[linux-2.6-block.git] / drivers / nvdimm / pmem.c

/*
 * Persistent Memory Driver
 *
 * Copyright (c) 2014-2015, Intel Corporation.
 * 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/badblocks.h>
#include <linux/memremap.h>
#include <linux/vmalloc.h>
#include <linux/pfn_t.h>
#include <linux/slab.h>
#include <linux/pmem.h>
#include <linux/nd.h>
#include "pmem.h"
#include "pfn.h"
#include "nd.h"

static struct device *to_dev(struct pmem_device *pmem)
{
	/*
	 * nvdimm bus services need a 'dev' parameter, and we record the device
	 * at init in bb.dev.
	 */
	return pmem->bb.dev;
}

static struct nd_region *to_region(struct pmem_device *pmem)
{
	return to_nd_region(to_dev(pmem)->parent);
}

static int pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset,
		unsigned int len)
{
	struct device *dev = to_dev(pmem);
	sector_t sector;
	long cleared;
	int rc = 0;

	sector = (offset - pmem->data_offset) / 512;

	cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len);
	if (cleared < len)
		rc = -EIO;
	if (cleared > 0 && cleared / 512) {
		cleared /= 512;
		dev_dbg(dev, "%s: %#llx clear %ld sector%s\n", __func__,
				(unsigned long long) sector, cleared,
				cleared > 1 ? "s" : "");
		badblocks_clear(&pmem->bb, sector, cleared);
	}

	invalidate_pmem(pmem->virt_addr + offset, len);

	return rc;
}

static void write_pmem(void *pmem_addr, struct page *page,
		unsigned int off, unsigned int len)
{
	void *mem = kmap_atomic(page);

	memcpy_to_pmem(pmem_addr, mem + off, len);
	kunmap_atomic(mem);
}

static int read_pmem(struct page *page, unsigned int off,
		void *pmem_addr, unsigned int len)
{
	int rc;
	void *mem = kmap_atomic(page);

	rc = memcpy_from_pmem(mem + off, pmem_addr, len);
	kunmap_atomic(mem);
	return rc;
}

static int pmem_do_bvec(struct pmem_device *pmem, struct page *page,
			unsigned int len, unsigned int off, bool is_write,
			sector_t sector)
{
	int rc = 0;
	bool bad_pmem = false;
	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
	void *pmem_addr = pmem->virt_addr + pmem_off;

	if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
		bad_pmem = true;

	if (!is_write) {
		if (unlikely(bad_pmem))
			rc = -EIO;
		else {
			rc = read_pmem(page, off, pmem_addr, len);
			flush_dcache_page(page);
		}
	} else {
		/*
		 * Note that we write the data both before and after
		 * clearing poison.  The write before clear poison
		 * handles situations where the latest written data is
		 * preserved and the clear poison operation simply marks
		 * the address range as valid without changing the data.
		 * In this case application software can assume that an
		 * interrupted write will either return the new good
		 * data or an error.
		 *
		 * However, if pmem_clear_poison() leaves the data in an
		 * indeterminate state we need to perform the write
		 * after clear poison.
		 */
		flush_dcache_page(page);
		write_pmem(pmem_addr, page, off, len);
		if (unlikely(bad_pmem)) {
			rc = pmem_clear_poison(pmem, pmem_off, len);
			write_pmem(pmem_addr, page, off, len);
		}
	}

	return rc;
}

/* account for REQ_FLUSH rename, replace with REQ_PREFLUSH after v4.8-rc1 */
#ifndef REQ_FLUSH
#define REQ_FLUSH REQ_PREFLUSH
#endif

static blk_qc_t pmem_make_request(struct request_queue *q, struct bio *bio)
{
	int rc = 0;
	bool do_acct;
	unsigned long start;
	struct bio_vec bvec;
	struct bvec_iter iter;
	struct pmem_device *pmem = q->queuedata;
	struct nd_region *nd_region = to_region(pmem);

	if (bio->bi_opf & REQ_FLUSH)
		nvdimm_flush(nd_region);

	do_acct = nd_iostat_start(bio, &start);
	bio_for_each_segment(bvec, bio, iter) {
		rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len,
				bvec.bv_offset, op_is_write(bio_op(bio)),
				iter.bi_sector);
		if (rc) {
			bio->bi_error = rc;
			break;
		}
	}
	if (do_acct)
		nd_iostat_end(bio, start);

	if (bio->bi_opf & REQ_FUA)
		nvdimm_flush(nd_region);

	bio_endio(bio);
	return BLK_QC_T_NONE;
}

static int pmem_rw_page(struct block_device *bdev, sector_t sector,
		       struct page *page, bool is_write)
{
	struct pmem_device *pmem = bdev->bd_queue->queuedata;
	int rc;

	rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, is_write, sector);

	/*
	 * The ->rw_page interface is subtle and tricky.  The core
	 * retries on any error, so we can only invoke page_endio() in
	 * the successful completion case.  Otherwise, we'll see crashes
	 * caused by double completion.
	 */
	if (rc == 0)
		page_endio(page, is_write, 0);

	return rc;
}

/* see "strong" declaration in tools/testing/nvdimm/pmem-dax.c */
__weak long pmem_direct_access(struct block_device *bdev, sector_t sector,
		      void **kaddr, pfn_t *pfn, long size)
{
	struct pmem_device *pmem = bdev->bd_queue->queuedata;
	resource_size_t offset = sector * 512 + pmem->data_offset;

	if (unlikely(is_bad_pmem(&pmem->bb, sector, size)))
		return -EIO;
	*kaddr = pmem->virt_addr + offset;
	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);

	/*
	 * If badblocks are present, limit known good range to the
	 * requested range.
	 */
	if (unlikely(pmem->bb.count))
		return size;
	return pmem->size - pmem->pfn_pad - offset;
}

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

static void pmem_release_queue(void *q)
{
	blk_cleanup_queue(q);
}

static void pmem_release_disk(void *disk)
{
	del_gendisk(disk);
	put_disk(disk);
}

static int pmem_attach_disk(struct device *dev,
		struct nd_namespace_common *ndns)
{
	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	struct nd_region *nd_region = to_nd_region(dev->parent);
	struct vmem_altmap __altmap, *altmap = NULL;
	struct resource *res = &nsio->res;
	struct nd_pfn *nd_pfn = NULL;
	int nid = dev_to_node(dev);
	struct nd_pfn_sb *pfn_sb;
	struct pmem_device *pmem;
	struct resource pfn_res;
	struct request_queue *q;
	struct gendisk *disk;
	void *addr;

	/* while nsio_rw_bytes is active, parse a pfn info block if present */
	if (is_nd_pfn(dev)) {
		nd_pfn = to_nd_pfn(dev);
		altmap = nvdimm_setup_pfn(nd_pfn, &pfn_res, &__altmap);
		if (IS_ERR(altmap))
			return PTR_ERR(altmap);
	}

	/* we're attaching a block device, disable raw namespace access */
	devm_nsio_disable(dev, nsio);

	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
	if (!pmem)
		return -ENOMEM;

	dev_set_drvdata(dev, pmem);
	pmem->phys_addr = res->start;
	pmem->size = resource_size(res);
	if (nvdimm_has_flush(nd_region) < 0)
		dev_warn(dev, "unable to guarantee persistence of writes\n");

	if (!devm_request_mem_region(dev, res->start, resource_size(res),
				dev_name(&ndns->dev))) {
		dev_warn(dev, "could not reserve region %pR\n", res);
		return -EBUSY;
	}

	q = blk_alloc_queue_node(GFP_KERNEL, dev_to_node(dev));
	if (!q)
		return -ENOMEM;

	pmem->pfn_flags = PFN_DEV;
	if (is_nd_pfn(dev)) {
		addr = devm_memremap_pages(dev, &pfn_res, &q->q_usage_counter,
				altmap);
		pfn_sb = nd_pfn->pfn_sb;
		pmem->data_offset = le64_to_cpu(pfn_sb->dataoff);
		pmem->pfn_pad = resource_size(res) - resource_size(&pfn_res);
		pmem->pfn_flags |= PFN_MAP;
		res = &pfn_res; /* for badblocks populate */
		res->start += pmem->data_offset;
	} else if (pmem_should_map_pages(dev)) {
		addr = devm_memremap_pages(dev, &nsio->res,
				&q->q_usage_counter, NULL);
		pmem->pfn_flags |= PFN_MAP;
	} else
		addr = devm_memremap(dev, pmem->phys_addr,
				pmem->size, ARCH_MEMREMAP_PMEM);

	/*
	 * At release time the queue must be dead before
	 * devm_memremap_pages is unwound
	 */
	if (devm_add_action_or_reset(dev, pmem_release_queue, q))
		return -ENOMEM;

	if (IS_ERR(addr))
		return PTR_ERR(addr);
	pmem->virt_addr = addr;

	blk_queue_write_cache(q, true, true);
	blk_queue_make_request(q, pmem_make_request);
	blk_queue_physical_block_size(q, PAGE_SIZE);
	blk_queue_max_hw_sectors(q, UINT_MAX);
	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q);
	queue_flag_set_unlocked(QUEUE_FLAG_DAX, q);
	q->queuedata = pmem;

	disk = alloc_disk_node(0, nid);
	if (!disk)
		return -ENOMEM;

	disk->fops		= &pmem_fops;
	disk->queue		= q;
	disk->flags		= GENHD_FL_EXT_DEVT;
	nvdimm_namespace_disk_name(ndns, disk->disk_name);
	set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset)
			/ 512);
	if (devm_init_badblocks(dev, &pmem->bb))
		return -ENOMEM;
	nvdimm_badblocks_populate(nd_region, &pmem->bb, res);
	disk->bb = &pmem->bb;
	device_add_disk(dev, disk);

	if (devm_add_action_or_reset(dev, pmem_release_disk, disk))
		return -ENOMEM;

	revalidate_disk(disk);

	return 0;
}

static int nd_pmem_probe(struct device *dev)
{
	struct nd_namespace_common *ndns;

	ndns = nvdimm_namespace_common_probe(dev);
	if (IS_ERR(ndns))
		return PTR_ERR(ndns);

	if (devm_nsio_enable(dev, to_nd_namespace_io(&ndns->dev)))
		return -ENXIO;

	if (is_nd_btt(dev))
		return nvdimm_namespace_attach_btt(ndns);

	if (is_nd_pfn(dev))
		return pmem_attach_disk(dev, ndns);

	/* if we find a valid info-block we'll come back as that personality */
	if (nd_btt_probe(dev, ndns) == 0 || nd_pfn_probe(dev, ndns) == 0
			|| nd_dax_probe(dev, ndns) == 0)
		return -ENXIO;

	/* ...otherwise we're just a raw pmem device */
	return pmem_attach_disk(dev, ndns);
}

static int nd_pmem_remove(struct device *dev)
{
	if (is_nd_btt(dev))
		nvdimm_namespace_detach_btt(to_nd_btt(dev));
	nvdimm_flush(to_nd_region(dev->parent));

	return 0;
}

static void nd_pmem_shutdown(struct device *dev)
{
	nvdimm_flush(to_nd_region(dev->parent));
}

static void nd_pmem_notify(struct device *dev, enum nvdimm_event event)
{
	struct pmem_device *pmem = dev_get_drvdata(dev);
	struct nd_region *nd_region = to_region(pmem);
	resource_size_t offset = 0, end_trunc = 0;
	struct nd_namespace_common *ndns;
	struct nd_namespace_io *nsio;
	struct resource res;

	if (event != NVDIMM_REVALIDATE_POISON)
		return;

	if (is_nd_btt(dev)) {
		struct nd_btt *nd_btt = to_nd_btt(dev);

		ndns = nd_btt->ndns;
	} else if (is_nd_pfn(dev)) {
		struct nd_pfn *nd_pfn = to_nd_pfn(dev);
		struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;

		ndns = nd_pfn->ndns;
		offset = pmem->data_offset + __le32_to_cpu(pfn_sb->start_pad);
		end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
	} else
		ndns = to_ndns(dev);

	nsio = to_nd_namespace_io(&ndns->dev);
	res.start = nsio->res.start + offset;
	res.end = nsio->res.end - end_trunc;
	nvdimm_badblocks_populate(nd_region, &pmem->bb, &res);
}

MODULE_ALIAS("pmem");
MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
static struct nd_device_driver nd_pmem_driver = {
	.probe = nd_pmem_probe,
	.remove = nd_pmem_remove,
	.notify = nd_pmem_notify,
	.shutdown = nd_pmem_shutdown,
	.drv = {
		.name = "nd_pmem",
	},
	.type = ND_DRIVER_NAMESPACE_IO | ND_DRIVER_NAMESPACE_PMEM,
};

static int __init pmem_init(void)
{
	return nd_driver_register(&nd_pmem_driver);
}
module_init(pmem_init);

static void pmem_exit(void)
{
	driver_unregister(&nd_pmem_driver.drv);
}
module_exit(pmem_exit);

MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
MODULE_LICENSE("GPL v2");
Commit	Line	Data
9e853f23 RZ	1	/*
	2	* Persistent Memory Driver
	3	*
9f53f9fa	4	* Copyright (c) 2014-2015, Intel Corporation.
9e853f23 RZ	5	* Copyright (c) 2015, Christoph Hellwig <hch@lst.de>.
	6	* Copyright (c) 2015, Boaz Harrosh <boaz@plexistor.com>.
	7	*
	8	* This program is free software; you can redistribute it and/or modify it
	9	* under the terms and conditions of the GNU General Public License,
	10	* version 2, as published by the Free Software Foundation.
	11	*
	12	* This program is distributed in the hope it will be useful, but WITHOUT
	13	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
	14	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
	15	* more details.
	16	*/
	17
	18	#include <asm/cacheflush.h>
	19	#include <linux/blkdev.h>
	20	#include <linux/hdreg.h>
	21	#include <linux/init.h>
	22	#include <linux/platform_device.h>
	23	#include <linux/module.h>
	24	#include <linux/moduleparam.h>
b95f5f43	25	#include <linux/badblocks.h>
9476df7d	26	#include <linux/memremap.h>
32ab0a3f	27	#include <linux/vmalloc.h>
34c0fd54	28	#include <linux/pfn_t.h>
9e853f23	29	#include <linux/slab.h>
61031952	30	#include <linux/pmem.h>
9f53f9fa	31	#include <linux/nd.h>
f295e53b	32	#include "pmem.h"
32ab0a3f	33	#include "pfn.h"
9f53f9fa	34	#include "nd.h"
9e853f23	35
f284a4f2 DW	36	static struct device to_dev(struct pmem_device pmem)
	37	{
	38	/*
	39	* nvdimm bus services need a 'dev' parameter, and we record the device
	40	* at init in bb.dev.
	41	*/
	42	return pmem->bb.dev;
	43	}
	44
	45	static struct nd_region to_region(struct pmem_device pmem)
	46	{
	47	return to_nd_region(to_dev(pmem)->parent);
	48	}
9e853f23	49
3115bb02	50	static int pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset,
59e64739 DW	51	unsigned int len)
59e64739 DW	52	{
f284a4f2	53	struct device *dev = to_dev(pmem);
59e64739 DW	54	sector_t sector;
59e64739 DW	55	long cleared;
868f036f	56	int rc = 0;
59e64739 DW	57
59e64739 DW	58	sector = (offset - pmem->data_offset) / 512;
59e64739	59
868f036f DW	60	cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len);
	61	if (cleared < len)
	62	rc = -EIO;
59e64739	63	if (cleared > 0 && cleared / 512) {
868f036f DW	64	cleared /= 512;
	65	dev_dbg(dev, "%s: %#llx clear %ld sector%s\n", __func__,
	66	(unsigned long long) sector, cleared,
	67	cleared > 1 ? "s" : "");
0a3f27b9	68	badblocks_clear(&pmem->bb, sector, cleared);
59e64739	69	}
3115bb02	70
59e64739	71	invalidate_pmem(pmem->virt_addr + offset, len);
868f036f DW	72
868f036f DW	73	return rc;
59e64739 DW	74	}
59e64739 DW	75
bd697a80 VV	76	static void write_pmem(void pmem_addr, struct page page,
	77	unsigned int off, unsigned int len)
	78	{
	79	void *mem = kmap_atomic(page);
	80
	81	memcpy_to_pmem(pmem_addr, mem + off, len);
	82	kunmap_atomic(mem);
	83	}
	84
	85	static int read_pmem(struct page *page, unsigned int off,
	86	void *pmem_addr, unsigned int len)
	87	{
	88	int rc;
	89	void *mem = kmap_atomic(page);
	90
	91	rc = memcpy_from_pmem(mem + off, pmem_addr, len);
	92	kunmap_atomic(mem);
	93	return rc;
	94	}
	95
e10624f8	96	static int pmem_do_bvec(struct pmem_device pmem, struct page page,
c11f0c0b	97	unsigned int len, unsigned int off, bool is_write,
9e853f23 RZ	98	sector_t sector)
9e853f23 RZ	99	{
b5ebc8ec	100	int rc = 0;
59e64739	101	bool bad_pmem = false;
32ab0a3f	102	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
7a9eb206	103	void *pmem_addr = pmem->virt_addr + pmem_off;
9e853f23	104
59e64739 DW	105	if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
	106	bad_pmem = true;
	107
c11f0c0b	108	if (!is_write) {
59e64739	109	if (unlikely(bad_pmem))
b5ebc8ec DW	110	rc = -EIO;
b5ebc8ec DW	111	else {
bd697a80	112	rc = read_pmem(page, off, pmem_addr, len);
b5ebc8ec DW	113	flush_dcache_page(page);
b5ebc8ec DW	114	}
9e853f23	115	} else {
0a370d26 DW	116	/*
	117	* Note that we write the data both before and after
	118	* clearing poison. The write before clear poison
	119	* handles situations where the latest written data is
	120	* preserved and the clear poison operation simply marks
	121	* the address range as valid without changing the data.
	122	* In this case application software can assume that an
	123	* interrupted write will either return the new good
	124	* data or an error.
	125	*
	126	* However, if pmem_clear_poison() leaves the data in an
	127	* indeterminate state we need to perform the write
	128	* after clear poison.
	129	*/
9e853f23	130	flush_dcache_page(page);
bd697a80	131	write_pmem(pmem_addr, page, off, len);
59e64739	132	if (unlikely(bad_pmem)) {
3115bb02	133	rc = pmem_clear_poison(pmem, pmem_off, len);
bd697a80	134	write_pmem(pmem_addr, page, off, len);
59e64739	135	}
9e853f23 RZ	136	}
9e853f23 RZ	137
b5ebc8ec	138	return rc;
9e853f23 RZ	139	}
9e853f23 RZ	140
7e267a8c DW	141	/* account for REQ_FLUSH rename, replace with REQ_PREFLUSH after v4.8-rc1 */
	142	#ifndef REQ_FLUSH
	143	#define REQ_FLUSH REQ_PREFLUSH
	144	#endif
	145
dece1635	146	static blk_qc_t pmem_make_request(struct request_queue q, struct bio bio)
9e853f23	147	{
e10624f8	148	int rc = 0;
f0dc089c DW	149	bool do_acct;
f0dc089c DW	150	unsigned long start;
9e853f23	151	struct bio_vec bvec;
9e853f23	152	struct bvec_iter iter;
bd842b8c	153	struct pmem_device *pmem = q->queuedata;
7e267a8c DW	154	struct nd_region *nd_region = to_region(pmem);
7e267a8c DW	155
1eff9d32	156	if (bio->bi_opf & REQ_FLUSH)
7e267a8c	157	nvdimm_flush(nd_region);
9e853f23	158
f0dc089c	159	do_acct = nd_iostat_start(bio, &start);
e10624f8 DW	160	bio_for_each_segment(bvec, bio, iter) {
e10624f8 DW	161	rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len,
c11f0c0b	162	bvec.bv_offset, op_is_write(bio_op(bio)),
e10624f8 DW	163	iter.bi_sector);
	164	if (rc) {
	165	bio->bi_error = rc;
	166	break;
	167	}
	168	}
f0dc089c DW	169	if (do_acct)
f0dc089c DW	170	nd_iostat_end(bio, start);
61031952	171
1eff9d32	172	if (bio->bi_opf & REQ_FUA)
7e267a8c	173	nvdimm_flush(nd_region);
61031952	174
4246a0b6	175	bio_endio(bio);
dece1635	176	return BLK_QC_T_NONE;
9e853f23 RZ	177	}
	178
	179	static int pmem_rw_page(struct block_device *bdev, sector_t sector,
c11f0c0b	180	struct page *page, bool is_write)
9e853f23	181	{
bd842b8c	182	struct pmem_device *pmem = bdev->bd_queue->queuedata;
e10624f8	183	int rc;
9e853f23	184
c11f0c0b	185	rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, is_write, sector);
9e853f23	186
e10624f8 DW	187	/*
	188	* The ->rw_page interface is subtle and tricky. The core
	189	* retries on any error, so we can only invoke page_endio() in
	190	* the successful completion case. Otherwise, we'll see crashes
	191	* caused by double completion.
	192	*/
	193	if (rc == 0)
c11f0c0b	194	page_endio(page, is_write, 0);
e10624f8 DW	195
e10624f8 DW	196	return rc;
9e853f23 RZ	197	}
9e853f23 RZ	198
f295e53b DW	199	/* see "strong" declaration in tools/testing/nvdimm/pmem-dax.c */
f295e53b DW	200	__weak long pmem_direct_access(struct block_device *bdev, sector_t sector,
7a9eb206	201	void *kaddr, pfn_t pfn, long size)
9e853f23	202	{
bd842b8c	203	struct pmem_device *pmem = bdev->bd_queue->queuedata;
32ab0a3f	204	resource_size_t offset = sector * 512 + pmem->data_offset;
589e75d1	205
0a70bd43 DW	206	if (unlikely(is_bad_pmem(&pmem->bb, sector, size)))
0a70bd43 DW	207	return -EIO;
e2e05394	208	*kaddr = pmem->virt_addr + offset;
34c0fd54	209	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);
9e853f23	210
0a70bd43 DW	211	/*
	212	* If badblocks are present, limit known good range to the
	213	* requested range.
	214	*/
	215	if (unlikely(pmem->bb.count))
	216	return size;
cfe30b87	217	return pmem->size - pmem->pfn_pad - offset;
9e853f23 RZ	218	}
	219
	220	static const struct block_device_operations pmem_fops = {
	221	.owner = THIS_MODULE,
	222	.rw_page = pmem_rw_page,
	223	.direct_access = pmem_direct_access,
58138820	224	.revalidate_disk = nvdimm_revalidate_disk,
9e853f23 RZ	225	};
9e853f23 RZ	226
030b99e3 DW	227	static void pmem_release_queue(void *q)
	228	{
	229	blk_cleanup_queue(q);
	230	}
	231
f02716db	232	static void pmem_release_disk(void *disk)
030b99e3 DW	233	{
	234	del_gendisk(disk);
	235	put_disk(disk);
	236	}
	237
200c79da DW	238	static int pmem_attach_disk(struct device *dev,
200c79da DW	239	struct nd_namespace_common *ndns)
9e853f23	240	{
200c79da	241	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
f284a4f2	242	struct nd_region *nd_region = to_nd_region(dev->parent);
200c79da DW	243	struct vmem_altmap __altmap, *altmap = NULL;
	244	struct resource *res = &nsio->res;
	245	struct nd_pfn *nd_pfn = NULL;
	246	int nid = dev_to_node(dev);
	247	struct nd_pfn_sb *pfn_sb;
9e853f23	248	struct pmem_device *pmem;
200c79da	249	struct resource pfn_res;
468ded03	250	struct request_queue *q;
200c79da DW	251	struct gendisk *disk;
	252	void *addr;
	253
	254	/* while nsio_rw_bytes is active, parse a pfn info block if present */
	255	if (is_nd_pfn(dev)) {
	256	nd_pfn = to_nd_pfn(dev);
	257	altmap = nvdimm_setup_pfn(nd_pfn, &pfn_res, &__altmap);
	258	if (IS_ERR(altmap))
	259	return PTR_ERR(altmap);
	260	}
	261
	262	/* we're attaching a block device, disable raw namespace access */
	263	devm_nsio_disable(dev, nsio);
9e853f23	264
708ab62b	265	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
9e853f23	266	if (!pmem)
200c79da	267	return -ENOMEM;
9e853f23	268
200c79da	269	dev_set_drvdata(dev, pmem);
9e853f23 RZ	270	pmem->phys_addr = res->start;
9e853f23 RZ	271	pmem->size = resource_size(res);
f284a4f2	272	if (nvdimm_has_flush(nd_region) < 0)
61031952	273	dev_warn(dev, "unable to guarantee persistence of writes\n");
9e853f23	274
947df02d	275	if (!devm_request_mem_region(dev, res->start, resource_size(res),
450c6633	276	dev_name(&ndns->dev))) {
947df02d	277	dev_warn(dev, "could not reserve region %pR\n", res);
200c79da	278	return -EBUSY;
9e853f23 RZ	279	}
9e853f23 RZ	280
468ded03 DW	281	q = blk_alloc_queue_node(GFP_KERNEL, dev_to_node(dev));
468ded03 DW	282	if (!q)
200c79da	283	return -ENOMEM;
468ded03	284
34c0fd54	285	pmem->pfn_flags = PFN_DEV;
200c79da DW	286	if (is_nd_pfn(dev)) {
	287	addr = devm_memremap_pages(dev, &pfn_res, &q->q_usage_counter,
	288	altmap);
	289	pfn_sb = nd_pfn->pfn_sb;
	290	pmem->data_offset = le64_to_cpu(pfn_sb->dataoff);
	291	pmem->pfn_pad = resource_size(res) - resource_size(&pfn_res);
	292	pmem->pfn_flags \|= PFN_MAP;
	293	res = &pfn_res; /* for badblocks populate */
	294	res->start += pmem->data_offset;
	295	} else if (pmem_should_map_pages(dev)) {
	296	addr = devm_memremap_pages(dev, &nsio->res,
5c2c2587	297	&q->q_usage_counter, NULL);
34c0fd54 DW	298	pmem->pfn_flags \|= PFN_MAP;
34c0fd54 DW	299	} else
200c79da DW	300	addr = devm_memremap(dev, pmem->phys_addr,
200c79da DW	301	pmem->size, ARCH_MEMREMAP_PMEM);
b36f4761	302
030b99e3 DW	303	/*
	304	* At release time the queue must be dead before
	305	* devm_memremap_pages is unwound
	306	*/
f02716db	307	if (devm_add_action_or_reset(dev, pmem_release_queue, q))
200c79da	308	return -ENOMEM;
8c2f7e86	309
200c79da DW	310	if (IS_ERR(addr))
200c79da DW	311	return PTR_ERR(addr);
7a9eb206	312	pmem->virt_addr = addr;
9e853f23	313
7e267a8c	314	blk_queue_write_cache(q, true, true);
5a92289f DW	315	blk_queue_make_request(q, pmem_make_request);
	316	blk_queue_physical_block_size(q, PAGE_SIZE);
	317	blk_queue_max_hw_sectors(q, UINT_MAX);
	318	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
	319	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q);
163d4baa	320	queue_flag_set_unlocked(QUEUE_FLAG_DAX, q);
5a92289f	321	q->queuedata = pmem;
9e853f23	322
538ea4aa	323	disk = alloc_disk_node(0, nid);
030b99e3 DW	324	if (!disk)
030b99e3 DW	325	return -ENOMEM;
9e853f23	326
9e853f23	327	disk->fops = &pmem_fops;
5a92289f	328	disk->queue = q;
9e853f23	329	disk->flags = GENHD_FL_EXT_DEVT;
5212e11f	330	nvdimm_namespace_disk_name(ndns, disk->disk_name);
cfe30b87 DW	331	set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset)
cfe30b87 DW	332	/ 512);
b95f5f43 DW	333	if (devm_init_badblocks(dev, &pmem->bb))
b95f5f43 DW	334	return -ENOMEM;
f284a4f2	335	nvdimm_badblocks_populate(nd_region, &pmem->bb, res);
57f7f317	336	disk->bb = &pmem->bb;
0d52c756	337	device_add_disk(dev, disk);
f02716db DW	338
	339	if (devm_add_action_or_reset(dev, pmem_release_disk, disk))
	340	return -ENOMEM;
	341
58138820	342	revalidate_disk(disk);
9e853f23	343
8c2f7e86 DW	344	return 0;
8c2f7e86 DW	345	}
9e853f23	346
9f53f9fa	347	static int nd_pmem_probe(struct device *dev)
9e853f23	348	{
8c2f7e86	349	struct nd_namespace_common *ndns;
9e853f23	350
8c2f7e86 DW	351	ndns = nvdimm_namespace_common_probe(dev);
	352	if (IS_ERR(ndns))
	353	return PTR_ERR(ndns);
bf9bccc1	354
200c79da DW	355	if (devm_nsio_enable(dev, to_nd_namespace_io(&ndns->dev)))
200c79da DW	356	return -ENXIO;
708ab62b	357
200c79da	358	if (is_nd_btt(dev))
708ab62b CH	359	return nvdimm_namespace_attach_btt(ndns);
708ab62b CH	360
32ab0a3f	361	if (is_nd_pfn(dev))
200c79da	362	return pmem_attach_disk(dev, ndns);
32ab0a3f	363
200c79da	364	/* if we find a valid info-block we'll come back as that personality */
c5ed9268 DW	365	if (nd_btt_probe(dev, ndns) == 0 \|\| nd_pfn_probe(dev, ndns) == 0
c5ed9268 DW	366	\|\| nd_dax_probe(dev, ndns) == 0)
32ab0a3f	367	return -ENXIO;
32ab0a3f	368
200c79da DW	369	/* ...otherwise we're just a raw pmem device */
200c79da DW	370	return pmem_attach_disk(dev, ndns);
9e853f23 RZ	371	}
9e853f23 RZ	372
9f53f9fa	373	static int nd_pmem_remove(struct device *dev)
9e853f23	374	{
8c2f7e86	375	if (is_nd_btt(dev))
298f2bc5	376	nvdimm_namespace_detach_btt(to_nd_btt(dev));
476f848a DW	377	nvdimm_flush(to_nd_region(dev->parent));
476f848a DW	378
9e853f23 RZ	379	return 0;
	380	}
	381
476f848a DW	382	static void nd_pmem_shutdown(struct device *dev)
	383	{
	384	nvdimm_flush(to_nd_region(dev->parent));
	385	}
	386
71999466 DW	387	static void nd_pmem_notify(struct device *dev, enum nvdimm_event event)
71999466 DW	388	{
298f2bc5	389	struct pmem_device *pmem = dev_get_drvdata(dev);
f284a4f2	390	struct nd_region *nd_region = to_region(pmem);
298f2bc5 DW	391	resource_size_t offset = 0, end_trunc = 0;
	392	struct nd_namespace_common *ndns;
	393	struct nd_namespace_io *nsio;
	394	struct resource res;
71999466 DW	395
	396	if (event != NVDIMM_REVALIDATE_POISON)
	397	return;
	398
298f2bc5 DW	399	if (is_nd_btt(dev)) {
	400	struct nd_btt *nd_btt = to_nd_btt(dev);
	401
	402	ndns = nd_btt->ndns;
	403	} else if (is_nd_pfn(dev)) {
a3901802 DW	404	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	405	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	406
298f2bc5 DW	407	ndns = nd_pfn->ndns;
	408	offset = pmem->data_offset + __le32_to_cpu(pfn_sb->start_pad);
	409	end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
	410	} else
	411	ndns = to_ndns(dev);
a3901802	412
298f2bc5 DW	413	nsio = to_nd_namespace_io(&ndns->dev);
	414	res.start = nsio->res.start + offset;
	415	res.end = nsio->res.end - end_trunc;
a3901802	416	nvdimm_badblocks_populate(nd_region, &pmem->bb, &res);
71999466 DW	417	}
71999466 DW	418
9f53f9fa DW	419	MODULE_ALIAS("pmem");
9f53f9fa DW	420	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
bf9bccc1	421	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
9f53f9fa DW	422	static struct nd_device_driver nd_pmem_driver = {
	423	.probe = nd_pmem_probe,
	424	.remove = nd_pmem_remove,
71999466	425	.notify = nd_pmem_notify,
476f848a	426	.shutdown = nd_pmem_shutdown,
9f53f9fa DW	427	.drv = {
9f53f9fa DW	428	.name = "nd_pmem",
9e853f23	429	},
bf9bccc1	430	.type = ND_DRIVER_NAMESPACE_IO \| ND_DRIVER_NAMESPACE_PMEM,
9e853f23 RZ	431	};
	432
	433	static int __init pmem_init(void)
	434	{
55155291	435	return nd_driver_register(&nd_pmem_driver);
9e853f23 RZ	436	}
	437	module_init(pmem_init);
	438
	439	static void pmem_exit(void)
	440	{
9f53f9fa	441	driver_unregister(&nd_pmem_driver.drv);
9e853f23 RZ	442	}
	443	module_exit(pmem_exit);
	444
	445	MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
	446	MODULE_LICENSE("GPL v2");