struct nd_pfn, dev);

	WARN_ON(!is_nd_pfn(dev));
	return nd_pfn;
}
EXPORT_SYMBOL(to_nd_pfn);

static ssize_t mode_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);

	switch (nd_pfn->mode) {
	case PFN_MODE_RAM:
		return sprintf(buf, "ram\n");
	case PFN_MODE_PMEM:
		return sprintf(buf, "pmem\n");
	default:
		return sprintf(buf, "none\n");
	}
}

static ssize_t mode_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t len)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc = 0;

	device_lock(dev);
	nvdimm_bus_lock(dev);
	if (dev->driver)
		rc = -EBUSY;
	else {
		size_t n = len - 1;

		if (strncmp(buf, "pmem\n", n) == 0
				|| strncmp(buf, "pmem", n) == 0) {
			nd_pfn->mode = PFN_MODE_PMEM;
		} else if (strncmp(buf, "ram\n", n) == 0
				|| strncmp(buf, "ram", n) == 0)
			nd_pfn->mode = PFN_MODE_RAM;
		else if (strncmp(buf, "none\n", n) == 0
				|| strncmp(buf, "none", n) == 0)
			nd_pfn->mode = PFN_MODE_NONE;
		else
			rc = -EINVAL;
	}
	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
			rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	nvdimm_bus_unlock(dev);
	device_unlock(dev);

	return rc ? rc : len;
}
static DEVICE_ATTR_RW(mode);

static ssize_t align_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);

	return sprintf(buf, "%lx\n", nd_pfn->align);
}

static ssize_t __align_store(struct nd_pfn *nd_pfn, const char *buf)
{
	unsigned long val;
	int rc;

	rc = kstrtoul(buf, 0, &val);
	if (rc)
		return rc;

	if (!is_power_of_2(val) || val < PAGE_SIZE || val > SZ_1G)
		return -EINVAL;

	if (nd_pfn->dev.driver)
		return -EBUSY;
	else
		nd_pfn->align = val;

	return 0;
}

static ssize_t align_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t len)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	device_lock(dev);
	nvdimm_bus_lock(dev);
	rc = __align_store(nd_pfn, buf);
	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
			rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	nvdimm_bus_unlock(dev);
	device_unlock(dev);

	return rc ? rc : len;
}
static DEVICE_ATTR_RW(align);

static ssize_t uuid_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);

	if (nd_pfn->uuid)
		return sprintf(buf, "%pUb\n", nd_pfn->uuid);
	return sprintf(buf, "\n");
}

static ssize_t uuid_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t len)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	device_lock(dev);
	rc = nd_uuid_store(dev, &nd_pfn->uuid, buf, len);
	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
			rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	device_unlock(dev);

	return rc ? rc : len;
}
static DEVICE_ATTR_RW(uuid);

static ssize_t namespace_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	nvdimm_bus_lock(dev);
	rc = sprintf(buf, "%s\n", nd_pfn->ndns
			? dev_name(&nd_pfn->ndns->dev) : "");
	nvdimm_bus_unlock(dev);
	return rc;
}

static ssize_t namespace_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t len)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	device_lock(dev);
	nvdimm_bus_lock(dev);
	rc = nd_namespace_store(dev, &nd_pfn->ndns, buf, len);
	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
			rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	nvdimm_bus_unlock(dev);
	device_unlock(dev);

	return rc;
}
static DEVICE_ATTR_RW(namespace);

static ssize_t resource_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	device_lock(dev);
	if (dev->driver) {
		struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
		u64 offset = __le64_to_cpu(pfn_sb->dataoff);
		struct nd_namespace_common *ndns = nd_pfn->ndns;
		u32 start_pad = __le32_to_cpu(pfn_sb->start_pad);
		struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);

		rc = sprintf(buf, "%#llx\n", (unsigned long long) nsio->res.start
				+ start_pad + offset);
	} else {
		/* no address to convey if the pfn instance is disabled */
		rc = -ENXIO;
	}
	device_unlock(dev);

	return rc;
}
static DEVICE_ATTR_RO(resource);

static ssize_t size_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	ssize_t rc;

	device_lock(dev);
	if (dev->driver) {
		struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
		u64 offset = __le64_to_cpu(pfn_sb->dataoff);
		struct nd_namespace_common *ndns = nd_pfn->ndns;
		u32 start_pad = __le32_to_cpu(pfn_sb->start_pad);
		u32 end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
		struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);

		rc = sprintf(buf, "%llu\n", (unsigned long long)
				resource_size(&nsio->res) - start_pad
				- end_trunc - offset);
	} else {
		/* no size to convey if the pfn instance is disabled */
		rc = -ENXIO;
	}
	device_unlock(dev);

	return rc;
}
static DEVICE_ATTR_RO(size);

static struct attribute *nd_pfn_attributes[] = {
	&dev_attr_mode.attr,
	&dev_attr_namespace.attr,
	&dev_attr_uuid.attr,
	&dev_attr_align.attr,
	&dev_attr_resource.attr,
	&dev_attr_size.attr,
	NULL,
};

static struct attribute_group nd_pfn_attribute_group = {
	.attrs = nd_pfn_attributes,
};

static const struct attribute_group *nd_pfn_attribute_groups[] = {
	&nd_pfn_attribute_group,
	&nd_device_attribute_group,
	&nd_numa_attribute_group,
	NULL,
};

static struct device *__nd_pfn_create(struct nd_region *nd_region,
		struct nd_namespace_common *ndns)
{
	struct nd_pfn *nd_pfn;
	struct device *dev;

	/* we can only create pages for contiguous ranged of pmem */
	if (!is_nd_pmem(&nd_region->dev))
		return NULL;

	nd_pfn = kzalloc(sizeof(*nd_pfn), GFP_KERNEL);
	if (!nd_pfn)
		return NULL;

	nd_pfn->id = ida_simple_get(&nd_region->pfn_ida, 0, 0, GFP_KERNEL);
	if (nd_pfn->id < 0) {
		kfree(nd_pfn);
		return NULL;
	}

	nd_pfn->mode = PFN_MODE_NONE;
	nd_pfn->align = HPAGE_SIZE;
	dev = &nd_pfn->dev;
	dev_set_name(dev, "pfn%d.%d", nd_region->id, nd_pfn->id);
	dev->parent = &nd_region->dev;
	dev->type = &nd_pfn_device_type;
	dev->groups = nd_pfn_attribute_groups;
	device_initialize(&nd_pfn->dev);
	if (ndns && !__nd_attach_ndns(&nd_pfn->dev, ndns, &nd_pfn->ndns)) {
		dev_dbg(&ndns->dev, "%s failed, already claimed by %s\n",
				__func__, dev_name(ndns->claim));
		put_device(dev);
		return NULL;
	}
	return dev;
}

struct device *nd_pfn_create(struct nd_region *nd_region)
{
	struct device *dev = __nd_pfn_create(nd_region, NULL);

	if (dev)
		__nd_device_register(dev);
	return dev;
}

int nd_pfn_validate(struct nd_pfn *nd_pfn)
{
	u64 checksum, offset;
	struct nd_namespace_io *nsio;
	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	struct nd_namespace_common *ndns = nd_pfn->ndns;
	const u8 *parent_uuid = nd_dev_to_uuid(&ndns->dev);

	if (!pfn_sb || !ndns)
		return -ENODEV;

	if (!is_nd_pmem(nd_pfn->dev.parent))
		return -ENODEV;

	if (nvdimm_read_bytes(ndns, SZ_4K, pfn_sb, sizeof(*pfn_sb)))
		return -ENXIO;

	if (memcmp(pfn_sb->signature, PFN_SIG, PFN_SIG_LEN) != 0)
		return -ENODEV;

	checksum = le64_to_cpu(pfn_sb->checksum);
	pfn_sb->checksum = 0;
	if (checksum != nd_sb_checksum((struct nd_gen_sb *) pfn_sb))
		return -ENODEV;
	pfn_sb->checksum = cpu_to_le64(checksum);

	if (memcmp(pfn_sb->parent_uuid, parent_uuid, 16) != 0)
		return -ENODEV;

	if (__le16_to_cpu(pfn_sb->version_minor) < 1) {
		pfn_sb->start_pad = 0;
		pfn_sb->end_trunc = 0;
	}

	switch (le32_to_cpu(pfn_sb->mode)) {
	case PFN_MODE_RAM:
	case PFN_MODE_PMEM:
		break;
	default:
		return -ENXIO;
	}

	if (!nd_pfn->uuid) {
		/* from probe we allocate */
		nd_pfn->uuid = kmemdup(pfn_sb->uuid, 16, GFP_KERNEL);
		if (!nd_pfn->uuid)
			return -ENOMEM;
	} else {
		/* from init we validate */
		if (memcmp(nd_pfn->uuid, pfn_sb->uuid, 16) != 0)
			return -ENODEV;
	}

	if (nd_pfn->align > nvdimm_namespace_capacity(ndns)) {
		dev_err(&nd_pfn->dev, "alignment: %lx exceeds capacity %llx\n",
				nd_pfn->align, nvdimm_namespace_capacity(ndns));
		return -EINVAL;
	}

	/*
	 * These warnings are verbose because they can only trigger in
	 * the case where the physical address alignment of the
	 * namespace has changed since the pfn superblock was
	 * established.
	 */
	offset = le64_to_cpu(pfn_sb->dataoff);
	nsio = to_nd_namespace_io(&ndns->dev);
	if (offset >= resource_size(&nsio->res)) {
		dev_err(&nd_pfn->dev, "pfn array size exceeds capacity of %s\n",
				dev_name(&ndns->dev));
		return -EBUSY;
	}

	nd_pfn->align = 1UL << ilog2(offset);
	if (!is_power_of_2(offset) || offset < PAGE_SIZE) {
		dev_err(&nd_pfn->dev, "bad offset: %#llx dax disabled\n",
				offset);
		return -ENXIO;
	}

	return 0;
}
EXPORT_SYMBOL(nd_pfn_validate);

int nd_pfn_probe(struct nd_namespace_common *ndns, void *drvdata)
{
	int rc;
	struct device *dev;
	struct nd_pfn *nd_pfn;
	struct nd_pfn_sb *pfn_sb;
	struct nd_region *nd_region = to_nd_region(ndns->dev.parent);

	if (ndns->force_raw)
		return -ENODEV;

	nvdimm_bus_lock(&ndns->dev);
	dev = __nd_pfn_create(nd_region, ndns);
	nvdimm_bus_unlock(&ndns->dev);
	if (!dev)
		return -ENOMEM;
	dev_set_drvdata(dev, drvdata);
	pfn_sb = kzalloc(sizeof(*pfn_sb), GFP_KERNEL);
	nd_pfn = to_nd_pfn(dev);
	nd_pfn->pfn_sb = pfn_sb;
	rc = nd_pfn_validate(nd_pfn);
	nd_pfn->pfn_sb = NULL;
	kfree(pfn_sb);
	dev_dbg(&ndns->dev, "%s: pfn: %s\n", __func__,
			rc == 0 ? dev_name(dev) : "<none>");
	if (rc < 0) {
		__nd_detach_ndns(dev, &nd_pfn->ndns);
		put_device(dev);
	} else
		__nd_device_register(&nd_pfn->dev);

	return rc;
}
EXPORT_SYMBOL(nd_pfn_probe);
Commit	Line	Data
e1455744 DW	1	/*
	2	* Copyright(c) 2013-2015 Intel Corporation. All rights reserved.
	3	*
	4	* This program is free software; you can redistribute it and/or modify
	5	* it under the terms of version 2 of the GNU General Public License as
	6	* published by the Free Software Foundation.
	7	*
	8	* This program is distributed in the hope that it will be useful, but
	9	* WITHOUT ANY WARRANTY; without even the implied warranty of
	10	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
	11	* General Public License for more details.
	12	*/
	13	#include <linux/blkdev.h>
	14	#include <linux/device.h>
	15	#include <linux/genhd.h>
	16	#include <linux/sizes.h>
	17	#include <linux/slab.h>
	18	#include <linux/fs.h>
	19	#include <linux/mm.h>
	20	#include "nd-core.h"
	21	#include "pfn.h"
	22	#include "nd.h"
	23
	24	static void nd_pfn_release(struct device *dev)
	25	{
	26	struct nd_region *nd_region = to_nd_region(dev->parent);
	27	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	28
	29	dev_dbg(dev, "%s\n", __func__);
	30	nd_detach_ndns(&nd_pfn->dev, &nd_pfn->ndns);
	31	ida_simple_remove(&nd_region->pfn_ida, nd_pfn->id);
	32	kfree(nd_pfn->uuid);
	33	kfree(nd_pfn);
	34	}
	35
	36	static struct device_type nd_pfn_device_type = {
	37	.name = "nd_pfn",
	38	.release = nd_pfn_release,
	39	};
	40
	41	bool is_nd_pfn(struct device *dev)
	42	{
	43	return dev ? dev->type == &nd_pfn_device_type : false;
	44	}
	45	EXPORT_SYMBOL(is_nd_pfn);
	46
	47	struct nd_pfn to_nd_pfn(struct device dev)
	48	{
	49	struct nd_pfn *nd_pfn = container_of(dev, struct nd_pfn, dev);
	50
	51	WARN_ON(!is_nd_pfn(dev));
	52	return nd_pfn;
	53	}
	54	EXPORT_SYMBOL(to_nd_pfn);
	55
	56	static ssize_t mode_show(struct device *dev,
	57	struct device_attribute attr, char buf)
	58	{
	59	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	60
	61	switch (nd_pfn->mode) {
	62	case PFN_MODE_RAM:
	63	return sprintf(buf, "ram\n");
	64	case PFN_MODE_PMEM:
65	return sprintf(buf, "pmem\n");
66	default:
67	return sprintf(buf, "none\n");
68	}
69	}
70
71	static ssize_t mode_store(struct device *dev,
72	struct device_attribute attr, const char buf, size_t len)
73	{
74	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
75	ssize_t rc = 0;
76
77	device_lock(dev);
78	nvdimm_bus_lock(dev);
79	if (dev->driver)
80	rc = -EBUSY;
81	else {
82	size_t n = len - 1;
83
84	if (strncmp(buf, "pmem\n", n) == 0
85	\|\| strncmp(buf, "pmem", n) == 0) {
d2c0f041	86	nd_pfn->mode = PFN_MODE_PMEM;
e1455744 DW	87	} else if (strncmp(buf, "ram\n", n) == 0
	88	\|\| strncmp(buf, "ram", n) == 0)
	89	nd_pfn->mode = PFN_MODE_RAM;
	90	else if (strncmp(buf, "none\n", n) == 0
	91	\|\| strncmp(buf, "none", n) == 0)
	92	nd_pfn->mode = PFN_MODE_NONE;
	93	else
	94	rc = -EINVAL;
	95	}
	96	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
	97	rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	98	nvdimm_bus_unlock(dev);
	99	device_unlock(dev);
	100
	101	return rc ? rc : len;
	102	}
	103	static DEVICE_ATTR_RW(mode);
	104
315c5625 DW	105	static ssize_t align_show(struct device *dev,
	106	struct device_attribute attr, char buf)
	107	{
	108	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	109
	110	return sprintf(buf, "%lx\n", nd_pfn->align);
	111	}
	112
	113	static ssize_t __align_store(struct nd_pfn nd_pfn, const char buf)
	114	{
	115	unsigned long val;
	116	int rc;
	117
	118	rc = kstrtoul(buf, 0, &val);
	119	if (rc)
	120	return rc;
	121
	122	if (!is_power_of_2(val) \|\| val < PAGE_SIZE \|\| val > SZ_1G)
	123	return -EINVAL;
	124
	125	if (nd_pfn->dev.driver)
	126	return -EBUSY;
	127	else
	128	nd_pfn->align = val;
	129
	130	return 0;
	131	}
	132
	133	static ssize_t align_store(struct device *dev,
	134	struct device_attribute attr, const char buf, size_t len)
	135	{
	136	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	137	ssize_t rc;
	138
	139	device_lock(dev);
	140	nvdimm_bus_lock(dev);
	141	rc = __align_store(nd_pfn, buf);
	142	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
	143	rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	144	nvdimm_bus_unlock(dev);
	145	device_unlock(dev);
	146
	147	return rc ? rc : len;
	148	}
	149	static DEVICE_ATTR_RW(align);
	150
e1455744 DW	151	static ssize_t uuid_show(struct device *dev,
	152	struct device_attribute attr, char buf)
	153	{
	154	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	155
	156	if (nd_pfn->uuid)
	157	return sprintf(buf, "%pUb\n", nd_pfn->uuid);
	158	return sprintf(buf, "\n");
	159	}
	160
	161	static ssize_t uuid_store(struct device *dev,
	162	struct device_attribute attr, const char buf, size_t len)
	163	{
	164	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	165	ssize_t rc;
	166
	167	device_lock(dev);
	168	rc = nd_uuid_store(dev, &nd_pfn->uuid, buf, len);
	169	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
	170	rc, buf, buf[len - 1] == '\n' ? "" : "\n");
	171	device_unlock(dev);
	172
	173	return rc ? rc : len;
	174	}
	175	static DEVICE_ATTR_RW(uuid);
	176
	177	static ssize_t namespace_show(struct device *dev,
	178	struct device_attribute attr, char buf)
	179	{
	180	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	181	ssize_t rc;
	182
	183	nvdimm_bus_lock(dev);
	184	rc = sprintf(buf, "%s\n", nd_pfn->ndns
	185	? dev_name(&nd_pfn->ndns->dev) : "");
	186	nvdimm_bus_unlock(dev);
	187	return rc;
	188	}
	189
	190	static ssize_t namespace_store(struct device *dev,
	191	struct device_attribute attr, const char buf, size_t len)
	192	{
	193	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	194	ssize_t rc;
	195
e1455744	196	device_lock(dev);
4ca8b57a	197	nvdimm_bus_lock(dev);
e1455744 DW	198	rc = nd_namespace_store(dev, &nd_pfn->ndns, buf, len);
	199	dev_dbg(dev, "%s: result: %zd wrote: %s%s", __func__,
	200	rc, buf, buf[len - 1] == '\n' ? "" : "\n");
e1455744	201	nvdimm_bus_unlock(dev);
4ca8b57a	202	device_unlock(dev);
e1455744 DW	203
	204	return rc;
	205	}
	206	static DEVICE_ATTR_RW(namespace);
	207
f6ed58c7 DW	208	static ssize_t resource_show(struct device *dev,
	209	struct device_attribute attr, char buf)
	210	{
	211	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	212	ssize_t rc;
	213
	214	device_lock(dev);
	215	if (dev->driver) {
	216	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	217	u64 offset = __le64_to_cpu(pfn_sb->dataoff);
	218	struct nd_namespace_common *ndns = nd_pfn->ndns;
	219	u32 start_pad = __le32_to_cpu(pfn_sb->start_pad);
	220	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	221
	222	rc = sprintf(buf, "%#llx\n", (unsigned long long) nsio->res.start
	223	+ start_pad + offset);
	224	} else {
	225	/* no address to convey if the pfn instance is disabled */
	226	rc = -ENXIO;
	227	}
	228	device_unlock(dev);
	229
	230	return rc;
	231	}
	232	static DEVICE_ATTR_RO(resource);
	233
	234	static ssize_t size_show(struct device *dev,
	235	struct device_attribute attr, char buf)
	236	{
	237	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	238	ssize_t rc;
	239
	240	device_lock(dev);
	241	if (dev->driver) {
	242	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	243	u64 offset = __le64_to_cpu(pfn_sb->dataoff);
	244	struct nd_namespace_common *ndns = nd_pfn->ndns;
	245	u32 start_pad = __le32_to_cpu(pfn_sb->start_pad);
	246	u32 end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
	247	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	248
	249	rc = sprintf(buf, "%llu\n", (unsigned long long)
	250	resource_size(&nsio->res) - start_pad
	251	- end_trunc - offset);
	252	} else {
	253	/* no size to convey if the pfn instance is disabled */
	254	rc = -ENXIO;
	255	}
	256	device_unlock(dev);
	257
	258	return rc;
	259	}
	260	static DEVICE_ATTR_RO(size);
	261
e1455744 DW	262	static struct attribute *nd_pfn_attributes[] = {
	263	&dev_attr_mode.attr,
	264	&dev_attr_namespace.attr,
	265	&dev_attr_uuid.attr,
315c5625	266	&dev_attr_align.attr,
f6ed58c7 DW	267	&dev_attr_resource.attr,
f6ed58c7 DW	268	&dev_attr_size.attr,
e1455744 DW	269	NULL,
	270	};
	271
	272	static struct attribute_group nd_pfn_attribute_group = {
	273	.attrs = nd_pfn_attributes,
	274	};
	275
	276	static const struct attribute_group *nd_pfn_attribute_groups[] = {
	277	&nd_pfn_attribute_group,
	278	&nd_device_attribute_group,
	279	&nd_numa_attribute_group,
	280	NULL,
	281	};
	282
	283	static struct device __nd_pfn_create(struct nd_region nd_region,
e1455744 DW	284	struct nd_namespace_common *ndns)
	285	{
	286	struct nd_pfn *nd_pfn;
	287	struct device *dev;
	288
	289	/* we can only create pages for contiguous ranged of pmem */
	290	if (!is_nd_pmem(&nd_region->dev))
	291	return NULL;
	292
	293	nd_pfn = kzalloc(sizeof(*nd_pfn), GFP_KERNEL);
	294	if (!nd_pfn)
	295	return NULL;
	296
	297	nd_pfn->id = ida_simple_get(&nd_region->pfn_ida, 0, 0, GFP_KERNEL);
	298	if (nd_pfn->id < 0) {
	299	kfree(nd_pfn);
	300	return NULL;
	301	}
	302
f7c6ab80	303	nd_pfn->mode = PFN_MODE_NONE;
315c5625	304	nd_pfn->align = HPAGE_SIZE;
e1455744 DW	305	dev = &nd_pfn->dev;
	306	dev_set_name(dev, "pfn%d.%d", nd_region->id, nd_pfn->id);
	307	dev->parent = &nd_region->dev;
	308	dev->type = &nd_pfn_device_type;
	309	dev->groups = nd_pfn_attribute_groups;
	310	device_initialize(&nd_pfn->dev);
	311	if (ndns && !__nd_attach_ndns(&nd_pfn->dev, ndns, &nd_pfn->ndns)) {
	312	dev_dbg(&ndns->dev, "%s failed, already claimed by %s\n",
	313	__func__, dev_name(ndns->claim));
	314	put_device(dev);
	315	return NULL;
	316	}
	317	return dev;
	318	}
	319
	320	struct device nd_pfn_create(struct nd_region nd_region)
	321	{
f7c6ab80	322	struct device *dev = __nd_pfn_create(nd_region, NULL);
e1455744 DW	323
	324	if (dev)
	325	__nd_device_register(dev);
	326	return dev;
	327	}
	328
32ab0a3f	329	int nd_pfn_validate(struct nd_pfn *nd_pfn)
e1455744	330	{
e1455744	331	u64 checksum, offset;
a34d5e8a DW	332	struct nd_namespace_io *nsio;
	333	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	334	struct nd_namespace_common *ndns = nd_pfn->ndns;
	335	const u8 *parent_uuid = nd_dev_to_uuid(&ndns->dev);
e1455744 DW	336
	337	if (!pfn_sb \|\| !ndns)
	338	return -ENODEV;
	339
	340	if (!is_nd_pmem(nd_pfn->dev.parent))
	341	return -ENODEV;
	342
e1455744 DW	343	if (nvdimm_read_bytes(ndns, SZ_4K, pfn_sb, sizeof(*pfn_sb)))
	344	return -ENXIO;
	345
	346	if (memcmp(pfn_sb->signature, PFN_SIG, PFN_SIG_LEN) != 0)
	347	return -ENODEV;
	348
	349	checksum = le64_to_cpu(pfn_sb->checksum);
	350	pfn_sb->checksum = 0;
	351	if (checksum != nd_sb_checksum((struct nd_gen_sb *) pfn_sb))
	352	return -ENODEV;
	353	pfn_sb->checksum = cpu_to_le64(checksum);
	354
a34d5e8a DW	355	if (memcmp(pfn_sb->parent_uuid, parent_uuid, 16) != 0)
	356	return -ENODEV;
	357
cfe30b87 DW	358	if (__le16_to_cpu(pfn_sb->version_minor) < 1) {
	359	pfn_sb->start_pad = 0;
	360	pfn_sb->end_trunc = 0;
	361	}
	362
e1455744 DW	363	switch (le32_to_cpu(pfn_sb->mode)) {
e1455744 DW	364	case PFN_MODE_RAM:
e1455744	365	case PFN_MODE_PMEM:
45eb570a	366	break;
e1455744 DW	367	default:
	368	return -ENXIO;
	369	}
	370
	371	if (!nd_pfn->uuid) {
	372	/* from probe we allocate */
	373	nd_pfn->uuid = kmemdup(pfn_sb->uuid, 16, GFP_KERNEL);
	374	if (!nd_pfn->uuid)
	375	return -ENOMEM;
	376	} else {
	377	/* from init we validate */
	378	if (memcmp(nd_pfn->uuid, pfn_sb->uuid, 16) != 0)
e5670563	379	return -ENODEV;
e1455744 DW	380	}
e1455744 DW	381
315c5625 DW	382	if (nd_pfn->align > nvdimm_namespace_capacity(ndns)) {
	383	dev_err(&nd_pfn->dev, "alignment: %lx exceeds capacity %llx\n",
	384	nd_pfn->align, nvdimm_namespace_capacity(ndns));
	385	return -EINVAL;
	386	}
	387
e1455744 DW	388	/*
	389	* These warnings are verbose because they can only trigger in
	390	* the case where the physical address alignment of the
	391	* namespace has changed since the pfn superblock was
	392	* established.
	393	*/
	394	offset = le64_to_cpu(pfn_sb->dataoff);
	395	nsio = to_nd_namespace_io(&ndns->dev);
9f1e8cee	396	if (offset >= resource_size(&nsio->res)) {
e1455744 DW	397	dev_err(&nd_pfn->dev, "pfn array size exceeds capacity of %s\n",
	398	dev_name(&ndns->dev));
	399	return -EBUSY;
	400	}
	401
315c5625 DW	402	nd_pfn->align = 1UL << ilog2(offset);
	403	if (!is_power_of_2(offset) \|\| offset < PAGE_SIZE) {
	404	dev_err(&nd_pfn->dev, "bad offset: %#llx dax disabled\n",
	405	offset);
	406	return -ENXIO;
	407	}
	408
e1455744 DW	409	return 0;
e1455744 DW	410	}
32ab0a3f	411	EXPORT_SYMBOL(nd_pfn_validate);
e1455744 DW	412
	413	int nd_pfn_probe(struct nd_namespace_common ndns, void drvdata)
	414	{
	415	int rc;
	416	struct device *dev;
	417	struct nd_pfn *nd_pfn;
	418	struct nd_pfn_sb *pfn_sb;
	419	struct nd_region *nd_region = to_nd_region(ndns->dev.parent);
	420
	421	if (ndns->force_raw)
	422	return -ENODEV;
	423
	424	nvdimm_bus_lock(&ndns->dev);
f7c6ab80	425	dev = __nd_pfn_create(nd_region, ndns);
e1455744 DW	426	nvdimm_bus_unlock(&ndns->dev);
	427	if (!dev)
	428	return -ENOMEM;
	429	dev_set_drvdata(dev, drvdata);
	430	pfn_sb = kzalloc(sizeof(*pfn_sb), GFP_KERNEL);
	431	nd_pfn = to_nd_pfn(dev);
	432	nd_pfn->pfn_sb = pfn_sb;
	433	rc = nd_pfn_validate(nd_pfn);
	434	nd_pfn->pfn_sb = NULL;
	435	kfree(pfn_sb);
	436	dev_dbg(&ndns->dev, "%s: pfn: %s\n", __func__,
	437	rc == 0 ? dev_name(dev) : "<none>");
	438	if (rc < 0) {
	439	__nd_detach_ndns(dev, &nd_pfn->ndns);
	440	put_device(dev);
	441	} else
	442	__nd_device_register(&nd_pfn->dev);
	443
	444	return rc;
	445	}
	446	EXPORT_SYMBOL(nd_pfn_probe);