[mirror_ubuntu-bionic-kernel.git] / drivers / nvdimm / dimm_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.
 */
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/vmalloc.h>
#include <linux/device.h>
#include <linux/ndctl.h>
#include <linux/slab.h>
#include <linux/io.h>
#include <linux/fs.h>
#include <linux/mm.h>
#include "nd-core.h"
#include "label.h"
#include "nd.h"

static DEFINE_IDA(dimm_ida);

/*
 * Retrieve bus and dimm handle and return if this bus supports
 * get_config_data commands
 */
int nvdimm_check_config_data(struct device *dev)
{
	struct nvdimm *nvdimm = to_nvdimm(dev);

	if (!nvdimm->cmd_mask ||
	    !test_bit(ND_CMD_GET_CONFIG_DATA, &nvdimm->cmd_mask)) {
		if (nvdimm->flags & NDD_ALIASING)
			return -ENXIO;
		else
			return -ENOTTY;
	}

	return 0;
}

static int validate_dimm(struct nvdimm_drvdata *ndd)
{
	int rc;

	if (!ndd)
		return -EINVAL;

	rc = nvdimm_check_config_data(ndd->dev);
	if (rc)
		dev_dbg(ndd->dev, "%pf: %s error: %d\n",
				__builtin_return_address(0), __func__, rc);
	return rc;
}

/**
 * nvdimm_init_nsarea - determine the geometry of a dimm's namespace area
 * @nvdimm: dimm to initialize
 */
int nvdimm_init_nsarea(struct nvdimm_drvdata *ndd)
{
	struct nd_cmd_get_config_size *cmd = &ndd->nsarea;
	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	struct nvdimm_bus_descriptor *nd_desc;
	int rc = validate_dimm(ndd);

	if (rc)
		return rc;

	if (cmd->config_size)
		return 0; /* already valid */

	memset(cmd, 0, sizeof(*cmd));
	nd_desc = nvdimm_bus->nd_desc;
	return nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
			ND_CMD_GET_CONFIG_SIZE, cmd, sizeof(*cmd), NULL);
}

int nvdimm_init_config_data(struct nvdimm_drvdata *ndd)
{
	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	struct nd_cmd_get_config_data_hdr *cmd;
	struct nvdimm_bus_descriptor *nd_desc;
	int rc = validate_dimm(ndd);
	u32 max_cmd_size, config_size;
	size_t offset;

	if (rc)
		return rc;

	if (ndd->data)
		return 0;

	if (ndd->nsarea.status || ndd->nsarea.max_xfer == 0
			|| ndd->nsarea.config_size < ND_LABEL_MIN_SIZE) {
		dev_dbg(ndd->dev, "failed to init config data area: (%d:%d)\n",
				ndd->nsarea.max_xfer, ndd->nsarea.config_size);
		return -ENXIO;
	}

	ndd->data = kmalloc(ndd->nsarea.config_size, GFP_KERNEL);
	if (!ndd->data)
		ndd->data = vmalloc(ndd->nsarea.config_size);

	if (!ndd->data)
		return -ENOMEM;

	max_cmd_size = min_t(u32, PAGE_SIZE, ndd->nsarea.max_xfer);
	cmd = kzalloc(max_cmd_size + sizeof(*cmd), GFP_KERNEL);
	if (!cmd)
		return -ENOMEM;

	nd_desc = nvdimm_bus->nd_desc;
	for (config_size = ndd->nsarea.config_size, offset = 0;
			config_size; config_size -= cmd->in_length,
			offset += cmd->in_length) {
		cmd->in_length = min(config_size, max_cmd_size);
		cmd->in_offset = offset;
		rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
				ND_CMD_GET_CONFIG_DATA, cmd,
				cmd->in_length + sizeof(*cmd), NULL);
		if (rc || cmd->status) {
			rc = -ENXIO;
			break;
		}
		memcpy(ndd->data + offset, cmd->out_buf, cmd->in_length);
	}
	dev_dbg(ndd->dev, "%s: len: %zu rc: %d\n", __func__, offset, rc);
	kfree(cmd);

	return rc;
}

int nvdimm_set_config_data(struct nvdimm_drvdata *ndd, size_t offset,
		void *buf, size_t len)
{
	int rc = validate_dimm(ndd);
	size_t max_cmd_size, buf_offset;
	struct nd_cmd_set_config_hdr *cmd;
	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	struct nvdimm_bus_descriptor *nd_desc = nvdimm_bus->nd_desc;

	if (rc)
		return rc;

	if (!ndd->data)
		return -ENXIO;

	if (offset + len > ndd->nsarea.config_size)
		return -ENXIO;

	max_cmd_size = min_t(u32, PAGE_SIZE, len);
	max_cmd_size = min_t(u32, max_cmd_size, ndd->nsarea.max_xfer);
	cmd = kzalloc(max_cmd_size + sizeof(*cmd) + sizeof(u32), GFP_KERNEL);
	if (!cmd)
		return -ENOMEM;

	for (buf_offset = 0; len; len -= cmd->in_length,
			buf_offset += cmd->in_length) {
		size_t cmd_size;
		u32 *status;

		cmd->in_offset = offset + buf_offset;
		cmd->in_length = min(max_cmd_size, len);
		memcpy(cmd->in_buf, buf + buf_offset, cmd->in_length);

		/* status is output in the last 4-bytes of the command buffer */
		cmd_size = sizeof(*cmd) + cmd->in_length + sizeof(u32);
		status = ((void *) cmd) + cmd_size - sizeof(u32);

		rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
				ND_CMD_SET_CONFIG_DATA, cmd, cmd_size, NULL);
		if (rc || *status) {
			rc = rc ? rc : -ENXIO;
			break;
		}
	}
	kfree(cmd);

	return rc;
}

void nvdimm_set_aliasing(struct device *dev)
{
	struct nvdimm *nvdimm = to_nvdimm(dev);

	nvdimm->flags |= NDD_ALIASING;
}

static void nvdimm_release(struct device *dev)
{
	struct nvdimm *nvdimm = to_nvdimm(dev);

	ida_simple_remove(&dimm_ida, nvdimm->id);
	kfree(nvdimm);
}

static struct device_type nvdimm_device_type = {
	.name = "nvdimm",
	.release = nvdimm_release,
};

bool is_nvdimm(struct device *dev)
{
	return dev->type == &nvdimm_device_type;
}

struct nvdimm *to_nvdimm(struct device *dev)
{
	struct nvdimm *nvdimm = container_of(dev, struct nvdimm, dev);

	WARN_ON(!is_nvdimm(dev));
	return nvdimm;
}
EXPORT_SYMBOL_GPL(to_nvdimm);

struct nvdimm *nd_blk_region_to_dimm(struct nd_blk_region *ndbr)
{
	struct nd_region *nd_region = &ndbr->nd_region;
	struct nd_mapping *nd_mapping = &nd_region->mapping[0];

	return nd_mapping->nvdimm;
}
EXPORT_SYMBOL_GPL(nd_blk_region_to_dimm);

struct nvdimm_drvdata *to_ndd(struct nd_mapping *nd_mapping)
{
	struct nvdimm *nvdimm = nd_mapping->nvdimm;

	WARN_ON_ONCE(!is_nvdimm_bus_locked(&nvdimm->dev));

	return dev_get_drvdata(&nvdimm->dev);
}
EXPORT_SYMBOL(to_ndd);

void nvdimm_drvdata_release(struct kref *kref)
{
	struct nvdimm_drvdata *ndd = container_of(kref, typeof(*ndd), kref);
	struct device *dev = ndd->dev;
	struct resource *res, *_r;

	dev_dbg(dev, "%s\n", __func__);

	nvdimm_bus_lock(dev);
	for_each_dpa_resource_safe(ndd, res, _r)
		nvdimm_free_dpa(ndd, res);
	nvdimm_bus_unlock(dev);

	kvfree(ndd->data);
	kfree(ndd);
	put_device(dev);
}

void get_ndd(struct nvdimm_drvdata *ndd)
{
	kref_get(&ndd->kref);
}

void put_ndd(struct nvdimm_drvdata *ndd)
{
	if (ndd)
		kref_put(&ndd->kref, nvdimm_drvdata_release);
}

const char *nvdimm_name(struct nvdimm *nvdimm)
{
	return dev_name(&nvdimm->dev);
}
EXPORT_SYMBOL_GPL(nvdimm_name);

struct kobject *nvdimm_kobj(struct nvdimm *nvdimm)
{
	return &nvdimm->dev.kobj;
}
EXPORT_SYMBOL_GPL(nvdimm_kobj);

unsigned long nvdimm_cmd_mask(struct nvdimm *nvdimm)
{
	return nvdimm->cmd_mask;
}
EXPORT_SYMBOL_GPL(nvdimm_cmd_mask);

void *nvdimm_provider_data(struct nvdimm *nvdimm)
{
	if (nvdimm)
		return nvdimm->provider_data;
	return NULL;
}
EXPORT_SYMBOL_GPL(nvdimm_provider_data);

static ssize_t commands_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nvdimm *nvdimm = to_nvdimm(dev);
	int cmd, len = 0;

	if (!nvdimm->cmd_mask)
		return sprintf(buf, "\n");

	for_each_set_bit(cmd, &nvdimm->cmd_mask, BITS_PER_LONG)
		len += sprintf(buf + len, "%s ", nvdimm_cmd_name(cmd));
	len += sprintf(buf + len, "\n");
	return len;
}
static DEVICE_ATTR_RO(commands);

static ssize_t state_show(struct device *dev, struct device_attribute *attr,
		char *buf)
{
	struct nvdimm *nvdimm = to_nvdimm(dev);

	/*
	 * The state may be in the process of changing, userspace should
	 * quiesce probing if it wants a static answer
	 */
	nvdimm_bus_lock(dev);
	nvdimm_bus_unlock(dev);
	return sprintf(buf, "%s\n", atomic_read(&nvdimm->busy)
			? "active" : "idle");
}
static DEVICE_ATTR_RO(state);

static ssize_t available_slots_show(struct device *dev,
		struct device_attribute *attr, char *buf)
{
	struct nvdimm_drvdata *ndd = dev_get_drvdata(dev);
	ssize_t rc;
	u32 nfree;

	if (!ndd)
		return -ENXIO;

	nvdimm_bus_lock(dev);
	nfree = nd_label_nfree(ndd);
	if (nfree - 1 > nfree) {
		dev_WARN_ONCE(dev, 1, "we ate our last label?\n");
		nfree = 0;
	} else
		nfree--;
	rc = sprintf(buf, "%d\n", nfree);
	nvdimm_bus_unlock(dev);
	return rc;
}
static DEVICE_ATTR_RO(available_slots);

static struct attribute *nvdimm_attributes[] = {
	&dev_attr_state.attr,
	&dev_attr_commands.attr,
	&dev_attr_available_slots.attr,
	NULL,
};

struct attribute_group nvdimm_attribute_group = {
	.attrs = nvdimm_attributes,
};
EXPORT_SYMBOL_GPL(nvdimm_attribute_group);

struct nvdimm *nvdimm_create(struct nvdimm_bus *nvdimm_bus, void *provider_data,
		const struct attribute_group **groups, unsigned long flags,
		unsigned long cmd_mask, int num_flush,
		struct resource *flush_wpq)
{
	struct nvdimm *nvdimm = kzalloc(sizeof(*nvdimm), GFP_KERNEL);
	struct device *dev;

	if (!nvdimm)
		return NULL;

	nvdimm->id = ida_simple_get(&dimm_ida, 0, 0, GFP_KERNEL);
	if (nvdimm->id < 0) {
		kfree(nvdimm);
		return NULL;
	}
	nvdimm->provider_data = provider_data;
	nvdimm->flags = flags;
	nvdimm->cmd_mask = cmd_mask;
	nvdimm->num_flush = num_flush;
	nvdimm->flush_wpq = flush_wpq;
	atomic_set(&nvdimm->busy, 0);
	dev = &nvdimm->dev;
	dev_set_name(dev, "nmem%d", nvdimm->id);
	dev->parent = &nvdimm_bus->dev;
	dev->type = &nvdimm_device_type;
	dev->devt = MKDEV(nvdimm_major, nvdimm->id);
	dev->groups = groups;
	nd_device_register(dev);

	return nvdimm;
}
EXPORT_SYMBOL_GPL(nvdimm_create);

int alias_dpa_busy(struct device *dev, void *data)
{
	resource_size_t map_end, blk_start, new;
	struct blk_alloc_info *info = data;
	struct nd_mapping *nd_mapping;
	struct nd_region *nd_region;
	struct nvdimm_drvdata *ndd;
	struct resource *res;
	int i;

	if (!is_nd_pmem(dev))
		return 0;

	nd_region = to_nd_region(dev);
	for (i = 0; i < nd_region->ndr_mappings; i++) {
		nd_mapping  = &nd_region->mapping[i];
		if (nd_mapping->nvdimm == info->nd_mapping->nvdimm)
			break;
	}

	if (i >= nd_region->ndr_mappings)
		return 0;

	ndd = to_ndd(nd_mapping);
	map_end = nd_mapping->start + nd_mapping->size - 1;
	blk_start = nd_mapping->start;

	/*
	 * In the allocation case ->res is set to free space that we are
	 * looking to validate against PMEM aliasing collision rules
	 * (i.e. BLK is allocated after all aliased PMEM).
	 */
	if (info->res) {
		if (info->res->start >= nd_mapping->start
				&& info->res->start < map_end)
			/* pass */;
		else
			return 0;
	}

 retry:
	/*
	 * Find the free dpa from the end of the last pmem allocation to
	 * the end of the interleave-set mapping.
	 */
	for_each_dpa_resource(ndd, res) {
		if (strncmp(res->name, "pmem", 4) != 0)
			continue;
		if ((res->start >= blk_start && res->start < map_end)
				|| (res->end >= blk_start
					&& res->end <= map_end)) {
			new = max(blk_start, min(map_end + 1, res->end + 1));
			if (new != blk_start) {
				blk_start = new;
				goto retry;
			}
		}
	}

	/* update the free space range with the probed blk_start */
	if (info->res && blk_start > info->res->start) {
		info->res->start = max(info->res->start, blk_start);
		if (info->res->start > info->res->end)
			info->res->end = info->res->start - 1;
		return 1;
	}

	info->available -= blk_start - nd_mapping->start;

	return 0;
}

/**
 * nd_blk_available_dpa - account the unused dpa of BLK region
 * @nd_mapping: container of dpa-resource-root + labels
 *
 * Unlike PMEM, BLK namespaces can occupy discontiguous DPA ranges, but
 * we arrange for them to never start at an lower dpa than the last
 * PMEM allocation in an aliased region.
 */
resource_size_t nd_blk_available_dpa(struct nd_region *nd_region)
{
	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(&nd_region->dev);
	struct nd_mapping *nd_mapping = &nd_region->mapping[0];
	struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
	struct blk_alloc_info info = {
		.nd_mapping = nd_mapping,
		.available = nd_mapping->size,
		.res = NULL,
	};
	struct resource *res;

	if (!ndd)
		return 0;

	device_for_each_child(&nvdimm_bus->dev, &info, alias_dpa_busy);

	/* now account for busy blk allocations in unaliased dpa */
	for_each_dpa_resource(ndd, res) {
		if (strncmp(res->name, "blk", 3) != 0)
			continue;
		info.available -= resource_size(res);
	}

	return info.available;
}

/**
 * nd_pmem_available_dpa - for the given dimm+region account unallocated dpa
 * @nd_mapping: container of dpa-resource-root + labels
 * @nd_region: constrain available space check to this reference region
 * @overlap: calculate available space assuming this level of overlap
 *
 * Validate that a PMEM label, if present, aligns with the start of an
 * interleave set and truncate the available size at the lowest BLK
 * overlap point.
 *
 * The expectation is that this routine is called multiple times as it
 * probes for the largest BLK encroachment for any single member DIMM of
 * the interleave set.  Once that value is determined the PMEM-limit for
 * the set can be established.
 */
resource_size_t nd_pmem_available_dpa(struct nd_region *nd_region,
		struct nd_mapping *nd_mapping, resource_size_t *overlap)
{
	resource_size_t map_start, map_end, busy = 0, available, blk_start;
	struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
	struct resource *res;
	const char *reason;

	if (!ndd)
		return 0;

	map_start = nd_mapping->start;
	map_end = map_start + nd_mapping->size - 1;
	blk_start = max(map_start, map_end + 1 - *overlap);
	for_each_dpa_resource(ndd, res) {
		if (res->start >= map_start && res->start < map_end) {
			if (strncmp(res->name, "blk", 3) == 0)
				blk_start = min(blk_start,
						max(map_start, res->start));
			else if (res->end > map_end) {
				reason = "misaligned to iset";
				goto err;
			} else
				busy += resource_size(res);
		} else if (res->end >= map_start && res->end <= map_end) {
			if (strncmp(res->name, "blk", 3) == 0) {
				/*
				 * If a BLK allocation overlaps the start of
				 * PMEM the entire interleave set may now only
				 * be used for BLK.
				 */
				blk_start = map_start;
			} else
				busy += resource_size(res);
		} else if (map_start > res->start && map_start < res->end) {
			/* total eclipse of the mapping */
			busy += nd_mapping->size;
			blk_start = map_start;
		}
	}

	*overlap = map_end + 1 - blk_start;
	available = blk_start - map_start;
	if (busy < available)
		return available - busy;
	return 0;

 err:
	nd_dbg_dpa(nd_region, ndd, res, "%s\n", reason);
	return 0;
}

void nvdimm_free_dpa(struct nvdimm_drvdata *ndd, struct resource *res)
{
	WARN_ON_ONCE(!is_nvdimm_bus_locked(ndd->dev));
	kfree(res->name);
	__release_region(&ndd->dpa, res->start, resource_size(res));
}

struct resource *nvdimm_allocate_dpa(struct nvdimm_drvdata *ndd,
		struct nd_label_id *label_id, resource_size_t start,
		resource_size_t n)
{
	char *name = kmemdup(label_id, sizeof(*label_id), GFP_KERNEL);
	struct resource *res;

	if (!name)
		return NULL;

	WARN_ON_ONCE(!is_nvdimm_bus_locked(ndd->dev));
	res = __request_region(&ndd->dpa, start, n, name, 0);
	if (!res)
		kfree(name);
	return res;
}

/**
 * nvdimm_allocated_dpa - sum up the dpa currently allocated to this label_id
 * @nvdimm: container of dpa-resource-root + labels
 * @label_id: dpa resource name of the form {pmem|blk}-<human readable uuid>
 */
resource_size_t nvdimm_allocated_dpa(struct nvdimm_drvdata *ndd,
		struct nd_label_id *label_id)
{
	resource_size_t allocated = 0;
	struct resource *res;

	for_each_dpa_resource(ndd, res)
		if (strcmp(res->name, label_id->id) == 0)
			allocated += resource_size(res);

	return allocated;
}

static int count_dimms(struct device *dev, void *c)
{
	int *count = c;

	if (is_nvdimm(dev))
		(*count)++;
	return 0;
}

int nvdimm_bus_check_dimm_count(struct nvdimm_bus *nvdimm_bus, int dimm_count)
{
	int count = 0;
	/* Flush any possible dimm registration failures */
	nd_synchronize();

	device_for_each_child(&nvdimm_bus->dev, &count, count_dimms);
	dev_dbg(&nvdimm_bus->dev, "%s: count: %d\n", __func__, count);
	if (count != dimm_count)
		return -ENXIO;
	return 0;
}
EXPORT_SYMBOL_GPL(nvdimm_bus_check_dimm_count);

void __exit nvdimm_devs_exit(void)
{
	ida_destroy(&dimm_ida);
}
Commit	Line	Data
e6dfb2de 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	#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
4d88a97a	14	#include <linux/vmalloc.h>
e6dfb2de	15	#include <linux/device.h>
62232e45	16	#include <linux/ndctl.h>
e6dfb2de DW	17	#include <linux/slab.h>
	18	#include <linux/io.h>
	19	#include <linux/fs.h>
	20	#include <linux/mm.h>
	21	#include "nd-core.h"
0ba1c634	22	#include "label.h"
4d88a97a	23	#include "nd.h"
e6dfb2de DW	24
	25	static DEFINE_IDA(dimm_ida);
	26
4d88a97a DW	27	/*
	28	* Retrieve bus and dimm handle and return if this bus supports
	29	* get_config_data commands
	30	*/
aee65987	31	int nvdimm_check_config_data(struct device *dev)
4d88a97a	32	{
aee65987	33	struct nvdimm *nvdimm = to_nvdimm(dev);
4d88a97a	34
aee65987 TK	35	if (!nvdimm->cmd_mask \|\|
	36	!test_bit(ND_CMD_GET_CONFIG_DATA, &nvdimm->cmd_mask)) {
	37	if (nvdimm->flags & NDD_ALIASING)
	38	return -ENXIO;
	39	else
	40	return -ENOTTY;
	41	}
4d88a97a DW	42
	43	return 0;
	44	}
	45
	46	static int validate_dimm(struct nvdimm_drvdata *ndd)
	47	{
aee65987	48	int rc;
4d88a97a	49
aee65987 TK	50	if (!ndd)
	51	return -EINVAL;
	52
	53	rc = nvdimm_check_config_data(ndd->dev);
	54	if (rc)
4d88a97a DW	55	dev_dbg(ndd->dev, "%pf: %s error: %d\n",
	56	__builtin_return_address(0), __func__, rc);
	57	return rc;
	58	}
	59
	60	/**
	61	* nvdimm_init_nsarea - determine the geometry of a dimm's namespace area
	62	* @nvdimm: dimm to initialize
	63	*/
	64	int nvdimm_init_nsarea(struct nvdimm_drvdata *ndd)
	65	{
	66	struct nd_cmd_get_config_size *cmd = &ndd->nsarea;
	67	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	68	struct nvdimm_bus_descriptor *nd_desc;
	69	int rc = validate_dimm(ndd);
	70
	71	if (rc)
	72	return rc;
	73
	74	if (cmd->config_size)
	75	return 0; /* already valid */
	76
	77	memset(cmd, 0, sizeof(*cmd));
	78	nd_desc = nvdimm_bus->nd_desc;
	79	return nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
aef25338	80	ND_CMD_GET_CONFIG_SIZE, cmd, sizeof(*cmd), NULL);
4d88a97a DW	81	}
	82
	83	int nvdimm_init_config_data(struct nvdimm_drvdata *ndd)
	84	{
	85	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	86	struct nd_cmd_get_config_data_hdr *cmd;
	87	struct nvdimm_bus_descriptor *nd_desc;
	88	int rc = validate_dimm(ndd);
	89	u32 max_cmd_size, config_size;
	90	size_t offset;
	91
	92	if (rc)
	93	return rc;
	94
	95	if (ndd->data)
	96	return 0;
	97
4a826c83 DW	98	if (ndd->nsarea.status \|\| ndd->nsarea.max_xfer == 0
	99	\|\| ndd->nsarea.config_size < ND_LABEL_MIN_SIZE) {
	100	dev_dbg(ndd->dev, "failed to init config data area: (%d:%d)\n",
	101	ndd->nsarea.max_xfer, ndd->nsarea.config_size);
4d88a97a	102	return -ENXIO;
4a826c83	103	}
4d88a97a DW	104
	105	ndd->data = kmalloc(ndd->nsarea.config_size, GFP_KERNEL);
	106	if (!ndd->data)
	107	ndd->data = vmalloc(ndd->nsarea.config_size);
	108
	109	if (!ndd->data)
	110	return -ENOMEM;
	111
	112	max_cmd_size = min_t(u32, PAGE_SIZE, ndd->nsarea.max_xfer);
	113	cmd = kzalloc(max_cmd_size + sizeof(*cmd), GFP_KERNEL);
	114	if (!cmd)
	115	return -ENOMEM;
	116
	117	nd_desc = nvdimm_bus->nd_desc;
	118	for (config_size = ndd->nsarea.config_size, offset = 0;
	119	config_size; config_size -= cmd->in_length,
	120	offset += cmd->in_length) {
	121	cmd->in_length = min(config_size, max_cmd_size);
	122	cmd->in_offset = offset;
	123	rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
	124	ND_CMD_GET_CONFIG_DATA, cmd,
aef25338	125	cmd->in_length + sizeof(*cmd), NULL);
4d88a97a DW	126	if (rc \|\| cmd->status) {
	127	rc = -ENXIO;
	128	break;
	129	}
	130	memcpy(ndd->data + offset, cmd->out_buf, cmd->in_length);
	131	}
	132	dev_dbg(ndd->dev, "%s: len: %zu rc: %d\n", __func__, offset, rc);
	133	kfree(cmd);
	134
	135	return rc;
	136	}
	137
f524bf27 DW	138	int nvdimm_set_config_data(struct nvdimm_drvdata *ndd, size_t offset,
	139	void *buf, size_t len)
	140	{
	141	int rc = validate_dimm(ndd);
	142	size_t max_cmd_size, buf_offset;
	143	struct nd_cmd_set_config_hdr *cmd;
	144	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(ndd->dev);
	145	struct nvdimm_bus_descriptor *nd_desc = nvdimm_bus->nd_desc;
	146
	147	if (rc)
	148	return rc;
	149
	150	if (!ndd->data)
	151	return -ENXIO;
	152
	153	if (offset + len > ndd->nsarea.config_size)
	154	return -ENXIO;
	155
	156	max_cmd_size = min_t(u32, PAGE_SIZE, len);
	157	max_cmd_size = min_t(u32, max_cmd_size, ndd->nsarea.max_xfer);
	158	cmd = kzalloc(max_cmd_size + sizeof(*cmd) + sizeof(u32), GFP_KERNEL);
	159	if (!cmd)
	160	return -ENOMEM;
	161
	162	for (buf_offset = 0; len; len -= cmd->in_length,
	163	buf_offset += cmd->in_length) {
	164	size_t cmd_size;
	165	u32 *status;
	166
	167	cmd->in_offset = offset + buf_offset;
	168	cmd->in_length = min(max_cmd_size, len);
	169	memcpy(cmd->in_buf, buf + buf_offset, cmd->in_length);
	170
	171	/* status is output in the last 4-bytes of the command buffer */
	172	cmd_size = sizeof(*cmd) + cmd->in_length + sizeof(u32);
	173	status = ((void *) cmd) + cmd_size - sizeof(u32);
	174
	175	rc = nd_desc->ndctl(nd_desc, to_nvdimm(ndd->dev),
aef25338	176	ND_CMD_SET_CONFIG_DATA, cmd, cmd_size, NULL);
f524bf27 DW	177	if (rc \|\| *status) {
	178	rc = rc ? rc : -ENXIO;
	179	break;
	180	}
	181	}
	182	kfree(cmd);
	183
	184	return rc;
	185	}
	186
42237e39 DW	187	void nvdimm_set_aliasing(struct device *dev)
	188	{
	189	struct nvdimm *nvdimm = to_nvdimm(dev);
	190
	191	nvdimm->flags \|= NDD_ALIASING;
	192	}
	193
e6dfb2de DW	194	static void nvdimm_release(struct device *dev)
	195	{
	196	struct nvdimm *nvdimm = to_nvdimm(dev);
	197
	198	ida_simple_remove(&dimm_ida, nvdimm->id);
	199	kfree(nvdimm);
	200	}
	201
	202	static struct device_type nvdimm_device_type = {
	203	.name = "nvdimm",
	204	.release = nvdimm_release,
	205	};
	206
62232e45	207	bool is_nvdimm(struct device *dev)
e6dfb2de DW	208	{
	209	return dev->type == &nvdimm_device_type;
	210	}
	211
	212	struct nvdimm to_nvdimm(struct device dev)
	213	{
	214	struct nvdimm *nvdimm = container_of(dev, struct nvdimm, dev);
	215
	216	WARN_ON(!is_nvdimm(dev));
	217	return nvdimm;
	218	}
	219	EXPORT_SYMBOL_GPL(to_nvdimm);
	220
047fc8a1 RZ	221	struct nvdimm nd_blk_region_to_dimm(struct nd_blk_region ndbr)
	222	{
	223	struct nd_region *nd_region = &ndbr->nd_region;
	224	struct nd_mapping *nd_mapping = &nd_region->mapping[0];
	225
	226	return nd_mapping->nvdimm;
	227	}
	228	EXPORT_SYMBOL_GPL(nd_blk_region_to_dimm);
	229
bf9bccc1 DW	230	struct nvdimm_drvdata to_ndd(struct nd_mapping nd_mapping)
	231	{
	232	struct nvdimm *nvdimm = nd_mapping->nvdimm;
	233
	234	WARN_ON_ONCE(!is_nvdimm_bus_locked(&nvdimm->dev));
	235
	236	return dev_get_drvdata(&nvdimm->dev);
	237	}
	238	EXPORT_SYMBOL(to_ndd);
	239
	240	void nvdimm_drvdata_release(struct kref *kref)
	241	{
	242	struct nvdimm_drvdata ndd = container_of(kref, typeof(ndd), kref);
	243	struct device *dev = ndd->dev;
	244	struct resource res, _r;
	245
	246	dev_dbg(dev, "%s\n", __func__);
	247
	248	nvdimm_bus_lock(dev);
	249	for_each_dpa_resource_safe(ndd, res, _r)
	250	nvdimm_free_dpa(ndd, res);
	251	nvdimm_bus_unlock(dev);
	252
a06a7576	253	kvfree(ndd->data);
bf9bccc1 DW	254	kfree(ndd);
	255	put_device(dev);
	256	}
	257
	258	void get_ndd(struct nvdimm_drvdata *ndd)
	259	{
	260	kref_get(&ndd->kref);
	261	}
	262
	263	void put_ndd(struct nvdimm_drvdata *ndd)
	264	{
	265	if (ndd)
	266	kref_put(&ndd->kref, nvdimm_drvdata_release);
	267	}
	268
e6dfb2de DW	269	const char nvdimm_name(struct nvdimm nvdimm)
	270	{
	271	return dev_name(&nvdimm->dev);
	272	}
	273	EXPORT_SYMBOL_GPL(nvdimm_name);
	274
ba9c8dd3 DW	275	struct kobject nvdimm_kobj(struct nvdimm nvdimm)
	276	{
	277	return &nvdimm->dev.kobj;
	278	}
	279	EXPORT_SYMBOL_GPL(nvdimm_kobj);
	280
e3654eca DW	281	unsigned long nvdimm_cmd_mask(struct nvdimm *nvdimm)
	282	{
	283	return nvdimm->cmd_mask;
	284	}
	285	EXPORT_SYMBOL_GPL(nvdimm_cmd_mask);
	286
e6dfb2de DW	287	void nvdimm_provider_data(struct nvdimm nvdimm)
e6dfb2de DW	288	{
62232e45 DW	289	if (nvdimm)
	290	return nvdimm->provider_data;
	291	return NULL;
e6dfb2de DW	292	}
	293	EXPORT_SYMBOL_GPL(nvdimm_provider_data);
	294
62232e45 DW	295	static ssize_t commands_show(struct device *dev,
	296	struct device_attribute attr, char buf)
	297	{
	298	struct nvdimm *nvdimm = to_nvdimm(dev);
	299	int cmd, len = 0;
	300
e3654eca	301	if (!nvdimm->cmd_mask)
62232e45 DW	302	return sprintf(buf, "\n");
62232e45 DW	303
e3654eca	304	for_each_set_bit(cmd, &nvdimm->cmd_mask, BITS_PER_LONG)
62232e45 DW	305	len += sprintf(buf + len, "%s ", nvdimm_cmd_name(cmd));
	306	len += sprintf(buf + len, "\n");
	307	return len;
	308	}
	309	static DEVICE_ATTR_RO(commands);
	310
eaf96153 DW	311	static ssize_t state_show(struct device dev, struct device_attribute attr,
	312	char *buf)
	313	{
	314	struct nvdimm *nvdimm = to_nvdimm(dev);
	315
	316	/*
	317	* The state may be in the process of changing, userspace should
	318	* quiesce probing if it wants a static answer
	319	*/
	320	nvdimm_bus_lock(dev);
	321	nvdimm_bus_unlock(dev);
	322	return sprintf(buf, "%s\n", atomic_read(&nvdimm->busy)
	323	? "active" : "idle");
	324	}
	325	static DEVICE_ATTR_RO(state);
	326
0ba1c634 DW	327	static ssize_t available_slots_show(struct device *dev,
	328	struct device_attribute attr, char buf)
	329	{
	330	struct nvdimm_drvdata *ndd = dev_get_drvdata(dev);
	331	ssize_t rc;
	332	u32 nfree;
	333
	334	if (!ndd)
	335	return -ENXIO;
	336
	337	nvdimm_bus_lock(dev);
	338	nfree = nd_label_nfree(ndd);
	339	if (nfree - 1 > nfree) {
	340	dev_WARN_ONCE(dev, 1, "we ate our last label?\n");
	341	nfree = 0;
	342	} else
	343	nfree--;
	344	rc = sprintf(buf, "%d\n", nfree);
	345	nvdimm_bus_unlock(dev);
	346	return rc;
	347	}
	348	static DEVICE_ATTR_RO(available_slots);
	349
62232e45	350	static struct attribute *nvdimm_attributes[] = {
eaf96153	351	&dev_attr_state.attr,
62232e45	352	&dev_attr_commands.attr,
0ba1c634	353	&dev_attr_available_slots.attr,
62232e45 DW	354	NULL,
	355	};
	356
	357	struct attribute_group nvdimm_attribute_group = {
	358	.attrs = nvdimm_attributes,
	359	};
	360	EXPORT_SYMBOL_GPL(nvdimm_attribute_group);
	361
e6dfb2de	362	struct nvdimm nvdimm_create(struct nvdimm_bus nvdimm_bus, void *provider_data,
62232e45	363	const struct attribute_group **groups, unsigned long flags,
e5ae3b25 DW	364	unsigned long cmd_mask, int num_flush,
e5ae3b25 DW	365	struct resource *flush_wpq)
e6dfb2de DW	366	{
	367	struct nvdimm nvdimm = kzalloc(sizeof(nvdimm), GFP_KERNEL);
	368	struct device *dev;
	369
	370	if (!nvdimm)
	371	return NULL;
	372
	373	nvdimm->id = ida_simple_get(&dimm_ida, 0, 0, GFP_KERNEL);
	374	if (nvdimm->id < 0) {
	375	kfree(nvdimm);
	376	return NULL;
	377	}
	378	nvdimm->provider_data = provider_data;
	379	nvdimm->flags = flags;
e3654eca	380	nvdimm->cmd_mask = cmd_mask;
e5ae3b25 DW	381	nvdimm->num_flush = num_flush;
e5ae3b25 DW	382	nvdimm->flush_wpq = flush_wpq;
eaf96153	383	atomic_set(&nvdimm->busy, 0);
e6dfb2de DW	384	dev = &nvdimm->dev;
	385	dev_set_name(dev, "nmem%d", nvdimm->id);
	386	dev->parent = &nvdimm_bus->dev;
	387	dev->type = &nvdimm_device_type;
62232e45	388	dev->devt = MKDEV(nvdimm_major, nvdimm->id);
e6dfb2de	389	dev->groups = groups;
4d88a97a	390	nd_device_register(dev);
e6dfb2de DW	391
	392	return nvdimm;
	393	}
	394	EXPORT_SYMBOL_GPL(nvdimm_create);
4d88a97a	395
762d067d	396	int alias_dpa_busy(struct device dev, void data)
a1f3e4d6	397	{
fe514739	398	resource_size_t map_end, blk_start, new;
a1f3e4d6 DW	399	struct blk_alloc_info *info = data;
	400	struct nd_mapping *nd_mapping;
	401	struct nd_region *nd_region;
	402	struct nvdimm_drvdata *ndd;
	403	struct resource *res;
	404	int i;
	405
	406	if (!is_nd_pmem(dev))
	407	return 0;
	408
	409	nd_region = to_nd_region(dev);
	410	for (i = 0; i < nd_region->ndr_mappings; i++) {
	411	nd_mapping = &nd_region->mapping[i];
	412	if (nd_mapping->nvdimm == info->nd_mapping->nvdimm)
	413	break;
	414	}
	415
	416	if (i >= nd_region->ndr_mappings)
	417	return 0;
	418
	419	ndd = to_ndd(nd_mapping);
	420	map_end = nd_mapping->start + nd_mapping->size - 1;
	421	blk_start = nd_mapping->start;
762d067d DW	422
	423	/*
	424	* In the allocation case ->res is set to free space that we are
	425	* looking to validate against PMEM aliasing collision rules
	426	* (i.e. BLK is allocated after all aliased PMEM).
	427	*/
	428	if (info->res) {
	429	if (info->res->start >= nd_mapping->start
	430	&& info->res->start < map_end)
	431	/* pass */;
	432	else
	433	return 0;
	434	}
	435
a1f3e4d6 DW	436	retry:
	437	/*
	438	* Find the free dpa from the end of the last pmem allocation to
fe514739	439	* the end of the interleave-set mapping.
a1f3e4d6	440	*/
a1f3e4d6	441	for_each_dpa_resource(ndd, res) {
fe514739 DW	442	if (strncmp(res->name, "pmem", 4) != 0)
fe514739 DW	443	continue;
a1f3e4d6 DW	444	if ((res->start >= blk_start && res->start < map_end)
	445	\|\| (res->end >= blk_start
	446	&& res->end <= map_end)) {
fe514739 DW	447	new = max(blk_start, min(map_end + 1, res->end + 1));
	448	if (new != blk_start) {
	449	blk_start = new;
	450	goto retry;
	451	}
a1f3e4d6 DW	452	}
	453	}
	454
762d067d DW	455	/* update the free space range with the probed blk_start */
	456	if (info->res && blk_start > info->res->start) {
	457	info->res->start = max(info->res->start, blk_start);
	458	if (info->res->start > info->res->end)
	459	info->res->end = info->res->start - 1;
	460	return 1;
	461	}
	462
fe514739	463	info->available -= blk_start - nd_mapping->start;
762d067d	464
a1f3e4d6 DW	465	return 0;
	466	}
	467
1b40e09a DW	468	/**
	469	* nd_blk_available_dpa - account the unused dpa of BLK region
	470	* @nd_mapping: container of dpa-resource-root + labels
	471	*
a1f3e4d6 DW	472	* Unlike PMEM, BLK namespaces can occupy discontiguous DPA ranges, but
	473	* we arrange for them to never start at an lower dpa than the last
	474	* PMEM allocation in an aliased region.
1b40e09a	475	*/
a1f3e4d6	476	resource_size_t nd_blk_available_dpa(struct nd_region *nd_region)
1b40e09a	477	{
a1f3e4d6 DW	478	struct nvdimm_bus *nvdimm_bus = walk_to_nvdimm_bus(&nd_region->dev);
a1f3e4d6 DW	479	struct nd_mapping *nd_mapping = &nd_region->mapping[0];
1b40e09a	480	struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
a1f3e4d6 DW	481	struct blk_alloc_info info = {
	482	.nd_mapping = nd_mapping,
	483	.available = nd_mapping->size,
762d067d	484	.res = NULL,
a1f3e4d6	485	};
1b40e09a DW	486	struct resource *res;
	487
	488	if (!ndd)
	489	return 0;
	490
a1f3e4d6	491	device_for_each_child(&nvdimm_bus->dev, &info, alias_dpa_busy);
1b40e09a	492
a1f3e4d6 DW	493	/* now account for busy blk allocations in unaliased dpa */
	494	for_each_dpa_resource(ndd, res) {
	495	if (strncmp(res->name, "blk", 3) != 0)
	496	continue;
fe514739	497	info.available -= resource_size(res);
a1f3e4d6 DW	498	}
	499
	500	return info.available;
1b40e09a DW	501	}
1b40e09a DW	502
bf9bccc1 DW	503	/**
	504	* nd_pmem_available_dpa - for the given dimm+region account unallocated dpa
	505	* @nd_mapping: container of dpa-resource-root + labels
	506	* @nd_region: constrain available space check to this reference region
	507	* @overlap: calculate available space assuming this level of overlap
	508	*
	509	* Validate that a PMEM label, if present, aligns with the start of an
	510	* interleave set and truncate the available size at the lowest BLK
	511	* overlap point.
	512	*
	513	* The expectation is that this routine is called multiple times as it
	514	* probes for the largest BLK encroachment for any single member DIMM of
	515	* the interleave set. Once that value is determined the PMEM-limit for
	516	* the set can be established.
	517	*/
	518	resource_size_t nd_pmem_available_dpa(struct nd_region *nd_region,
	519	struct nd_mapping nd_mapping, resource_size_t overlap)
	520	{
	521	resource_size_t map_start, map_end, busy = 0, available, blk_start;
	522	struct nvdimm_drvdata *ndd = to_ndd(nd_mapping);
	523	struct resource *res;
	524	const char *reason;
	525
	526	if (!ndd)
	527	return 0;
	528
	529	map_start = nd_mapping->start;
	530	map_end = map_start + nd_mapping->size - 1;
	531	blk_start = max(map_start, map_end + 1 - *overlap);
a1f3e4d6	532	for_each_dpa_resource(ndd, res) {
bf9bccc1 DW	533	if (res->start >= map_start && res->start < map_end) {
bf9bccc1 DW	534	if (strncmp(res->name, "blk", 3) == 0)
a1f3e4d6 DW	535	blk_start = min(blk_start,
	536	max(map_start, res->start));
	537	else if (res->end > map_end) {
bf9bccc1 DW	538	reason = "misaligned to iset";
bf9bccc1 DW	539	goto err;
a1f3e4d6	540	} else
bf9bccc1	541	busy += resource_size(res);
bf9bccc1 DW	542	} else if (res->end >= map_start && res->end <= map_end) {
	543	if (strncmp(res->name, "blk", 3) == 0) {
	544	/*
	545	* If a BLK allocation overlaps the start of
	546	* PMEM the entire interleave set may now only
	547	* be used for BLK.
	548	*/
	549	blk_start = map_start;
a1f3e4d6 DW	550	} else
a1f3e4d6 DW	551	busy += resource_size(res);
bf9bccc1 DW	552	} else if (map_start > res->start && map_start < res->end) {
	553	/* total eclipse of the mapping */
	554	busy += nd_mapping->size;
	555	blk_start = map_start;
	556	}
a1f3e4d6	557	}
bf9bccc1 DW	558
	559	*overlap = map_end + 1 - blk_start;
	560	available = blk_start - map_start;
	561	if (busy < available)
	562	return available - busy;
	563	return 0;
	564
	565	err:
bf9bccc1 DW	566	nd_dbg_dpa(nd_region, ndd, res, "%s\n", reason);
	567	return 0;
	568	}
	569
4a826c83 DW	570	void nvdimm_free_dpa(struct nvdimm_drvdata ndd, struct resource res)
	571	{
	572	WARN_ON_ONCE(!is_nvdimm_bus_locked(ndd->dev));
	573	kfree(res->name);
	574	__release_region(&ndd->dpa, res->start, resource_size(res));
	575	}
	576
	577	struct resource nvdimm_allocate_dpa(struct nvdimm_drvdata ndd,
	578	struct nd_label_id *label_id, resource_size_t start,
	579	resource_size_t n)
	580	{
	581	char name = kmemdup(label_id, sizeof(label_id), GFP_KERNEL);
	582	struct resource *res;
	583
	584	if (!name)
	585	return NULL;
	586
	587	WARN_ON_ONCE(!is_nvdimm_bus_locked(ndd->dev));
	588	res = __request_region(&ndd->dpa, start, n, name, 0);
	589	if (!res)
	590	kfree(name);
	591	return res;
	592	}
	593
bf9bccc1 DW	594	/**
	595	* nvdimm_allocated_dpa - sum up the dpa currently allocated to this label_id
	596	* @nvdimm: container of dpa-resource-root + labels
	597	* @label_id: dpa resource name of the form {pmem\|blk}-<human readable uuid>
	598	*/
	599	resource_size_t nvdimm_allocated_dpa(struct nvdimm_drvdata *ndd,
	600	struct nd_label_id *label_id)
	601	{
	602	resource_size_t allocated = 0;
	603	struct resource *res;
	604
	605	for_each_dpa_resource(ndd, res)
	606	if (strcmp(res->name, label_id->id) == 0)
	607	allocated += resource_size(res);
	608
	609	return allocated;
	610	}
	611
4d88a97a DW	612	static int count_dimms(struct device dev, void c)
	613	{
	614	int *count = c;
	615
	616	if (is_nvdimm(dev))
	617	(*count)++;
	618	return 0;
	619	}
	620
	621	int nvdimm_bus_check_dimm_count(struct nvdimm_bus *nvdimm_bus, int dimm_count)
	622	{
	623	int count = 0;
	624	/* Flush any possible dimm registration failures */
	625	nd_synchronize();
	626
	627	device_for_each_child(&nvdimm_bus->dev, &count, count_dimms);
	628	dev_dbg(&nvdimm_bus->dev, "%s: count: %d\n", __func__, count);
	629	if (count != dimm_count)
	630	return -ENXIO;
	631	return 0;
	632	}
	633	EXPORT_SYMBOL_GPL(nvdimm_bus_check_dimm_count);
b354aba0 DW	634
	635	void __exit nvdimm_devs_exit(void)
	636	{
	637	ida_destroy(&dimm_ida);
	638	}