pmem.c 7.0 KB
Newer Older
1 2 3
/*
 * Persistent Memory Driver
 *
4
 * Copyright (c) 2014-2015, Intel Corporation.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
 * Copyright (c) 2015, Christoph Hellwig <hch@lst.de>.
 * Copyright (c) 2015, Boaz Harrosh <boaz@plexistor.com>.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope 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 <asm/cacheflush.h>
#include <linux/blkdev.h>
#include <linux/hdreg.h>
#include <linux/init.h>
#include <linux/platform_device.h>
#include <linux/module.h>
#include <linux/moduleparam.h>
#include <linux/slab.h>
26
#include <linux/pmem.h>
27 28
#include <linux/nd.h>
#include "nd.h"
29 30 31 32 33 34 35

struct pmem_device {
	struct request_queue	*pmem_queue;
	struct gendisk		*pmem_disk;

	/* One contiguous memory region per device */
	phys_addr_t		phys_addr;
36
	void __pmem		*virt_addr;
37 38 39 40 41 42 43 44 45 46 47
	size_t			size;
};

static int pmem_major;

static void pmem_do_bvec(struct pmem_device *pmem, struct page *page,
			unsigned int len, unsigned int off, int rw,
			sector_t sector)
{
	void *mem = kmap_atomic(page);
	size_t pmem_off = sector << 9;
48
	void __pmem *pmem_addr = pmem->virt_addr + pmem_off;
49 50

	if (rw == READ) {
51
		memcpy_from_pmem(mem + off, pmem_addr, len);
52 53 54
		flush_dcache_page(page);
	} else {
		flush_dcache_page(page);
55
		memcpy_to_pmem(pmem_addr, mem + off, len);
56 57 58 59 60 61 62
	}

	kunmap_atomic(mem);
}

static void pmem_make_request(struct request_queue *q, struct bio *bio)
{
D
Dan Williams 已提交
63 64
	bool do_acct;
	unsigned long start;
65 66
	struct bio_vec bvec;
	struct bvec_iter iter;
D
Dan Williams 已提交
67 68
	struct block_device *bdev = bio->bi_bdev;
	struct pmem_device *pmem = bdev->bd_disk->private_data;
69

D
Dan Williams 已提交
70
	do_acct = nd_iostat_start(bio, &start);
D
Dan Williams 已提交
71
	bio_for_each_segment(bvec, bio, iter)
72
		pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len, bvec.bv_offset,
D
Dan Williams 已提交
73
				bio_data_dir(bio), iter.bi_sector);
D
Dan Williams 已提交
74 75
	if (do_acct)
		nd_iostat_end(bio, start);
76 77 78 79

	if (bio_data_dir(bio))
		wmb_pmem();

D
Dan Williams 已提交
80
	bio_endio(bio, 0);
81 82 83 84 85 86 87 88 89 90 91 92 93 94
}

static int pmem_rw_page(struct block_device *bdev, sector_t sector,
		       struct page *page, int rw)
{
	struct pmem_device *pmem = bdev->bd_disk->private_data;

	pmem_do_bvec(pmem, page, PAGE_CACHE_SIZE, 0, rw, sector);
	page_endio(page, rw & WRITE, 0);

	return 0;
}

static long pmem_direct_access(struct block_device *bdev, sector_t sector,
95
		      void __pmem **kaddr, unsigned long *pfn)
96 97 98 99 100 101 102
{
	struct pmem_device *pmem = bdev->bd_disk->private_data;
	size_t offset = sector << 9;

	if (!pmem)
		return -ENODEV;

103
	/* FIXME convert DAX to comprehend that this mapping has a lifetime */
104
	*kaddr = pmem->virt_addr + offset;
105 106 107 108 109 110 111 112 113
	*pfn = (pmem->phys_addr + offset) >> PAGE_SHIFT;

	return pmem->size - offset;
}

static const struct block_device_operations pmem_fops = {
	.owner =		THIS_MODULE,
	.rw_page =		pmem_rw_page,
	.direct_access =	pmem_direct_access,
114
	.revalidate_disk =	nvdimm_revalidate_disk,
115 116
};

117 118
static struct pmem_device *pmem_alloc(struct device *dev,
		struct resource *res, int id)
119 120 121
{
	struct pmem_device *pmem;

122
	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
123
	if (!pmem)
124
		return ERR_PTR(-ENOMEM);
125 126 127

	pmem->phys_addr = res->start;
	pmem->size = resource_size(res);
128 129
	if (!arch_has_pmem_api())
		dev_warn(dev, "unable to guarantee persistence of writes\n");
130

131 132
	if (!devm_request_mem_region(dev, pmem->phys_addr, pmem->size,
			dev_name(dev))) {
133 134
		dev_warn(dev, "could not reserve region [0x%pa:0x%zx]\n",
				&pmem->phys_addr, pmem->size);
135
		return ERR_PTR(-EBUSY);
136 137
	}

138 139
	pmem->virt_addr = memremap_pmem(dev, pmem->phys_addr, pmem->size);
	if (!pmem->virt_addr)
140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155
		return ERR_PTR(-ENXIO);

	return pmem;
}

static void pmem_detach_disk(struct pmem_device *pmem)
{
	del_gendisk(pmem->pmem_disk);
	put_disk(pmem->pmem_disk);
	blk_cleanup_queue(pmem->pmem_queue);
}

static int pmem_attach_disk(struct nd_namespace_common *ndns,
		struct pmem_device *pmem)
{
	struct gendisk *disk;
156 157 158

	pmem->pmem_queue = blk_alloc_queue(GFP_KERNEL);
	if (!pmem->pmem_queue)
159
		return -ENOMEM;
160 161

	blk_queue_make_request(pmem->pmem_queue, pmem_make_request);
162
	blk_queue_physical_block_size(pmem->pmem_queue, PAGE_SIZE);
163
	blk_queue_max_hw_sectors(pmem->pmem_queue, UINT_MAX);
164
	blk_queue_bounce_limit(pmem->pmem_queue, BLK_BOUNCE_ANY);
165
	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, pmem->pmem_queue);
166

167
	disk = alloc_disk(0);
168 169 170 171
	if (!disk) {
		blk_cleanup_queue(pmem->pmem_queue);
		return -ENOMEM;
	}
172 173

	disk->major		= pmem_major;
174
	disk->first_minor	= 0;
175 176 177 178
	disk->fops		= &pmem_fops;
	disk->private_data	= pmem;
	disk->queue		= pmem->pmem_queue;
	disk->flags		= GENHD_FL_EXT_DEVT;
V
Vishal Verma 已提交
179
	nvdimm_namespace_disk_name(ndns, disk->disk_name);
180
	disk->driverfs_dev = &ndns->dev;
181 182 183 184
	set_capacity(disk, pmem->size >> 9);
	pmem->pmem_disk = disk;

	add_disk(disk);
185
	revalidate_disk(disk);
186

187 188
	return 0;
}
189

190 191 192 193 194 195 196 197 198 199 200
static int pmem_rw_bytes(struct nd_namespace_common *ndns,
		resource_size_t offset, void *buf, size_t size, int rw)
{
	struct pmem_device *pmem = dev_get_drvdata(ndns->claim);

	if (unlikely(offset + size > pmem->size)) {
		dev_WARN_ONCE(&ndns->dev, 1, "request out of range\n");
		return -EFAULT;
	}

	if (rw == READ)
201 202 203 204 205
		memcpy_from_pmem(buf, pmem->virt_addr + offset, size);
	else {
		memcpy_to_pmem(pmem->virt_addr + offset, buf, size);
		wmb_pmem();
	}
206 207 208 209

	return 0;
}

210
static int nd_pmem_probe(struct device *dev)
211
{
212
	struct nd_region *nd_region = to_nd_region(dev->parent);
213 214
	struct nd_namespace_common *ndns;
	struct nd_namespace_io *nsio;
215 216
	struct pmem_device *pmem;

217 218 219
	ndns = nvdimm_namespace_common_probe(dev);
	if (IS_ERR(ndns))
		return PTR_ERR(ndns);
220

221
	nsio = to_nd_namespace_io(&ndns->dev);
222
	pmem = pmem_alloc(dev, &nsio->res, nd_region->id);
223 224 225
	if (IS_ERR(pmem))
		return PTR_ERR(pmem);

226
	dev_set_drvdata(dev, pmem);
227
	ndns->rw_bytes = pmem_rw_bytes;
228

229
	if (is_nd_btt(dev))
230 231 232
		return nvdimm_namespace_attach_btt(ndns);

	if (nd_btt_probe(ndns, pmem) == 0)
233
		/* we'll come back as btt-pmem */
234 235
		return -ENXIO;
	return pmem_attach_disk(ndns, pmem);
236 237
}

238
static int nd_pmem_remove(struct device *dev)
239
{
240
	struct pmem_device *pmem = dev_get_drvdata(dev);
241

242 243 244 245 246
	if (is_nd_btt(dev))
		nvdimm_namespace_detach_btt(to_nd_btt(dev)->ndns);
	else
		pmem_detach_disk(pmem);

247 248 249
	return 0;
}

250 251
MODULE_ALIAS("pmem");
MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
252
MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
253 254 255 256 257
static struct nd_device_driver nd_pmem_driver = {
	.probe = nd_pmem_probe,
	.remove = nd_pmem_remove,
	.drv = {
		.name = "nd_pmem",
258
	},
259
	.type = ND_DRIVER_NAMESPACE_IO | ND_DRIVER_NAMESPACE_PMEM,
260 261 262 263 264 265 266 267 268 269
};

static int __init pmem_init(void)
{
	int error;

	pmem_major = register_blkdev(0, "pmem");
	if (pmem_major < 0)
		return pmem_major;

270 271
	error = nd_driver_register(&nd_pmem_driver);
	if (error) {
272
		unregister_blkdev(pmem_major, "pmem");
273 274 275 276
		return error;
	}

	return 0;
277 278 279 280 281
}
module_init(pmem_init);

static void pmem_exit(void)
{
282
	driver_unregister(&nd_pmem_driver.drv);
283 284 285 286 287 288
	unregister_blkdev(pmem_major, "pmem");
}
module_exit(pmem_exit);

MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
MODULE_LICENSE("GPL v2");