[mirror_ubuntu-zesty-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/nd.h>
#include "pfn.h"
#include "nd.h"

struct pmem_device {
	/* One contiguous memory region per device */
	phys_addr_t		phys_addr;
	/* when non-zero this device is hosting a 'pfn' instance */
	phys_addr_t		data_offset;
	u64			pfn_flags;
	void __pmem		*virt_addr;
	/* immutable base size of the namespace */
	size_t			size;
	/* trim size when namespace capacity has been section aligned */
	u32			pfn_pad;
	struct badblocks	bb;
};

static void pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset,
		unsigned int len)
{
	struct device *dev = pmem->bb.dev;
	sector_t sector;
	long cleared;

	sector = (offset - pmem->data_offset) / 512;
	cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len);

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

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

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

	if (rw == READ) {
		if (unlikely(bad_pmem))
			rc = -EIO;
		else {
			rc = memcpy_from_pmem(mem + 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);
		memcpy_to_pmem(pmem_addr, mem + off, len);
		if (unlikely(bad_pmem)) {
			pmem_clear_poison(pmem, pmem_off, len);
			memcpy_to_pmem(pmem_addr, mem + off, len);
		}
	}

	kunmap_atomic(mem);
	return rc;
}

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;

	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, bio_data_dir(bio),
				iter.bi_sector);
		if (rc) {
			bio->bi_error = rc;
			break;
		}
	}
	if (do_acct)
		nd_iostat_end(bio, start);

	if (bio_data_dir(bio))
		wmb_pmem();

	bio_endio(bio);
	return BLK_QC_T_NONE;
}

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

	rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, rw, sector);
	if (rw & WRITE)
		wmb_pmem();

	/*
	 * 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, rw & WRITE, 0);

	return rc;
}

static long pmem_direct_access(struct block_device *bdev, sector_t sector,
		      void __pmem **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);
}

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 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 (!arch_has_wmb_pmem())
		dev_warn(dev, "unable to guarantee persistence of writes\n");

	if (!devm_request_mem_region(dev, res->start, resource_size(res),
				dev_name(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(dev, pmem_release_queue, q)) {
		blk_cleanup_queue(q);
		return -ENOMEM;
	}

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

	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);
	q->queuedata = pmem;

	disk = alloc_disk_node(0, nid);
	if (!disk)
		return -ENOMEM;
	if (devm_add_action(dev, pmem_release_disk, disk)) {
		put_disk(disk);
		return -ENOMEM;
	}

	disk->fops		= &pmem_fops;
	disk->queue		= q;
	disk->flags		= GENHD_FL_EXT_DEVT;
	nvdimm_namespace_disk_name(ndns, disk->disk_name);
	disk->driverfs_dev = dev;
	set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset)
			/ 512);
	if (devm_init_badblocks(dev, &pmem->bb))
		return -ENOMEM;
	nvdimm_badblocks_populate(to_nd_region(dev->parent), &pmem->bb, res);
	disk->bb = &pmem->bb;
	add_disk(disk);
	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));
	return 0;
}

static void nd_pmem_notify(struct device *dev, enum nvdimm_event event)
{
	struct nd_region *nd_region = to_nd_region(dev->parent);
	struct pmem_device *pmem = dev_get_drvdata(dev);
	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,
	.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>
32ab0a3f	32	#include "pfn.h"
9f53f9fa	33	#include "nd.h"
9e853f23 RZ	34
9e853f23 RZ	35	struct pmem_device {
9e853f23 RZ	36	/* One contiguous memory region per device */
9e853f23 RZ	37	phys_addr_t phys_addr;
32ab0a3f DW	38	/* when non-zero this device is hosting a 'pfn' instance */
32ab0a3f DW	39	phys_addr_t data_offset;
c4544205	40	u64 pfn_flags;
61031952	41	void __pmem *virt_addr;
cfe30b87	42	/* immutable base size of the namespace */
9e853f23	43	size_t size;
cfe30b87 DW	44	/* trim size when namespace capacity has been section aligned */
cfe30b87 DW	45	u32 pfn_pad;
b95f5f43	46	struct badblocks bb;
9e853f23 RZ	47	};
9e853f23 RZ	48
59e64739 DW	49	static void pmem_clear_poison(struct pmem_device *pmem, phys_addr_t offset,
	50	unsigned int len)
	51	{
5a92289f	52	struct device *dev = pmem->bb.dev;
59e64739 DW	53	sector_t sector;
	54	long cleared;
	55
	56	sector = (offset - pmem->data_offset) / 512;
	57	cleared = nvdimm_clear_poison(dev, pmem->phys_addr + offset, len);
	58
	59	if (cleared > 0 && cleared / 512) {
	60	dev_dbg(dev, "%s: %llx clear %ld sector%s\n",
	61	__func__, (unsigned long long) sector,
	62	cleared / 512, cleared / 512 > 1 ? "s" : "");
	63	badblocks_clear(&pmem->bb, sector, cleared / 512);
	64	}
	65	invalidate_pmem(pmem->virt_addr + offset, len);
	66	}
	67
e10624f8	68	static int pmem_do_bvec(struct pmem_device pmem, struct page page,
9e853f23 RZ	69	unsigned int len, unsigned int off, int rw,
	70	sector_t sector)
	71	{
b5ebc8ec	72	int rc = 0;
59e64739	73	bool bad_pmem = false;
9e853f23	74	void *mem = kmap_atomic(page);
32ab0a3f	75	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
61031952	76	void __pmem *pmem_addr = pmem->virt_addr + pmem_off;
9e853f23	77
59e64739 DW	78	if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
	79	bad_pmem = true;
	80
9e853f23	81	if (rw == READ) {
59e64739	82	if (unlikely(bad_pmem))
b5ebc8ec DW	83	rc = -EIO;
b5ebc8ec DW	84	else {
fc0c2028	85	rc = memcpy_from_pmem(mem + off, pmem_addr, len);
b5ebc8ec DW	86	flush_dcache_page(page);
b5ebc8ec DW	87	}
9e853f23	88	} else {
0a370d26 DW	89	/*
	90	* Note that we write the data both before and after
	91	* clearing poison. The write before clear poison
	92	* handles situations where the latest written data is
	93	* preserved and the clear poison operation simply marks
	94	* the address range as valid without changing the data.
	95	* In this case application software can assume that an
	96	* interrupted write will either return the new good
	97	* data or an error.
	98	*
	99	* However, if pmem_clear_poison() leaves the data in an
	100	* indeterminate state we need to perform the write
	101	* after clear poison.
	102	*/
9e853f23	103	flush_dcache_page(page);
61031952	104	memcpy_to_pmem(pmem_addr, mem + off, len);
59e64739 DW	105	if (unlikely(bad_pmem)) {
	106	pmem_clear_poison(pmem, pmem_off, len);
	107	memcpy_to_pmem(pmem_addr, mem + off, len);
	108	}
9e853f23 RZ	109	}
	110
	111	kunmap_atomic(mem);
b5ebc8ec	112	return rc;
9e853f23 RZ	113	}
9e853f23 RZ	114
dece1635	115	static blk_qc_t pmem_make_request(struct request_queue q, struct bio bio)
9e853f23	116	{
e10624f8	117	int rc = 0;
f0dc089c DW	118	bool do_acct;
f0dc089c DW	119	unsigned long start;
9e853f23	120	struct bio_vec bvec;
9e853f23	121	struct bvec_iter iter;
bd842b8c	122	struct pmem_device *pmem = q->queuedata;
9e853f23	123
f0dc089c	124	do_acct = nd_iostat_start(bio, &start);
e10624f8 DW	125	bio_for_each_segment(bvec, bio, iter) {
	126	rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len,
	127	bvec.bv_offset, bio_data_dir(bio),
	128	iter.bi_sector);
	129	if (rc) {
	130	bio->bi_error = rc;
	131	break;
	132	}
	133	}
f0dc089c DW	134	if (do_acct)
f0dc089c DW	135	nd_iostat_end(bio, start);
61031952 RZ	136
	137	if (bio_data_dir(bio))
	138	wmb_pmem();
	139
4246a0b6	140	bio_endio(bio);
dece1635	141	return BLK_QC_T_NONE;
9e853f23 RZ	142	}
	143
	144	static int pmem_rw_page(struct block_device *bdev, sector_t sector,
	145	struct page *page, int rw)
	146	{
bd842b8c	147	struct pmem_device *pmem = bdev->bd_queue->queuedata;
e10624f8	148	int rc;
9e853f23	149
09cbfeaf	150	rc = pmem_do_bvec(pmem, page, PAGE_SIZE, 0, rw, sector);
ba8fe0f8 RZ	151	if (rw & WRITE)
ba8fe0f8 RZ	152	wmb_pmem();
9e853f23	153
e10624f8 DW	154	/*
	155	* The ->rw_page interface is subtle and tricky. The core
	156	* retries on any error, so we can only invoke page_endio() in
	157	* the successful completion case. Otherwise, we'll see crashes
	158	* caused by double completion.
	159	*/
	160	if (rc == 0)
	161	page_endio(page, rw & WRITE, 0);
	162
	163	return rc;
9e853f23 RZ	164	}
	165
	166	static long pmem_direct_access(struct block_device *bdev, sector_t sector,
0a70bd43	167	void __pmem *kaddr, pfn_t pfn, long size)
9e853f23	168	{
bd842b8c	169	struct pmem_device *pmem = bdev->bd_queue->queuedata;
32ab0a3f	170	resource_size_t offset = sector * 512 + pmem->data_offset;
589e75d1	171
0a70bd43 DW	172	if (unlikely(is_bad_pmem(&pmem->bb, sector, size)))
0a70bd43 DW	173	return -EIO;
e2e05394	174	*kaddr = pmem->virt_addr + offset;
34c0fd54	175	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);
9e853f23	176
0a70bd43 DW	177	/*
	178	* If badblocks are present, limit known good range to the
	179	* requested range.
	180	*/
	181	if (unlikely(pmem->bb.count))
	182	return size;
cfe30b87	183	return pmem->size - pmem->pfn_pad - offset;
9e853f23 RZ	184	}
	185
	186	static const struct block_device_operations pmem_fops = {
	187	.owner = THIS_MODULE,
	188	.rw_page = pmem_rw_page,
	189	.direct_access = pmem_direct_access,
58138820	190	.revalidate_disk = nvdimm_revalidate_disk,
9e853f23 RZ	191	};
9e853f23 RZ	192
030b99e3 DW	193	static void pmem_release_queue(void *q)
	194	{
	195	blk_cleanup_queue(q);
	196	}
	197
	198	void pmem_release_disk(void *disk)
	199	{
	200	del_gendisk(disk);
	201	put_disk(disk);
	202	}
	203
200c79da DW	204	static int pmem_attach_disk(struct device *dev,
200c79da DW	205	struct nd_namespace_common *ndns)
9e853f23	206	{
200c79da DW	207	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	208	struct vmem_altmap __altmap, *altmap = NULL;
	209	struct resource *res = &nsio->res;
	210	struct nd_pfn *nd_pfn = NULL;
	211	int nid = dev_to_node(dev);
	212	struct nd_pfn_sb *pfn_sb;
9e853f23	213	struct pmem_device *pmem;
200c79da	214	struct resource pfn_res;
468ded03	215	struct request_queue *q;
200c79da DW	216	struct gendisk *disk;
	217	void *addr;
	218
	219	/* while nsio_rw_bytes is active, parse a pfn info block if present */
	220	if (is_nd_pfn(dev)) {
	221	nd_pfn = to_nd_pfn(dev);
	222	altmap = nvdimm_setup_pfn(nd_pfn, &pfn_res, &__altmap);
	223	if (IS_ERR(altmap))
	224	return PTR_ERR(altmap);
	225	}
	226
	227	/* we're attaching a block device, disable raw namespace access */
	228	devm_nsio_disable(dev, nsio);
9e853f23	229
708ab62b	230	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
9e853f23	231	if (!pmem)
200c79da	232	return -ENOMEM;
9e853f23	233
200c79da	234	dev_set_drvdata(dev, pmem);
9e853f23 RZ	235	pmem->phys_addr = res->start;
9e853f23 RZ	236	pmem->size = resource_size(res);
96601adb	237	if (!arch_has_wmb_pmem())
61031952	238	dev_warn(dev, "unable to guarantee persistence of writes\n");
9e853f23	239
947df02d DW	240	if (!devm_request_mem_region(dev, res->start, resource_size(res),
	241	dev_name(dev))) {
	242	dev_warn(dev, "could not reserve region %pR\n", res);
200c79da	243	return -EBUSY;
9e853f23 RZ	244	}
9e853f23 RZ	245
468ded03 DW	246	q = blk_alloc_queue_node(GFP_KERNEL, dev_to_node(dev));
468ded03 DW	247	if (!q)
200c79da	248	return -ENOMEM;
468ded03	249
34c0fd54	250	pmem->pfn_flags = PFN_DEV;
200c79da DW	251	if (is_nd_pfn(dev)) {
	252	addr = devm_memremap_pages(dev, &pfn_res, &q->q_usage_counter,
	253	altmap);
	254	pfn_sb = nd_pfn->pfn_sb;
	255	pmem->data_offset = le64_to_cpu(pfn_sb->dataoff);
	256	pmem->pfn_pad = resource_size(res) - resource_size(&pfn_res);
	257	pmem->pfn_flags \|= PFN_MAP;
	258	res = &pfn_res; /* for badblocks populate */
	259	res->start += pmem->data_offset;
	260	} else if (pmem_should_map_pages(dev)) {
	261	addr = devm_memremap_pages(dev, &nsio->res,
5c2c2587	262	&q->q_usage_counter, NULL);
34c0fd54 DW	263	pmem->pfn_flags \|= PFN_MAP;
34c0fd54 DW	264	} else
200c79da DW	265	addr = devm_memremap(dev, pmem->phys_addr,
200c79da DW	266	pmem->size, ARCH_MEMREMAP_PMEM);
b36f4761	267
030b99e3 DW	268	/*
	269	* At release time the queue must be dead before
	270	* devm_memremap_pages is unwound
	271	*/
	272	if (devm_add_action(dev, pmem_release_queue, q)) {
468ded03	273	blk_cleanup_queue(q);
200c79da	274	return -ENOMEM;
468ded03	275	}
8c2f7e86	276
200c79da DW	277	if (IS_ERR(addr))
	278	return PTR_ERR(addr);
	279	pmem->virt_addr = (void __pmem *) addr;
9e853f23	280
5a92289f DW	281	blk_queue_make_request(q, pmem_make_request);
	282	blk_queue_physical_block_size(q, PAGE_SIZE);
	283	blk_queue_max_hw_sectors(q, UINT_MAX);
	284	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
	285	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, q);
	286	q->queuedata = pmem;
9e853f23	287
538ea4aa	288	disk = alloc_disk_node(0, nid);
030b99e3 DW	289	if (!disk)
	290	return -ENOMEM;
	291	if (devm_add_action(dev, pmem_release_disk, disk)) {
	292	put_disk(disk);
8c2f7e86 DW	293	return -ENOMEM;
8c2f7e86 DW	294	}
9e853f23	295
9e853f23	296	disk->fops = &pmem_fops;
5a92289f	297	disk->queue = q;
9e853f23	298	disk->flags = GENHD_FL_EXT_DEVT;
5212e11f	299	nvdimm_namespace_disk_name(ndns, disk->disk_name);
32ab0a3f	300	disk->driverfs_dev = dev;
cfe30b87 DW	301	set_capacity(disk, (pmem->size - pmem->pfn_pad - pmem->data_offset)
cfe30b87 DW	302	/ 512);
b95f5f43 DW	303	if (devm_init_badblocks(dev, &pmem->bb))
b95f5f43 DW	304	return -ENOMEM;
200c79da	305	nvdimm_badblocks_populate(to_nd_region(dev->parent), &pmem->bb, res);
57f7f317	306	disk->bb = &pmem->bb;
9e853f23	307	add_disk(disk);
58138820	308	revalidate_disk(disk);
9e853f23	309
8c2f7e86 DW	310	return 0;
8c2f7e86 DW	311	}
9e853f23	312
9f53f9fa	313	static int nd_pmem_probe(struct device *dev)
9e853f23	314	{
8c2f7e86	315	struct nd_namespace_common *ndns;
9e853f23	316
8c2f7e86 DW	317	ndns = nvdimm_namespace_common_probe(dev);
	318	if (IS_ERR(ndns))
	319	return PTR_ERR(ndns);
bf9bccc1	320
200c79da DW	321	if (devm_nsio_enable(dev, to_nd_namespace_io(&ndns->dev)))
200c79da DW	322	return -ENXIO;
708ab62b	323
200c79da	324	if (is_nd_btt(dev))
708ab62b CH	325	return nvdimm_namespace_attach_btt(ndns);
708ab62b CH	326
32ab0a3f	327	if (is_nd_pfn(dev))
200c79da	328	return pmem_attach_disk(dev, ndns);
32ab0a3f	329
200c79da	330	/* if we find a valid info-block we'll come back as that personality */
c5ed9268 DW	331	if (nd_btt_probe(dev, ndns) == 0 \|\| nd_pfn_probe(dev, ndns) == 0
c5ed9268 DW	332	\|\| nd_dax_probe(dev, ndns) == 0)
32ab0a3f	333	return -ENXIO;
32ab0a3f	334
200c79da DW	335	/* ...otherwise we're just a raw pmem device */
200c79da DW	336	return pmem_attach_disk(dev, ndns);
9e853f23 RZ	337	}
9e853f23 RZ	338
9f53f9fa	339	static int nd_pmem_remove(struct device *dev)
9e853f23	340	{
8c2f7e86	341	if (is_nd_btt(dev))
298f2bc5	342	nvdimm_namespace_detach_btt(to_nd_btt(dev));
9e853f23 RZ	343	return 0;
	344	}
	345
71999466 DW	346	static void nd_pmem_notify(struct device *dev, enum nvdimm_event event)
71999466 DW	347	{
a3901802	348	struct nd_region *nd_region = to_nd_region(dev->parent);
298f2bc5 DW	349	struct pmem_device *pmem = dev_get_drvdata(dev);
	350	resource_size_t offset = 0, end_trunc = 0;
	351	struct nd_namespace_common *ndns;
	352	struct nd_namespace_io *nsio;
	353	struct resource res;
71999466 DW	354
	355	if (event != NVDIMM_REVALIDATE_POISON)
	356	return;
	357
298f2bc5 DW	358	if (is_nd_btt(dev)) {
	359	struct nd_btt *nd_btt = to_nd_btt(dev);
	360
	361	ndns = nd_btt->ndns;
	362	} else if (is_nd_pfn(dev)) {
a3901802 DW	363	struct nd_pfn *nd_pfn = to_nd_pfn(dev);
	364	struct nd_pfn_sb *pfn_sb = nd_pfn->pfn_sb;
	365
298f2bc5 DW	366	ndns = nd_pfn->ndns;
	367	offset = pmem->data_offset + __le32_to_cpu(pfn_sb->start_pad);
	368	end_trunc = __le32_to_cpu(pfn_sb->end_trunc);
	369	} else
	370	ndns = to_ndns(dev);
a3901802	371
298f2bc5 DW	372	nsio = to_nd_namespace_io(&ndns->dev);
	373	res.start = nsio->res.start + offset;
	374	res.end = nsio->res.end - end_trunc;
a3901802	375	nvdimm_badblocks_populate(nd_region, &pmem->bb, &res);
71999466 DW	376	}
71999466 DW	377
9f53f9fa DW	378	MODULE_ALIAS("pmem");
9f53f9fa DW	379	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
bf9bccc1	380	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
9f53f9fa DW	381	static struct nd_device_driver nd_pmem_driver = {
	382	.probe = nd_pmem_probe,
	383	.remove = nd_pmem_remove,
71999466	384	.notify = nd_pmem_notify,
9f53f9fa DW	385	.drv = {
9f53f9fa DW	386	.name = "nd_pmem",
9e853f23	387	},
bf9bccc1	388	.type = ND_DRIVER_NAMESPACE_IO \| ND_DRIVER_NAMESPACE_PMEM,
9e853f23 RZ	389	};
	390
	391	static int __init pmem_init(void)
	392	{
55155291	393	return nd_driver_register(&nd_pmem_driver);
9e853f23 RZ	394	}
	395	module_init(pmem_init);
	396
	397	static void pmem_exit(void)
	398	{
9f53f9fa	399	driver_unregister(&nd_pmem_driver.drv);
9e853f23 RZ	400	}
	401	module_exit(pmem_exit);
	402
	403	MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
	404	MODULE_LICENSE("GPL v2");