[mirror_ubuntu-bionic-kernel.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/dax.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);
	if (rc)
		return -EIO;
	return 0;
}

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 pmem_device *pmem, pgoff_t pgoff,
		long nr_pages, void **kaddr, pfn_t *pfn)
{
	resource_size_t offset = PFN_PHYS(pgoff) + pmem->data_offset;

	if (unlikely(is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) / 512,
					PFN_PHYS(nr_pages))))
		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 nr_pages;
	return PHYS_PFN(pmem->size - pmem->pfn_pad - offset);
}

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

static long pmem_dax_direct_access(struct dax_device *dax_dev,
		pgoff_t pgoff, long nr_pages, void **kaddr, pfn_t *pfn)
{
	struct pmem_device *pmem = dax_get_private(dax_dev);

	return __pmem_direct_access(pmem, pgoff, nr_pages, kaddr, pfn);
}

static const struct dax_operations pmem_dax_ops = {
	.direct_access = pmem_dax_direct_access,
};

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

static void pmem_release_disk(void *__pmem)
{
	struct pmem_device *pmem = __pmem;

	kill_dax(pmem->dax_dev);
	put_dax(pmem->dax_dev);
	del_gendisk(pmem->disk);
	put_disk(pmem->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;
	struct dax_device *dax_dev;
	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;
	pmem->disk = disk;

	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;

	dax_dev = alloc_dax(pmem, disk->disk_name, &pmem_dax_ops);
	if (!dax_dev) {
		put_disk(disk);
		return -ENOMEM;
	}
	pmem->dax_dev = dax_dev;

	device_add_disk(dev, disk);
	if (devm_add_action_or_reset(dev, pmem_release_disk, pmem))
		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>
c1d6e828	31	#include <linux/dax.h>
9f53f9fa	32	#include <linux/nd.h>
f295e53b	33	#include "pmem.h"
32ab0a3f	34	#include "pfn.h"
9f53f9fa	35	#include "nd.h"
9e853f23	36
f284a4f2 DW	37	static struct device to_dev(struct pmem_device pmem)
	38	{
	39	/*
	40	* nvdimm bus services need a 'dev' parameter, and we record the device
	41	* at init in bb.dev.
	42	*/
	43	return pmem->bb.dev;
	44	}
	45
	46	static struct nd_region to_region(struct pmem_device pmem)
	47	{
	48	return to_nd_region(to_dev(pmem)->parent);
	49	}
9e853f23	50
3115bb02	51	static int pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset,
59e64739 DW	52	unsigned int len)
59e64739 DW	53	{
f284a4f2	54	struct device *dev = to_dev(pmem);
59e64739 DW	55	sector_t sector;
59e64739 DW	56	long cleared;
868f036f	57	int rc = 0;
59e64739 DW	58
59e64739 DW	59	sector = (offset - pmem->data_offset) / 512;
59e64739	60
868f036f DW	61	cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len);
	62	if (cleared < len)
	63	rc = -EIO;
59e64739	64	if (cleared > 0 && cleared / 512) {
868f036f DW	65	cleared /= 512;
	66	dev_dbg(dev, "%s: %#llx clear %ld sector%s\n", __func__,
	67	(unsigned long long) sector, cleared,
	68	cleared > 1 ? "s" : "");
0a3f27b9	69	badblocks_clear(&pmem->bb, sector, cleared);
59e64739	70	}
3115bb02	71
59e64739	72	invalidate_pmem(pmem->virt_addr + offset, len);
868f036f DW	73
868f036f DW	74	return rc;
59e64739 DW	75	}
59e64739 DW	76
bd697a80 VV	77	static void write_pmem(void pmem_addr, struct page page,
	78	unsigned int off, unsigned int len)
	79	{
	80	void *mem = kmap_atomic(page);
	81
	82	memcpy_to_pmem(pmem_addr, mem + off, len);
	83	kunmap_atomic(mem);
	84	}
	85
	86	static int read_pmem(struct page *page, unsigned int off,
	87	void *pmem_addr, unsigned int len)
	88	{
	89	int rc;
	90	void *mem = kmap_atomic(page);
	91
	92	rc = memcpy_from_pmem(mem + off, pmem_addr, len);
	93	kunmap_atomic(mem);
d47d1d27 SH	94	if (rc)
	95	return -EIO;
	96	return 0;
bd697a80 VV	97	}
bd697a80 VV	98
e10624f8	99	static int pmem_do_bvec(struct pmem_device pmem, struct page page,
c11f0c0b	100	unsigned int len, unsigned int off, bool is_write,
9e853f23 RZ	101	sector_t sector)
9e853f23 RZ	102	{
b5ebc8ec	103	int rc = 0;
59e64739	104	bool bad_pmem = false;
32ab0a3f	105	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
7a9eb206	106	void *pmem_addr = pmem->virt_addr + pmem_off;
9e853f23	107
59e64739 DW	108	if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
	109	bad_pmem = true;
	110
c11f0c0b	111	if (!is_write) {
59e64739	112	if (unlikely(bad_pmem))
b5ebc8ec DW	113	rc = -EIO;
b5ebc8ec DW	114	else {
bd697a80	115	rc = read_pmem(page, off, pmem_addr, len);
b5ebc8ec DW	116	flush_dcache_page(page);
b5ebc8ec DW	117	}
9e853f23	118	} else {
0a370d26 DW	119	/*
	120	* Note that we write the data both before and after
	121	* clearing poison. The write before clear poison
	122	* handles situations where the latest written data is
	123	* preserved and the clear poison operation simply marks
	124	* the address range as valid without changing the data.
	125	* In this case application software can assume that an
	126	* interrupted write will either return the new good
	127	* data or an error.
	128	*
	129	* However, if pmem_clear_poison() leaves the data in an
	130	* indeterminate state we need to perform the write
	131	* after clear poison.
	132	*/
9e853f23	133	flush_dcache_page(page);
bd697a80	134	write_pmem(pmem_addr, page, off, len);
59e64739	135	if (unlikely(bad_pmem)) {
3115bb02	136	rc = pmem_clear_poison(pmem, pmem_off, len);
bd697a80	137	write_pmem(pmem_addr, page, off, len);
59e64739	138	}
9e853f23 RZ	139	}
9e853f23 RZ	140
b5ebc8ec	141	return rc;
9e853f23 RZ	142	}
9e853f23 RZ	143
7e267a8c DW	144	/* account for REQ_FLUSH rename, replace with REQ_PREFLUSH after v4.8-rc1 */
	145	#ifndef REQ_FLUSH
	146	#define REQ_FLUSH REQ_PREFLUSH
	147	#endif
	148
dece1635	149	static blk_qc_t pmem_make_request(struct request_queue q, struct bio bio)
9e853f23	150	{
e10624f8	151	int rc = 0;
f0dc089c DW	152	bool do_acct;
f0dc089c DW	153	unsigned long start;
9e853f23	154	struct bio_vec bvec;
9e853f23	155	struct bvec_iter iter;
bd842b8c	156	struct pmem_device *pmem = q->queuedata;
7e267a8c DW	157	struct nd_region *nd_region = to_region(pmem);
7e267a8c DW	158
1eff9d32	159	if (bio->bi_opf & REQ_FLUSH)
7e267a8c	160	nvdimm_flush(nd_region);
9e853f23	161
f0dc089c	162	do_acct = nd_iostat_start(bio, &start);
e10624f8 DW	163	bio_for_each_segment(bvec, bio, iter) {
e10624f8 DW	164	rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len,
c11f0c0b	165	bvec.bv_offset, op_is_write(bio_op(bio)),
e10624f8 DW	166	iter.bi_sector);
	167	if (rc) {
	168	bio->bi_error = rc;
	169	break;
	170	}
	171	}
f0dc089c DW	172	if (do_acct)
f0dc089c DW	173	nd_iostat_end(bio, start);
61031952	174
1eff9d32	175	if (bio->bi_opf & REQ_FUA)
7e267a8c	176	nvdimm_flush(nd_region);
61031952	177
4246a0b6	178	bio_endio(bio);
dece1635	179	return BLK_QC_T_NONE;
9e853f23 RZ	180	}
	181
	182	static int pmem_rw_page(struct block_device *bdev, sector_t sector,
c11f0c0b	183	struct page *page, bool is_write)
9e853f23	184	{
bd842b8c	185	struct pmem_device *pmem = bdev->bd_queue->queuedata;
e10624f8	186	int rc;
9e853f23	187
c11f0c0b	188	rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, is_write, sector);
9e853f23	189
e10624f8 DW	190	/*
	191	* The ->rw_page interface is subtle and tricky. The core
	192	* retries on any error, so we can only invoke page_endio() in
	193	* the successful completion case. Otherwise, we'll see crashes
	194	* caused by double completion.
	195	*/
	196	if (rc == 0)
c11f0c0b	197	page_endio(page, is_write, 0);
e10624f8 DW	198
e10624f8 DW	199	return rc;
9e853f23 RZ	200	}
9e853f23 RZ	201
f295e53b	202	/* see "strong" declaration in tools/testing/nvdimm/pmem-dax.c */
c1d6e828 DW	203	__weak long __pmem_direct_access(struct pmem_device *pmem, pgoff_t pgoff,
c1d6e828 DW	204	long nr_pages, void *kaddr, pfn_t pfn)
9e853f23	205	{
c1d6e828	206	resource_size_t offset = PFN_PHYS(pgoff) + pmem->data_offset;
589e75d1	207
c1d6e828 DW	208	if (unlikely(is_bad_pmem(&pmem->bb, PFN_PHYS(pgoff) / 512,
c1d6e828 DW	209	PFN_PHYS(nr_pages))))
0a70bd43	210	return -EIO;
e2e05394	211	*kaddr = pmem->virt_addr + offset;
34c0fd54	212	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);
9e853f23	213
0a70bd43 DW	214	/*
	215	* If badblocks are present, limit known good range to the
	216	* requested range.
	217	*/
	218	if (unlikely(pmem->bb.count))
c1d6e828 DW	219	return nr_pages;
	220	return PHYS_PFN(pmem->size - pmem->pfn_pad - offset);
	221	}
	222
9e853f23 RZ	223	static const struct block_device_operations pmem_fops = {
	224	.owner = THIS_MODULE,
	225	.rw_page = pmem_rw_page,
58138820	226	.revalidate_disk = nvdimm_revalidate_disk,
9e853f23 RZ	227	};
9e853f23 RZ	228
c1d6e828 DW	229	static long pmem_dax_direct_access(struct dax_device *dax_dev,
	230	pgoff_t pgoff, long nr_pages, void *kaddr, pfn_t pfn)
	231	{
	232	struct pmem_device *pmem = dax_get_private(dax_dev);
	233
	234	return __pmem_direct_access(pmem, pgoff, nr_pages, kaddr, pfn);
	235	}
	236
	237	static const struct dax_operations pmem_dax_ops = {
	238	.direct_access = pmem_dax_direct_access,
	239	};
	240
030b99e3 DW	241	static void pmem_release_queue(void *q)
	242	{
	243	blk_cleanup_queue(q);
	244	}
	245
c1d6e828	246	static void pmem_release_disk(void *__pmem)
030b99e3	247	{
c1d6e828 DW	248	struct pmem_device *pmem = __pmem;
	249
	250	kill_dax(pmem->dax_dev);
	251	put_dax(pmem->dax_dev);
	252	del_gendisk(pmem->disk);
	253	put_disk(pmem->disk);
030b99e3 DW	254	}
030b99e3 DW	255
200c79da DW	256	static int pmem_attach_disk(struct device *dev,
200c79da DW	257	struct nd_namespace_common *ndns)
9e853f23	258	{
200c79da	259	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
f284a4f2	260	struct nd_region *nd_region = to_nd_region(dev->parent);
200c79da DW	261	struct vmem_altmap __altmap, *altmap = NULL;
	262	struct resource *res = &nsio->res;
	263	struct nd_pfn *nd_pfn = NULL;
c1d6e828	264	struct dax_device *dax_dev;
200c79da DW	265	int nid = dev_to_node(dev);
200c79da DW	266	struct nd_pfn_sb *pfn_sb;
9e853f23	267	struct pmem_device *pmem;
200c79da	268	struct resource pfn_res;
468ded03	269	struct request_queue *q;
200c79da DW	270	struct gendisk *disk;
	271	void *addr;
	272
	273	/* while nsio_rw_bytes is active, parse a pfn info block if present */
	274	if (is_nd_pfn(dev)) {
	275	nd_pfn = to_nd_pfn(dev);
	276	altmap = nvdimm_setup_pfn(nd_pfn, &pfn_res, &__altmap);
	277	if (IS_ERR(altmap))
	278	return PTR_ERR(altmap);
	279	}
	280
	281	/* we're attaching a block device, disable raw namespace access */
	282	devm_nsio_disable(dev, nsio);
9e853f23	283
708ab62b	284	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
9e853f23	285	if (!pmem)
200c79da	286	return -ENOMEM;
9e853f23	287
200c79da	288	dev_set_drvdata(dev, pmem);
9e853f23 RZ	289	pmem->phys_addr = res->start;
9e853f23 RZ	290	pmem->size = resource_size(res);
f284a4f2	291	if (nvdimm_has_flush(nd_region) < 0)
61031952	292	dev_warn(dev, "unable to guarantee persistence of writes\n");
9e853f23	293
947df02d	294	if (!devm_request_mem_region(dev, res->start, resource_size(res),
450c6633	295	dev_name(&ndns->dev))) {
947df02d	296	dev_warn(dev, "could not reserve region %pR\n", res);
200c79da	297	return -EBUSY;
9e853f23 RZ	298	}
9e853f23 RZ	299
468ded03 DW	300	q = blk_alloc_queue_node(GFP_KERNEL, dev_to_node(dev));
468ded03 DW	301	if (!q)
200c79da	302	return -ENOMEM;
468ded03	303
34c0fd54	304	pmem->pfn_flags = PFN_DEV;
200c79da DW	305	if (is_nd_pfn(dev)) {
	306	addr = devm_memremap_pages(dev, &pfn_res, &q->q_usage_counter,
	307	altmap);
	308	pfn_sb = nd_pfn->pfn_sb;
	309	pmem->data_offset = le64_to_cpu(pfn_sb->dataoff);
	310	pmem->pfn_pad = resource_size(res) - resource_size(&pfn_res);
	311	pmem->pfn_flags \|= PFN_MAP;
	312	res = &pfn_res; /* for badblocks populate */
	313	res->start += pmem->data_offset;
	314	} else if (pmem_should_map_pages(dev)) {
	315	addr = devm_memremap_pages(dev, &nsio->res,
5c2c2587	316	&q->q_usage_counter, NULL);
34c0fd54 DW	317	pmem->pfn_flags \|= PFN_MAP;
34c0fd54 DW	318	} else
200c79da DW	319	addr = devm_memremap(dev, pmem->phys_addr,
200c79da DW	320	pmem->size, ARCH_MEMREMAP_PMEM);
b36f4761	321
030b99e3 DW	322	/*
	323	* At release time the queue must be dead before
	324	* devm_memremap_pages is unwound
	325	*/
f02716db	326	if (devm_add_action_or_reset(dev, pmem_release_queue, q))
200c79da	327	return -ENOMEM;
8c2f7e86	328
200c79da DW	329	if (IS_ERR(addr))
200c79da DW	330	return PTR_ERR(addr);
7a9eb206	331	pmem->virt_addr = addr;
9e853f23	332
7e267a8c	333	blk_queue_write_cache(q, true, true);
5a92289f DW	334	blk_queue_make_request(q, pmem_make_request);
	335	blk_queue_physical_block_size(q, PAGE_SIZE);
	336	blk_queue_max_hw_sectors(q, UINT_MAX);
	337	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
	338	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q);
163d4baa	339	queue_flag_set_unlocked(QUEUE_FLAG_DAX, q);
5a92289f	340	q->queuedata = pmem;
9e853f23	341
538ea4aa	342	disk = alloc_disk_node(0, nid);
030b99e3 DW	343	if (!disk)
030b99e3 DW	344	return -ENOMEM;
c1d6e828	345	pmem->disk = disk;
9e853f23	346
9e853f23	347	disk->fops = &pmem_fops;
5a92289f	348	disk->queue = q;
9e853f23	349	disk->flags = GENHD_FL_EXT_DEVT;
5212e11f	350	nvdimm_namespace_disk_name(ndns, disk->disk_name);
cfe30b87 DW	351	set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset)
cfe30b87 DW	352	/ 512);
b95f5f43 DW	353	if (devm_init_badblocks(dev, &pmem->bb))
b95f5f43 DW	354	return -ENOMEM;
f284a4f2	355	nvdimm_badblocks_populate(nd_region, &pmem->bb, res);
57f7f317	356	disk->bb = &pmem->bb;
f02716db	357
c1d6e828 DW	358	dax_dev = alloc_dax(pmem, disk->disk_name, &pmem_dax_ops);
	359	if (!dax_dev) {
	360	put_disk(disk);
	361	return -ENOMEM;
	362	}
	363	pmem->dax_dev = dax_dev;
	364
	365	device_add_disk(dev, disk);
	366	if (devm_add_action_or_reset(dev, pmem_release_disk, pmem))
f02716db DW	367	return -ENOMEM;
f02716db DW	368
58138820	369	revalidate_disk(disk);
9e853f23	370
8c2f7e86 DW	371	return 0;
8c2f7e86 DW	372	}
9e853f23	373
9f53f9fa	374	static int nd_pmem_probe(struct device *dev)
9e853f23	375	{
8c2f7e86	376	struct nd_namespace_common *ndns;
9e853f23	377
8c2f7e86 DW	378	ndns = nvdimm_namespace_common_probe(dev);
	379	if (IS_ERR(ndns))
	380	return PTR_ERR(ndns);
bf9bccc1	381
200c79da DW	382	if (devm_nsio_enable(dev, to_nd_namespace_io(&ndns->dev)))
200c79da DW	383	return -ENXIO;
708ab62b	384
200c79da	385	if (is_nd_btt(dev))
708ab62b CH	386	return nvdimm_namespace_attach_btt(ndns);
708ab62b CH	387
32ab0a3f	388	if (is_nd_pfn(dev))
200c79da	389	return pmem_attach_disk(dev, ndns);
32ab0a3f	390
200c79da	391	/* if we find a valid info-block we'll come back as that personality */
c5ed9268 DW	392	if (nd_btt_probe(dev, ndns) == 0 \|\| nd_pfn_probe(dev, ndns) == 0
c5ed9268 DW	393	\|\| nd_dax_probe(dev, ndns) == 0)
32ab0a3f	394	return -ENXIO;
32ab0a3f	395
200c79da DW	396	/* ...otherwise we're just a raw pmem device */
200c79da DW	397	return pmem_attach_disk(dev, ndns);
9e853f23 RZ	398	}
9e853f23 RZ	399
9f53f9fa	400	static int nd_pmem_remove(struct device *dev)
9e853f23	401	{
8c2f7e86	402	if (is_nd_btt(dev))
298f2bc5	403	nvdimm_namespace_detach_btt(to_nd_btt(dev));
476f848a DW	404	nvdimm_flush(to_nd_region(dev->parent));
476f848a DW	405
9e853f23 RZ	406	return 0;
	407	}
	408
476f848a DW	409	static void nd_pmem_shutdown(struct device *dev)
	410	{
	411	nvdimm_flush(to_nd_region(dev->parent));
	412	}
	413
71999466 DW	414	static void nd_pmem_notify(struct device *dev, enum nvdimm_event event)
71999466 DW	415	{
298f2bc5	416	struct pmem_device *pmem = dev_get_drvdata(dev);
f284a4f2	417	struct nd_region *nd_region = to_region(pmem);
298f2bc5 DW	418	resource_size_t offset = 0, end_trunc = 0;
	419	struct nd_namespace_common *ndns;
	420	struct nd_namespace_io *nsio;
	421	struct resource res;
71999466 DW	422
	423	if (event != NVDIMM_REVALIDATE_POISON)
	424	return;
	425
298f2bc5 DW	426	if (is_nd_btt(dev)) {
	427	struct nd_btt *nd_btt = to_nd_btt(dev);
	428
	429	ndns = nd_btt->ndns;
	430	} else if (is_nd_pfn(dev)) {
a3901802 DW	431	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	432	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	433
298f2bc5 DW	434	ndns = nd_pfn->ndns;
	435	offset = pmem->data_offset + __le32_to_cpu(pfn_sb->start_pad);
	436	end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
	437	} else
	438	ndns = to_ndns(dev);
a3901802	439
298f2bc5 DW	440	nsio = to_nd_namespace_io(&ndns->dev);
	441	res.start = nsio->res.start + offset;
	442	res.end = nsio->res.end - end_trunc;
a3901802	443	nvdimm_badblocks_populate(nd_region, &pmem->bb, &res);
71999466 DW	444	}
71999466 DW	445
9f53f9fa DW	446	MODULE_ALIAS("pmem");
9f53f9fa DW	447	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
bf9bccc1	448	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
9f53f9fa DW	449	static struct nd_device_driver nd_pmem_driver = {
	450	.probe = nd_pmem_probe,
	451	.remove = nd_pmem_remove,
71999466	452	.notify = nd_pmem_notify,
476f848a	453	.shutdown = nd_pmem_shutdown,
9f53f9fa DW	454	.drv = {
9f53f9fa DW	455	.name = "nd_pmem",
9e853f23	456	},
bf9bccc1	457	.type = ND_DRIVER_NAMESPACE_IO \| ND_DRIVER_NAMESPACE_PMEM,
9e853f23 RZ	458	};
	459
	460	static int __init pmem_init(void)
	461	{
55155291	462	return nd_driver_register(&nd_pmem_driver);
9e853f23 RZ	463	}
	464	module_init(pmem_init);
	465
	466	static void pmem_exit(void)
	467	{
9f53f9fa	468	driver_unregister(&nd_pmem_driver.drv);
9e853f23 RZ	469	}
	470	module_exit(pmem_exit);
	471
	472	MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
	473	MODULE_LICENSE("GPL v2");