[deliverable/linux.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 {
	struct request_queue	*pmem_queue;
	struct gendisk		*pmem_disk;
	struct nd_namespace_common *ndns;

	/* 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;
	unsigned long		pfn_flags;
	void __pmem		*virt_addr;
	size_t			size;
	struct badblocks	bb;
};

static int pmem_major;

static bool is_bad_pmem(struct badblocks *bb, sector_t sector, unsigned int len)
{
	if (bb->count) {
		sector_t first_bad;
		int num_bad;

		return !!badblocks_check(bb, sector, len / 512, &first_bad,
				&num_bad);
	}

	return false;
}

static int 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);
	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
	void __pmem *pmem_addr = pmem->virt_addr + pmem_off;

	if (rw == READ) {
		if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
			return -EIO;
		memcpy_from_pmem(mem + off, pmem_addr, len);
		flush_dcache_page(page);
	} else {
		flush_dcache_page(page);
		memcpy_to_pmem(pmem_addr, mem + off, len);
	}

	kunmap_atomic(mem);
	return 0;
}

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 block_device *bdev = bio->bi_bdev;
	struct pmem_device *pmem = bdev->bd_disk->private_data;

	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_disk->private_data;
	int rc;

	rc = pmem_do_bvec(pmem, page, PAGE_CACHE_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)
{
	struct pmem_device *pmem = bdev->bd_disk->private_data;
	resource_size_t offset = sector * 512 + pmem->data_offset;

	*kaddr = pmem->virt_addr + offset;
	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);

	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,
	.revalidate_disk =	nvdimm_revalidate_disk,
};

static struct pmem_device *pmem_alloc(struct device *dev,
		struct resource *res, int id)
{
	struct pmem_device *pmem;

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

	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, pmem->phys_addr, pmem->size,
			dev_name(dev))) {
		dev_warn(dev, "could not reserve region [0x%pa:0x%zx]\n",
				&pmem->phys_addr, pmem->size);
		return ERR_PTR(-EBUSY);
	}

	pmem->pfn_flags = PFN_DEV;
	if (pmem_should_map_pages(dev)) {
		pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, res);
		pmem->pfn_flags |= PFN_MAP;
	} else
		pmem->virt_addr = (void __pmem *) devm_memremap(dev,
				pmem->phys_addr, pmem->size,
				ARCH_MEMREMAP_PMEM);

	if (IS_ERR(pmem->virt_addr))
		return (void __force *) pmem->virt_addr;

	return pmem;
}

static void pmem_detach_disk(struct pmem_device *pmem)
{
	if (!pmem->pmem_disk)
		return;

	del_gendisk(pmem->pmem_disk);
	put_disk(pmem->pmem_disk);
	blk_cleanup_queue(pmem->pmem_queue);
}

static int pmem_attach_disk(struct device *dev,
		struct nd_namespace_common *ndns, struct pmem_device *pmem)
{
	int nid = dev_to_node(dev);
	struct gendisk *disk;

	pmem->pmem_queue = blk_alloc_queue_node(GFP_KERNEL, nid);
	if (!pmem->pmem_queue)
		return -ENOMEM;

	blk_queue_make_request(pmem->pmem_queue, pmem_make_request);
	blk_queue_physical_block_size(pmem->pmem_queue, PAGE_SIZE);
	blk_queue_max_hw_sectors(pmem->pmem_queue, UINT_MAX);
	blk_queue_bounce_limit(pmem->pmem_queue, BLK_BOUNCE_ANY);
	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, pmem->pmem_queue);

	disk = alloc_disk_node(0, nid);
	if (!disk) {
		blk_cleanup_queue(pmem->pmem_queue);
		return -ENOMEM;
	}

	disk->major		= pmem_major;
	disk->first_minor	= 0;
	disk->fops		= &pmem_fops;
	disk->private_data	= pmem;
	disk->queue		= pmem->pmem_queue;
	disk->flags		= GENHD_FL_EXT_DEVT;
	nvdimm_namespace_disk_name(ndns, disk->disk_name);
	disk->driverfs_dev = dev;
	set_capacity(disk, (pmem->size - pmem->data_offset) / 512);
	pmem->pmem_disk = disk;
	devm_exit_badblocks(dev, &pmem->bb);
	if (devm_init_badblocks(dev, &pmem->bb))
		return -ENOMEM;
	nvdimm_namespace_add_poison(ndns, &pmem->bb, pmem->data_offset);

	disk->bb = &pmem->bb;
	add_disk(disk);
	revalidate_disk(disk);

	return 0;
}

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) {
		unsigned int sz_align = ALIGN(size + (offset & (512 - 1)), 512);

		if (unlikely(is_bad_pmem(&pmem->bb, offset / 512, sz_align)))
			return -EIO;
		memcpy_from_pmem(buf, pmem->virt_addr + offset, size);
	} else {
		memcpy_to_pmem(pmem->virt_addr + offset, buf, size);
		wmb_pmem();
	}

	return 0;
}

static int nd_pfn_init(struct nd_pfn *nd_pfn)
{
	struct nd_pfn_sb *pfn_sb = kzalloc(sizeof(*pfn_sb), GFP_KERNEL);
	struct pmem_device *pmem = dev_get_drvdata(&nd_pfn->dev);
	struct nd_namespace_common *ndns = nd_pfn->ndns;
	struct nd_region *nd_region;
	unsigned long npfns;
	phys_addr_t offset;
	u64 checksum;
	int rc;

	if (!pfn_sb)
		return -ENOMEM;

	nd_pfn->pfn_sb = pfn_sb;
	rc = nd_pfn_validate(nd_pfn);
	if (rc == -ENODEV)
		/* no info block, do init */;
	else
		return rc;

	nd_region = to_nd_region(nd_pfn->dev.parent);
	if (nd_region->ro) {
		dev_info(&nd_pfn->dev,
				"%s is read-only, unable to init metadata\n",
				dev_name(&nd_region->dev));
		goto err;
	}

	memset(pfn_sb, 0, sizeof(*pfn_sb));
	npfns = (pmem->size - SZ_8K) / SZ_4K;
	/*
	 * Note, we use 64 here for the standard size of struct page,
	 * debugging options may cause it to be larger in which case the
	 * implementation will limit the pfns advertised through
	 * ->direct_access() to those that are included in the memmap.
	 */
	if (nd_pfn->mode == PFN_MODE_PMEM)
		offset = ALIGN(SZ_8K + 64 * npfns, nd_pfn->align);
	else if (nd_pfn->mode == PFN_MODE_RAM)
		offset = ALIGN(SZ_8K, nd_pfn->align);
	else
		goto err;

	npfns = (pmem->size - offset) / SZ_4K;
	pfn_sb->mode = cpu_to_le32(nd_pfn->mode);
	pfn_sb->dataoff = cpu_to_le64(offset);
	pfn_sb->npfns = cpu_to_le64(npfns);
	memcpy(pfn_sb->signature, PFN_SIG, PFN_SIG_LEN);
	memcpy(pfn_sb->uuid, nd_pfn->uuid, 16);
	memcpy(pfn_sb->parent_uuid, nd_dev_to_uuid(&ndns->dev), 16);
	pfn_sb->version_major = cpu_to_le16(1);
	checksum = nd_sb_checksum((struct nd_gen_sb *) pfn_sb);
	pfn_sb->checksum = cpu_to_le64(checksum);

	rc = nvdimm_write_bytes(ndns, SZ_4K, pfn_sb, sizeof(*pfn_sb));
	if (rc)
		goto err;

	return 0;
 err:
	nd_pfn->pfn_sb = NULL;
	kfree(pfn_sb);
	return -ENXIO;
}

static int nvdimm_namespace_detach_pfn(struct nd_namespace_common *ndns)
{
	struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim);
	struct pmem_device *pmem;

	/* free pmem disk */
	pmem = dev_get_drvdata(&nd_pfn->dev);
	pmem_detach_disk(pmem);

	/* release nd_pfn resources */
	kfree(nd_pfn->pfn_sb);
	nd_pfn->pfn_sb = NULL;

	return 0;
}

static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns)
{
	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim);
	struct device *dev = &nd_pfn->dev;
	struct vmem_altmap *altmap;
	struct nd_region *nd_region;
	struct nd_pfn_sb *pfn_sb;
	struct pmem_device *pmem;
	phys_addr_t offset;
	int rc;

	if (!nd_pfn->uuid || !nd_pfn->ndns)
		return -ENODEV;

	nd_region = to_nd_region(dev->parent);
	rc = nd_pfn_init(nd_pfn);
	if (rc)
		return rc;

	pfn_sb = nd_pfn->pfn_sb;
	offset = le64_to_cpu(pfn_sb->dataoff);
	nd_pfn->mode = le32_to_cpu(nd_pfn->pfn_sb->mode);
	if (nd_pfn->mode == PFN_MODE_RAM) {
		if (offset < SZ_8K)
			return -EINVAL;
		nd_pfn->npfns = le64_to_cpu(pfn_sb->npfns);
		altmap = NULL;
	} else {
		rc = -ENXIO;
		goto err;
	}

	/* establish pfn range for lookup, and switch to direct map */
	pmem = dev_get_drvdata(dev);
	devm_memunmap(dev, (void __force *) pmem->virt_addr);
	pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, &nsio->res);
	pmem->pfn_flags |= PFN_MAP;
	if (IS_ERR(pmem->virt_addr)) {
		rc = PTR_ERR(pmem->virt_addr);
		goto err;
	}

	/* attach pmem disk in "pfn-mode" */
	pmem->data_offset = offset;
	rc = pmem_attach_disk(dev, ndns, pmem);
	if (rc)
		goto err;

	return rc;
 err:
	nvdimm_namespace_detach_pfn(ndns);
	return rc;
}

static int nd_pmem_probe(struct device *dev)
{
	struct nd_region *nd_region = to_nd_region(dev->parent);
	struct nd_namespace_common *ndns;
	struct nd_namespace_io *nsio;
	struct pmem_device *pmem;

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

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

	pmem->ndns = ndns;
	dev_set_drvdata(dev, pmem);
	ndns->rw_bytes = pmem_rw_bytes;
	if (devm_init_badblocks(dev, &pmem->bb))
		return -ENOMEM;
	nvdimm_namespace_add_poison(ndns, &pmem->bb, 0);

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

	if (is_nd_pfn(dev))
		return nvdimm_namespace_attach_pfn(ndns);

	if (nd_btt_probe(ndns, pmem) == 0) {
		/* we'll come back as btt-pmem */
		return -ENXIO;
	}

	if (nd_pfn_probe(ndns, pmem) == 0) {
		/* we'll come back as pfn-pmem */
		return -ENXIO;
	}

	return pmem_attach_disk(dev, ndns, pmem);
}

static int nd_pmem_remove(struct device *dev)
{
	struct pmem_device *pmem = dev_get_drvdata(dev);

	if (is_nd_btt(dev))
		nvdimm_namespace_detach_btt(pmem->ndns);
	else if (is_nd_pfn(dev))
		nvdimm_namespace_detach_pfn(pmem->ndns);
	else
		pmem_detach_disk(pmem);

	return 0;
}

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,
	.drv = {
		.name = "nd_pmem",
	},
	.type = ND_DRIVER_NAMESPACE_IO | ND_DRIVER_NAMESPACE_PMEM,
};

static int __init pmem_init(void)
{
	int error;

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

	error = nd_driver_register(&nd_pmem_driver);
	if (error) {
		unregister_blkdev(pmem_major, "pmem");
		return error;
	}

	return 0;
}
module_init(pmem_init);

static void pmem_exit(void)
{
	driver_unregister(&nd_pmem_driver.drv);
	unregister_blkdev(pmem_major, "pmem");
}
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
	35	struct pmem_device {
	36	struct request_queue *pmem_queue;
	37	struct gendisk *pmem_disk;
32ab0a3f	38	struct nd_namespace_common *ndns;
9e853f23 RZ	39
	40	/* One contiguous memory region per device */
	41	phys_addr_t phys_addr;
32ab0a3f DW	42	/* when non-zero this device is hosting a 'pfn' instance */
32ab0a3f DW	43	phys_addr_t data_offset;
34c0fd54	44	unsigned long pfn_flags;
61031952	45	void __pmem *virt_addr;
9e853f23	46	size_t size;
b95f5f43	47	struct badblocks bb;
9e853f23 RZ	48	};
	49
	50	static int pmem_major;
9e853f23	51
e10624f8 DW	52	static bool is_bad_pmem(struct badblocks *bb, sector_t sector, unsigned int len)
	53	{
	54	if (bb->count) {
	55	sector_t first_bad;
	56	int num_bad;
	57
	58	return !!badblocks_check(bb, sector, len / 512, &first_bad,
	59	&num_bad);
	60	}
	61
	62	return false;
	63	}
	64
	65	static int pmem_do_bvec(struct pmem_device pmem, struct page page,
9e853f23 RZ	66	unsigned int len, unsigned int off, int rw,
	67	sector_t sector)
	68	{
	69	void *mem = kmap_atomic(page);
32ab0a3f	70	phys_addr_t pmem_off = sector * 512 + pmem->data_offset;
61031952	71	void __pmem *pmem_addr = pmem->virt_addr + pmem_off;
9e853f23 RZ	72
9e853f23 RZ	73	if (rw == READ) {
e10624f8 DW	74	if (unlikely(is_bad_pmem(&pmem->bb, sector, len)))
e10624f8 DW	75	return -EIO;
61031952	76	memcpy_from_pmem(mem + off, pmem_addr, len);
9e853f23 RZ	77	flush_dcache_page(page);
	78	} else {
	79	flush_dcache_page(page);
61031952	80	memcpy_to_pmem(pmem_addr, mem + off, len);
9e853f23 RZ	81	}
	82
	83	kunmap_atomic(mem);
e10624f8	84	return 0;
9e853f23 RZ	85	}
9e853f23 RZ	86
dece1635	87	static blk_qc_t pmem_make_request(struct request_queue q, struct bio bio)
9e853f23	88	{
e10624f8	89	int rc = 0;
f0dc089c DW	90	bool do_acct;
f0dc089c DW	91	unsigned long start;
9e853f23	92	struct bio_vec bvec;
9e853f23	93	struct bvec_iter iter;
edc870e5 DW	94	struct block_device *bdev = bio->bi_bdev;
edc870e5 DW	95	struct pmem_device *pmem = bdev->bd_disk->private_data;
9e853f23	96
f0dc089c	97	do_acct = nd_iostat_start(bio, &start);
e10624f8 DW	98	bio_for_each_segment(bvec, bio, iter) {
	99	rc = pmem_do_bvec(pmem, bvec.bv_page, bvec.bv_len,
	100	bvec.bv_offset, bio_data_dir(bio),
	101	iter.bi_sector);
	102	if (rc) {
	103	bio->bi_error = rc;
	104	break;
	105	}
	106	}
f0dc089c DW	107	if (do_acct)
f0dc089c DW	108	nd_iostat_end(bio, start);
61031952 RZ	109
	110	if (bio_data_dir(bio))
	111	wmb_pmem();
	112
4246a0b6	113	bio_endio(bio);
dece1635	114	return BLK_QC_T_NONE;
9e853f23 RZ	115	}
	116
	117	static int pmem_rw_page(struct block_device *bdev, sector_t sector,
	118	struct page *page, int rw)
	119	{
	120	struct pmem_device *pmem = bdev->bd_disk->private_data;
e10624f8	121	int rc;
9e853f23	122
e10624f8	123	rc = pmem_do_bvec(pmem, page, PAGE_CACHE_SIZE, 0, rw, sector);
ba8fe0f8 RZ	124	if (rw & WRITE)
ba8fe0f8 RZ	125	wmb_pmem();
9e853f23	126
e10624f8 DW	127	/*
	128	* The ->rw_page interface is subtle and tricky. The core
	129	* retries on any error, so we can only invoke page_endio() in
	130	* the successful completion case. Otherwise, we'll see crashes
	131	* caused by double completion.
	132	*/
	133	if (rc == 0)
	134	page_endio(page, rw & WRITE, 0);
	135
	136	return rc;
9e853f23 RZ	137	}
	138
	139	static long pmem_direct_access(struct block_device *bdev, sector_t sector,
34c0fd54	140	void __pmem *kaddr, pfn_t pfn)
9e853f23 RZ	141	{
9e853f23 RZ	142	struct pmem_device *pmem = bdev->bd_disk->private_data;
32ab0a3f	143	resource_size_t offset = sector * 512 + pmem->data_offset;
589e75d1	144
e2e05394	145	*kaddr = pmem->virt_addr + offset;
34c0fd54	146	*pfn = phys_to_pfn_t(pmem->phys_addr + offset, pmem->pfn_flags);
9e853f23	147
589e75d1	148	return pmem->size - offset;
9e853f23 RZ	149	}
	150
	151	static const struct block_device_operations pmem_fops = {
	152	.owner = THIS_MODULE,
	153	.rw_page = pmem_rw_page,
	154	.direct_access = pmem_direct_access,
58138820	155	.revalidate_disk = nvdimm_revalidate_disk,
9e853f23 RZ	156	};
9e853f23 RZ	157
9f53f9fa DW	158	static struct pmem_device pmem_alloc(struct device dev,
9f53f9fa DW	159	struct resource *res, int id)
9e853f23 RZ	160	{
9e853f23 RZ	161	struct pmem_device *pmem;
9e853f23	162
708ab62b	163	pmem = devm_kzalloc(dev, sizeof(*pmem), GFP_KERNEL);
9e853f23	164	if (!pmem)
8c2f7e86	165	return ERR_PTR(-ENOMEM);
9e853f23 RZ	166
	167	pmem->phys_addr = res->start;
	168	pmem->size = resource_size(res);
96601adb	169	if (!arch_has_wmb_pmem())
61031952	170	dev_warn(dev, "unable to guarantee persistence of writes\n");
9e853f23	171
708ab62b CH	172	if (!devm_request_mem_region(dev, pmem->phys_addr, pmem->size,
708ab62b CH	173	dev_name(dev))) {
9f53f9fa DW	174	dev_warn(dev, "could not reserve region [0x%pa:0x%zx]\n",
9f53f9fa DW	175	&pmem->phys_addr, pmem->size);
8c2f7e86	176	return ERR_PTR(-EBUSY);
9e853f23 RZ	177	}
9e853f23 RZ	178
34c0fd54 DW	179	pmem->pfn_flags = PFN_DEV;
34c0fd54 DW	180	if (pmem_should_map_pages(dev)) {
b36f4761	181	pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, res);
34c0fd54 DW	182	pmem->pfn_flags \|= PFN_MAP;
34c0fd54 DW	183	} else
a639315d DW	184	pmem->virt_addr = (void __pmem *) devm_memremap(dev,
	185	pmem->phys_addr, pmem->size,
	186	ARCH_MEMREMAP_PMEM);
b36f4761 DW	187
	188	if (IS_ERR(pmem->virt_addr))
	189	return (void __force *) pmem->virt_addr;
8c2f7e86 DW	190
	191	return pmem;
	192	}
	193
	194	static void pmem_detach_disk(struct pmem_device *pmem)
	195	{
32ab0a3f DW	196	if (!pmem->pmem_disk)
	197	return;
	198
8c2f7e86 DW	199	del_gendisk(pmem->pmem_disk);
	200	put_disk(pmem->pmem_disk);
	201	blk_cleanup_queue(pmem->pmem_queue);
	202	}
	203
32ab0a3f DW	204	static int pmem_attach_disk(struct device *dev,
32ab0a3f DW	205	struct nd_namespace_common ndns, struct pmem_device pmem)
8c2f7e86	206	{
538ea4aa	207	int nid = dev_to_node(dev);
8c2f7e86	208	struct gendisk *disk;
9e853f23	209
538ea4aa	210	pmem->pmem_queue = blk_alloc_queue_node(GFP_KERNEL, nid);
9e853f23	211	if (!pmem->pmem_queue)
8c2f7e86	212	return -ENOMEM;
9e853f23 RZ	213
9e853f23 RZ	214	blk_queue_make_request(pmem->pmem_queue, pmem_make_request);
6b47496a	215	blk_queue_physical_block_size(pmem->pmem_queue, PAGE_SIZE);
43d3fa3a	216	blk_queue_max_hw_sectors(pmem->pmem_queue, UINT_MAX);
9e853f23	217	blk_queue_bounce_limit(pmem->pmem_queue, BLK_BOUNCE_ANY);
0f51c4fa	218	queue_flag_set_unlocked(QUEUE_FLAG_NONROT, pmem->pmem_queue);
9e853f23	219
538ea4aa	220	disk = alloc_disk_node(0, nid);
8c2f7e86 DW	221	if (!disk) {
	222	blk_cleanup_queue(pmem->pmem_queue);
	223	return -ENOMEM;
	224	}
9e853f23	225
9e853f23	226	disk->major = pmem_major;
9f53f9fa	227	disk->first_minor = 0;
9e853f23 RZ	228	disk->fops = &pmem_fops;
	229	disk->private_data = pmem;
	230	disk->queue = pmem->pmem_queue;
	231	disk->flags = GENHD_FL_EXT_DEVT;
5212e11f	232	nvdimm_namespace_disk_name(ndns, disk->disk_name);
32ab0a3f DW	233	disk->driverfs_dev = dev;
32ab0a3f DW	234	set_capacity(disk, (pmem->size - pmem->data_offset) / 512);
9e853f23	235	pmem->pmem_disk = disk;
710d69cc	236	devm_exit_badblocks(dev, &pmem->bb);
b95f5f43 DW	237	if (devm_init_badblocks(dev, &pmem->bb))
	238	return -ENOMEM;
	239	nvdimm_namespace_add_poison(ndns, &pmem->bb, pmem->data_offset);
9e853f23	240
57f7f317	241	disk->bb = &pmem->bb;
9e853f23	242	add_disk(disk);
58138820	243	revalidate_disk(disk);
9e853f23	244
8c2f7e86 DW	245	return 0;
8c2f7e86 DW	246	}
9e853f23	247
8c2f7e86 DW	248	static int pmem_rw_bytes(struct nd_namespace_common *ndns,
	249	resource_size_t offset, void *buf, size_t size, int rw)
	250	{
	251	struct pmem_device *pmem = dev_get_drvdata(ndns->claim);
	252
	253	if (unlikely(offset + size > pmem->size)) {
	254	dev_WARN_ONCE(&ndns->dev, 1, "request out of range\n");
	255	return -EFAULT;
	256	}
	257
710d69cc DW	258	if (rw == READ) {
	259	unsigned int sz_align = ALIGN(size + (offset & (512 - 1)), 512);
	260
	261	if (unlikely(is_bad_pmem(&pmem->bb, offset / 512, sz_align)))
	262	return -EIO;
61031952	263	memcpy_from_pmem(buf, pmem->virt_addr + offset, size);
710d69cc	264	} else {
61031952 RZ	265	memcpy_to_pmem(pmem->virt_addr + offset, buf, size);
	266	wmb_pmem();
	267	}
8c2f7e86 DW	268
	269	return 0;
	270	}
	271
32ab0a3f DW	272	static int nd_pfn_init(struct nd_pfn *nd_pfn)
	273	{
	274	struct nd_pfn_sb pfn_sb = kzalloc(sizeof(pfn_sb), GFP_KERNEL);
	275	struct pmem_device *pmem = dev_get_drvdata(&nd_pfn->dev);
	276	struct nd_namespace_common *ndns = nd_pfn->ndns;
	277	struct nd_region *nd_region;
	278	unsigned long npfns;
	279	phys_addr_t offset;
	280	u64 checksum;
	281	int rc;
	282
	283	if (!pfn_sb)
	284	return -ENOMEM;
	285
	286	nd_pfn->pfn_sb = pfn_sb;
	287	rc = nd_pfn_validate(nd_pfn);
3fa96268 DW	288	if (rc == -ENODEV)
	289	/* no info block, do init */;
	290	else
32ab0a3f DW	291	return rc;
32ab0a3f DW	292
32ab0a3f DW	293	nd_region = to_nd_region(nd_pfn->dev.parent);
	294	if (nd_region->ro) {
	295	dev_info(&nd_pfn->dev,
	296	"%s is read-only, unable to init metadata\n",
	297	dev_name(&nd_region->dev));
	298	goto err;
	299	}
	300
	301	memset(pfn_sb, 0, sizeof(*pfn_sb));
	302	npfns = (pmem->size - SZ_8K) / SZ_4K;
	303	/*
	304	* Note, we use 64 here for the standard size of struct page,
	305	* debugging options may cause it to be larger in which case the
	306	* implementation will limit the pfns advertised through
	307	* ->direct_access() to those that are included in the memmap.
	308	*/
	309	if (nd_pfn->mode == PFN_MODE_PMEM)
315c5625	310	offset = ALIGN(SZ_8K + 64 * npfns, nd_pfn->align);
32ab0a3f	311	else if (nd_pfn->mode == PFN_MODE_RAM)
315c5625	312	offset = ALIGN(SZ_8K, nd_pfn->align);
32ab0a3f DW	313	else
	314	goto err;
	315
	316	npfns = (pmem->size - offset) / SZ_4K;
	317	pfn_sb->mode = cpu_to_le32(nd_pfn->mode);
	318	pfn_sb->dataoff = cpu_to_le64(offset);
	319	pfn_sb->npfns = cpu_to_le64(npfns);
	320	memcpy(pfn_sb->signature, PFN_SIG, PFN_SIG_LEN);
	321	memcpy(pfn_sb->uuid, nd_pfn->uuid, 16);
a34d5e8a	322	memcpy(pfn_sb->parent_uuid, nd_dev_to_uuid(&ndns->dev), 16);
32ab0a3f DW	323	pfn_sb->version_major = cpu_to_le16(1);
	324	checksum = nd_sb_checksum((struct nd_gen_sb *) pfn_sb);
	325	pfn_sb->checksum = cpu_to_le64(checksum);
	326
	327	rc = nvdimm_write_bytes(ndns, SZ_4K, pfn_sb, sizeof(*pfn_sb));
	328	if (rc)
	329	goto err;
	330
	331	return 0;
	332	err:
	333	nd_pfn->pfn_sb = NULL;
	334	kfree(pfn_sb);
	335	return -ENXIO;
	336	}
	337
	338	static int nvdimm_namespace_detach_pfn(struct nd_namespace_common *ndns)
	339	{
	340	struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim);
	341	struct pmem_device *pmem;
	342
	343	/* free pmem disk */
	344	pmem = dev_get_drvdata(&nd_pfn->dev);
	345	pmem_detach_disk(pmem);
	346
	347	/* release nd_pfn resources */
	348	kfree(nd_pfn->pfn_sb);
	349	nd_pfn->pfn_sb = NULL;
	350
	351	return 0;
	352	}
	353
	354	static int nvdimm_namespace_attach_pfn(struct nd_namespace_common *ndns)
9e853f23	355	{
32ab0a3f DW	356	struct nd_namespace_io *nsio = to_nd_namespace_io(&ndns->dev);
	357	struct nd_pfn *nd_pfn = to_nd_pfn(ndns->claim);
	358	struct device *dev = &nd_pfn->dev;
	359	struct vmem_altmap *altmap;
	360	struct nd_region *nd_region;
	361	struct nd_pfn_sb *pfn_sb;
	362	struct pmem_device *pmem;
	363	phys_addr_t offset;
	364	int rc;
	365
	366	if (!nd_pfn->uuid \|\| !nd_pfn->ndns)
	367	return -ENODEV;
	368
	369	nd_region = to_nd_region(dev->parent);
	370	rc = nd_pfn_init(nd_pfn);
	371	if (rc)
	372	return rc;
	373
32ab0a3f DW	374	pfn_sb = nd_pfn->pfn_sb;
	375	offset = le64_to_cpu(pfn_sb->dataoff);
	376	nd_pfn->mode = le32_to_cpu(nd_pfn->pfn_sb->mode);
	377	if (nd_pfn->mode == PFN_MODE_RAM) {
315c5625	378	if (offset < SZ_8K)
32ab0a3f DW	379	return -EINVAL;
	380	nd_pfn->npfns = le64_to_cpu(pfn_sb->npfns);
	381	altmap = NULL;
	382	} else {
	383	rc = -ENXIO;
	384	goto err;
	385	}
	386
	387	/* establish pfn range for lookup, and switch to direct map */
	388	pmem = dev_get_drvdata(dev);
a639315d DW	389	devm_memunmap(dev, (void __force *) pmem->virt_addr);
a639315d DW	390	pmem->virt_addr = (void __pmem *) devm_memremap_pages(dev, &nsio->res);
34c0fd54	391	pmem->pfn_flags \|= PFN_MAP;
32ab0a3f DW	392	if (IS_ERR(pmem->virt_addr)) {
	393	rc = PTR_ERR(pmem->virt_addr);
	394	goto err;
	395	}
	396
	397	/* attach pmem disk in "pfn-mode" */
	398	pmem->data_offset = offset;
	399	rc = pmem_attach_disk(dev, ndns, pmem);
	400	if (rc)
	401	goto err;
	402
	403	return rc;
	404	err:
	405	nvdimm_namespace_detach_pfn(ndns);
	406	return rc;
9e853f23 RZ	407	}
9e853f23 RZ	408
9f53f9fa	409	static int nd_pmem_probe(struct device *dev)
9e853f23	410	{
9f53f9fa	411	struct nd_region *nd_region = to_nd_region(dev->parent);
8c2f7e86 DW	412	struct nd_namespace_common *ndns;
8c2f7e86 DW	413	struct nd_namespace_io *nsio;
9e853f23	414	struct pmem_device *pmem;
9e853f23	415
8c2f7e86 DW	416	ndns = nvdimm_namespace_common_probe(dev);
	417	if (IS_ERR(ndns))
	418	return PTR_ERR(ndns);
bf9bccc1	419
8c2f7e86	420	nsio = to_nd_namespace_io(&ndns->dev);
9f53f9fa	421	pmem = pmem_alloc(dev, &nsio->res, nd_region->id);
9e853f23 RZ	422	if (IS_ERR(pmem))
	423	return PTR_ERR(pmem);
	424
32ab0a3f	425	pmem->ndns = ndns;
9f53f9fa	426	dev_set_drvdata(dev, pmem);
8c2f7e86	427	ndns->rw_bytes = pmem_rw_bytes;
710d69cc DW	428	if (devm_init_badblocks(dev, &pmem->bb))
	429	return -ENOMEM;
	430	nvdimm_namespace_add_poison(ndns, &pmem->bb, 0);
708ab62b	431
8c2f7e86	432	if (is_nd_btt(dev))
708ab62b CH	433	return nvdimm_namespace_attach_btt(ndns);
708ab62b CH	434
32ab0a3f DW	435	if (is_nd_pfn(dev))
	436	return nvdimm_namespace_attach_pfn(ndns);
	437
	438	if (nd_btt_probe(ndns, pmem) == 0) {
8c2f7e86	439	/* we'll come back as btt-pmem */
708ab62b	440	return -ENXIO;
32ab0a3f DW	441	}
	442
	443	if (nd_pfn_probe(ndns, pmem) == 0) {
	444	/* we'll come back as pfn-pmem */
	445	return -ENXIO;
	446	}
	447
	448	return pmem_attach_disk(dev, ndns, pmem);
9e853f23 RZ	449	}
9e853f23 RZ	450
9f53f9fa	451	static int nd_pmem_remove(struct device *dev)
9e853f23	452	{
9f53f9fa	453	struct pmem_device *pmem = dev_get_drvdata(dev);
9e853f23	454
8c2f7e86	455	if (is_nd_btt(dev))
32ab0a3f DW	456	nvdimm_namespace_detach_btt(pmem->ndns);
	457	else if (is_nd_pfn(dev))
	458	nvdimm_namespace_detach_pfn(pmem->ndns);
8c2f7e86 DW	459	else
8c2f7e86 DW	460	pmem_detach_disk(pmem);
8c2f7e86	461
9e853f23 RZ	462	return 0;
	463	}
	464
9f53f9fa DW	465	MODULE_ALIAS("pmem");
9f53f9fa DW	466	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_IO);
bf9bccc1	467	MODULE_ALIAS_ND_DEVICE(ND_DEVICE_NAMESPACE_PMEM);
9f53f9fa DW	468	static struct nd_device_driver nd_pmem_driver = {
	469	.probe = nd_pmem_probe,
	470	.remove = nd_pmem_remove,
	471	.drv = {
	472	.name = "nd_pmem",
9e853f23	473	},
bf9bccc1	474	.type = ND_DRIVER_NAMESPACE_IO \| ND_DRIVER_NAMESPACE_PMEM,
9e853f23 RZ	475	};
	476
	477	static int __init pmem_init(void)
	478	{
	479	int error;
	480
	481	pmem_major = register_blkdev(0, "pmem");
	482	if (pmem_major < 0)
	483	return pmem_major;
	484
9f53f9fa DW	485	error = nd_driver_register(&nd_pmem_driver);
9f53f9fa DW	486	if (error) {
9e853f23	487	unregister_blkdev(pmem_major, "pmem");
9f53f9fa DW	488	return error;
	489	}
	490
	491	return 0;
9e853f23 RZ	492	}
	493	module_init(pmem_init);
	494
	495	static void pmem_exit(void)
	496	{
9f53f9fa	497	driver_unregister(&nd_pmem_driver.drv);
9e853f23 RZ	498	unregister_blkdev(pmem_major, "pmem");
	499	}
	500	module_exit(pmem_exit);
	501
	502	MODULE_AUTHOR("Ross Zwisler <ross.zwisler@linux.intel.com>");
	503	MODULE_LICENSE("GPL v2");