[linux-2.6-block.git] / fs / erofs / data.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright (C) 2017-2018 HUAWEI, Inc.
 *             https://www.huawei.com/
 * Copyright (C) 2021, Alibaba Cloud
 */
#include "internal.h"
#include <linux/prefetch.h>
#include <linux/dax.h>
#include <trace/events/erofs.h>

void erofs_unmap_metabuf(struct erofs_buf *buf)
{
	if (buf->kmap_type == EROFS_KMAP)
		kunmap(buf->page);
	else if (buf->kmap_type == EROFS_KMAP_ATOMIC)
		kunmap_atomic(buf->base);
	buf->base = NULL;
	buf->kmap_type = EROFS_NO_KMAP;
}

void erofs_put_metabuf(struct erofs_buf *buf)
{
	if (!buf->page)
		return;
	erofs_unmap_metabuf(buf);
	put_page(buf->page);
	buf->page = NULL;
}

void *erofs_bread(struct erofs_buf *buf, struct inode *inode,
		  erofs_blk_t blkaddr, enum erofs_kmap_type type)
{
	struct address_space *const mapping = inode->i_mapping;
	erofs_off_t offset = blknr_to_addr(blkaddr);
	pgoff_t index = offset >> PAGE_SHIFT;
	struct page *page = buf->page;

	if (!page || page->index != index) {
		erofs_put_metabuf(buf);
		page = read_cache_page_gfp(mapping, index,
				mapping_gfp_constraint(mapping, ~__GFP_FS));
		if (IS_ERR(page))
			return page;
		/* should already be PageUptodate, no need to lock page */
		buf->page = page;
	}
	if (buf->kmap_type == EROFS_NO_KMAP) {
		if (type == EROFS_KMAP)
			buf->base = kmap(page);
		else if (type == EROFS_KMAP_ATOMIC)
			buf->base = kmap_atomic(page);
		buf->kmap_type = type;
	} else if (buf->kmap_type != type) {
		DBG_BUGON(1);
		return ERR_PTR(-EFAULT);
	}
	if (type == EROFS_NO_KMAP)
		return NULL;
	return buf->base + (offset & ~PAGE_MASK);
}

void *erofs_read_metabuf(struct erofs_buf *buf, struct super_block *sb,
			 erofs_blk_t blkaddr, enum erofs_kmap_type type)
{
	return erofs_bread(buf, sb->s_bdev->bd_inode, blkaddr, type);
}

static int erofs_map_blocks_flatmode(struct inode *inode,
				     struct erofs_map_blocks *map,
				     int flags)
{
	erofs_blk_t nblocks, lastblk;
	u64 offset = map->m_la;
	struct erofs_inode *vi = EROFS_I(inode);
	bool tailendpacking = (vi->datalayout == EROFS_INODE_FLAT_INLINE);

	nblocks = DIV_ROUND_UP(inode->i_size, EROFS_BLKSIZ);
	lastblk = nblocks - tailendpacking;

	/* there is no hole in flatmode */
	map->m_flags = EROFS_MAP_MAPPED;
	if (offset < blknr_to_addr(lastblk)) {
		map->m_pa = blknr_to_addr(vi->raw_blkaddr) + map->m_la;
		map->m_plen = blknr_to_addr(lastblk) - offset;
	} else if (tailendpacking) {
		/* 2 - inode inline B: inode, [xattrs], inline last blk... */
		struct erofs_sb_info *sbi = EROFS_SB(inode->i_sb);

		map->m_pa = iloc(sbi, vi->nid) + vi->inode_isize +
			vi->xattr_isize + erofs_blkoff(map->m_la);
		map->m_plen = inode->i_size - offset;

		/* inline data should be located in the same meta block */
		if (erofs_blkoff(map->m_pa) + map->m_plen > EROFS_BLKSIZ) {
			erofs_err(inode->i_sb,
				  "inline data cross block boundary @ nid %llu",
				  vi->nid);
			DBG_BUGON(1);
			return -EFSCORRUPTED;
		}
		map->m_flags |= EROFS_MAP_META;
	} else {
		erofs_err(inode->i_sb,
			  "internal error @ nid: %llu (size %llu), m_la 0x%llx",
			  vi->nid, inode->i_size, map->m_la);
		DBG_BUGON(1);
		return -EIO;
	}
	return 0;
}

static int erofs_map_blocks(struct inode *inode,
			    struct erofs_map_blocks *map, int flags)
{
	struct super_block *sb = inode->i_sb;
	struct erofs_inode *vi = EROFS_I(inode);
	struct erofs_inode_chunk_index *idx;
	struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
	u64 chunknr;
	unsigned int unit;
	erofs_off_t pos;
	void *kaddr;
	int err = 0;

	trace_erofs_map_blocks_enter(inode, map, flags);
	map->m_deviceid = 0;
	if (map->m_la >= inode->i_size) {
		/* leave out-of-bound access unmapped */
		map->m_flags = 0;
		map->m_plen = 0;
		goto out;
	}

	if (vi->datalayout != EROFS_INODE_CHUNK_BASED) {
		err = erofs_map_blocks_flatmode(inode, map, flags);
		goto out;
	}

	if (vi->chunkformat & EROFS_CHUNK_FORMAT_INDEXES)
		unit = sizeof(*idx);			/* chunk index */
	else
		unit = EROFS_BLOCK_MAP_ENTRY_SIZE;	/* block map */

	chunknr = map->m_la >> vi->chunkbits;
	pos = ALIGN(iloc(EROFS_SB(sb), vi->nid) + vi->inode_isize +
		    vi->xattr_isize, unit) + unit * chunknr;

	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos), EROFS_KMAP);
	if (IS_ERR(kaddr)) {
		err = PTR_ERR(kaddr);
		goto out;
	}
	map->m_la = chunknr << vi->chunkbits;
	map->m_plen = min_t(erofs_off_t, 1UL << vi->chunkbits,
			    roundup(inode->i_size - map->m_la, EROFS_BLKSIZ));

	/* handle block map */
	if (!(vi->chunkformat & EROFS_CHUNK_FORMAT_INDEXES)) {
		__le32 *blkaddr = kaddr + erofs_blkoff(pos);

		if (le32_to_cpu(*blkaddr) == EROFS_NULL_ADDR) {
			map->m_flags = 0;
		} else {
			map->m_pa = blknr_to_addr(le32_to_cpu(*blkaddr));
			map->m_flags = EROFS_MAP_MAPPED;
		}
		goto out_unlock;
	}
	/* parse chunk indexes */
	idx = kaddr + erofs_blkoff(pos);
	switch (le32_to_cpu(idx->blkaddr)) {
	case EROFS_NULL_ADDR:
		map->m_flags = 0;
		break;
	default:
		map->m_deviceid = le16_to_cpu(idx->device_id) &
			EROFS_SB(sb)->device_id_mask;
		map->m_pa = blknr_to_addr(le32_to_cpu(idx->blkaddr));
		map->m_flags = EROFS_MAP_MAPPED;
		break;
	}
out_unlock:
	erofs_put_metabuf(&buf);
out:
	if (!err)
		map->m_llen = map->m_plen;
	trace_erofs_map_blocks_exit(inode, map, flags, 0);
	return err;
}

int erofs_map_dev(struct super_block *sb, struct erofs_map_dev *map)
{
	struct erofs_dev_context *devs = EROFS_SB(sb)->devs;
	struct erofs_device_info *dif;
	int id;

	/* primary device by default */
	map->m_bdev = sb->s_bdev;
	map->m_daxdev = EROFS_SB(sb)->dax_dev;
	map->m_dax_part_off = EROFS_SB(sb)->dax_part_off;

	if (map->m_deviceid) {
		down_read(&devs->rwsem);
		dif = idr_find(&devs->tree, map->m_deviceid - 1);
		if (!dif) {
			up_read(&devs->rwsem);
			return -ENODEV;
		}
		map->m_bdev = dif->bdev;
		map->m_daxdev = dif->dax_dev;
		map->m_dax_part_off = dif->dax_part_off;
		up_read(&devs->rwsem);
	} else if (devs->extra_devices) {
		down_read(&devs->rwsem);
		idr_for_each_entry(&devs->tree, dif, id) {
			erofs_off_t startoff, length;

			if (!dif->mapped_blkaddr)
				continue;
			startoff = blknr_to_addr(dif->mapped_blkaddr);
			length = blknr_to_addr(dif->blocks);

			if (map->m_pa >= startoff &&
			    map->m_pa < startoff + length) {
				map->m_pa -= startoff;
				map->m_bdev = dif->bdev;
				map->m_daxdev = dif->dax_dev;
				map->m_dax_part_off = dif->dax_part_off;
				break;
			}
		}
		up_read(&devs->rwsem);
	}
	return 0;
}

static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
		unsigned int flags, struct iomap *iomap, struct iomap *srcmap)
{
	int ret;
	struct erofs_map_blocks map;
	struct erofs_map_dev mdev;

	map.m_la = offset;
	map.m_llen = length;

	ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
	if (ret < 0)
		return ret;

	mdev = (struct erofs_map_dev) {
		.m_deviceid = map.m_deviceid,
		.m_pa = map.m_pa,
	};
	ret = erofs_map_dev(inode->i_sb, &mdev);
	if (ret)
		return ret;

	iomap->offset = map.m_la;
	if (flags & IOMAP_DAX)
		iomap->dax_dev = mdev.m_daxdev;
	else
		iomap->bdev = mdev.m_bdev;
	iomap->length = map.m_llen;
	iomap->flags = 0;
	iomap->private = NULL;

	if (!(map.m_flags & EROFS_MAP_MAPPED)) {
		iomap->type = IOMAP_HOLE;
		iomap->addr = IOMAP_NULL_ADDR;
		if (!iomap->length)
			iomap->length = length;
		return 0;
	}

	if (map.m_flags & EROFS_MAP_META) {
		void *ptr;
		struct erofs_buf buf = __EROFS_BUF_INITIALIZER;

		iomap->type = IOMAP_INLINE;
		ptr = erofs_read_metabuf(&buf, inode->i_sb,
					 erofs_blknr(mdev.m_pa), EROFS_KMAP);
		if (IS_ERR(ptr))
			return PTR_ERR(ptr);
		iomap->inline_data = ptr + erofs_blkoff(mdev.m_pa);
		iomap->private = buf.base;
	} else {
		iomap->type = IOMAP_MAPPED;
		iomap->addr = mdev.m_pa;
		if (flags & IOMAP_DAX)
			iomap->addr += mdev.m_dax_part_off;
	}
	return 0;
}

static int erofs_iomap_end(struct inode *inode, loff_t pos, loff_t length,
		ssize_t written, unsigned int flags, struct iomap *iomap)
{
	void *ptr = iomap->private;

	if (ptr) {
		struct erofs_buf buf = {
			.page = kmap_to_page(ptr),
			.base = ptr,
			.kmap_type = EROFS_KMAP,
		};

		DBG_BUGON(iomap->type != IOMAP_INLINE);
		erofs_put_metabuf(&buf);
	} else {
		DBG_BUGON(iomap->type == IOMAP_INLINE);
	}
	return written;
}

static const struct iomap_ops erofs_iomap_ops = {
	.iomap_begin = erofs_iomap_begin,
	.iomap_end = erofs_iomap_end,
};

int erofs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
		 u64 start, u64 len)
{
	if (erofs_inode_is_data_compressed(EROFS_I(inode)->datalayout)) {
#ifdef CONFIG_EROFS_FS_ZIP
		return iomap_fiemap(inode, fieinfo, start, len,
				    &z_erofs_iomap_report_ops);
#else
		return -EOPNOTSUPP;
#endif
	}
	return iomap_fiemap(inode, fieinfo, start, len, &erofs_iomap_ops);
}

/*
 * since we dont have write or truncate flows, so no inode
 * locking needs to be held at the moment.
 */
static int erofs_readpage(struct file *file, struct page *page)
{
	return iomap_readpage(page, &erofs_iomap_ops);
}

static void erofs_readahead(struct readahead_control *rac)
{
	return iomap_readahead(rac, &erofs_iomap_ops);
}

static sector_t erofs_bmap(struct address_space *mapping, sector_t block)
{
	return iomap_bmap(mapping, block, &erofs_iomap_ops);
}

static int erofs_prepare_dio(struct kiocb *iocb, struct iov_iter *to)
{
	struct inode *inode = file_inode(iocb->ki_filp);
	loff_t align = iocb->ki_pos | iov_iter_count(to) |
		iov_iter_alignment(to);
	struct block_device *bdev = inode->i_sb->s_bdev;
	unsigned int blksize_mask;

	if (bdev)
		blksize_mask = (1 << ilog2(bdev_logical_block_size(bdev))) - 1;
	else
		blksize_mask = (1 << inode->i_blkbits) - 1;

	if (align & blksize_mask)
		return -EINVAL;
	return 0;
}

static ssize_t erofs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
{
	/* no need taking (shared) inode lock since it's a ro filesystem */
	if (!iov_iter_count(to))
		return 0;

#ifdef CONFIG_FS_DAX
	if (IS_DAX(iocb->ki_filp->f_mapping->host))
		return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
#endif
	if (iocb->ki_flags & IOCB_DIRECT) {
		int err = erofs_prepare_dio(iocb, to);

		if (!err)
			return iomap_dio_rw(iocb, to, &erofs_iomap_ops,
					    NULL, 0, 0);
		if (err < 0)
			return err;
	}
	return filemap_read(iocb, to, 0);
}

/* for uncompressed (aligned) files and raw access for other files */
const struct address_space_operations erofs_raw_access_aops = {
	.readpage = erofs_readpage,
	.readahead = erofs_readahead,
	.bmap = erofs_bmap,
	.direct_IO = noop_direct_IO,
};

#ifdef CONFIG_FS_DAX
static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
		enum page_entry_size pe_size)
{
	return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
}

static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
{
	return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
}

static const struct vm_operations_struct erofs_dax_vm_ops = {
	.fault		= erofs_dax_fault,
	.huge_fault	= erofs_dax_huge_fault,
};

static int erofs_file_mmap(struct file *file, struct vm_area_struct *vma)
{
	if (!IS_DAX(file_inode(file)))
		return generic_file_readonly_mmap(file, vma);

	if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
		return -EINVAL;

	vma->vm_ops = &erofs_dax_vm_ops;
	vma->vm_flags |= VM_HUGEPAGE;
	return 0;
}
#else
#define erofs_file_mmap	generic_file_readonly_mmap
#endif

const struct file_operations erofs_file_fops = {
	.llseek		= generic_file_llseek,
	.read_iter	= erofs_file_read_iter,
	.mmap		= erofs_file_mmap,
	.splice_read	= generic_file_splice_read,
};
Commit	Line	Data
29b24f6c	1	// SPDX-License-Identifier: GPL-2.0-only
81781b02	2	/*
81781b02	3	* Copyright (C) 2017-2018 HUAWEI, Inc.
592e7cd0	4	* https://www.huawei.com/
c5aa903a	5	* Copyright (C) 2021, Alibaba Cloud
81781b02 GX	6	*/
	7	#include "internal.h"
	8	#include <linux/prefetch.h>
06252e9c	9	#include <linux/dax.h>
13f06f48 CY	10	#include <trace/events/erofs.h>
13f06f48 CY	11
fdf80a47 GX	12	void erofs_unmap_metabuf(struct erofs_buf *buf)
	13	{
	14	if (buf->kmap_type == EROFS_KMAP)
	15	kunmap(buf->page);
	16	else if (buf->kmap_type == EROFS_KMAP_ATOMIC)
	17	kunmap_atomic(buf->base);
	18	buf->base = NULL;
	19	buf->kmap_type = EROFS_NO_KMAP;
	20	}
	21
	22	void erofs_put_metabuf(struct erofs_buf *buf)
	23	{
	24	if (!buf->page)
	25	return;
	26	erofs_unmap_metabuf(buf);
	27	put_page(buf->page);
	28	buf->page = NULL;
	29	}
	30
fe5de585 GX	31	void erofs_bread(struct erofs_buf buf, struct inode *inode,
fe5de585 GX	32	erofs_blk_t blkaddr, enum erofs_kmap_type type)
fdf80a47	33	{
fe5de585	34	struct address_space *const mapping = inode->i_mapping;
fdf80a47 GX	35	erofs_off_t offset = blknr_to_addr(blkaddr);
	36	pgoff_t index = offset >> PAGE_SHIFT;
	37	struct page *page = buf->page;
	38
	39	if (!page \|\| page->index != index) {
	40	erofs_put_metabuf(buf);
	41	page = read_cache_page_gfp(mapping, index,
	42	mapping_gfp_constraint(mapping, ~__GFP_FS));
	43	if (IS_ERR(page))
	44	return page;
	45	/* should already be PageUptodate, no need to lock page */
	46	buf->page = page;
	47	}
	48	if (buf->kmap_type == EROFS_NO_KMAP) {
	49	if (type == EROFS_KMAP)
	50	buf->base = kmap(page);
	51	else if (type == EROFS_KMAP_ATOMIC)
	52	buf->base = kmap_atomic(page);
	53	buf->kmap_type = type;
	54	} else if (buf->kmap_type != type) {
	55	DBG_BUGON(1);
	56	return ERR_PTR(-EFAULT);
	57	}
	58	if (type == EROFS_NO_KMAP)
	59	return NULL;
	60	return buf->base + (offset & ~PAGE_MASK);
	61	}
	62
fe5de585 GX	63	void erofs_read_metabuf(struct erofs_buf buf, struct super_block *sb,
	64	erofs_blk_t blkaddr, enum erofs_kmap_type type)
	65	{
	66	return erofs_bread(buf, sb->s_bdev->bd_inode, blkaddr, type);
	67	}
	68
81781b02	69	static int erofs_map_blocks_flatmode(struct inode *inode,
f0950b02 BD	70	struct erofs_map_blocks *map,
f0950b02 BD	71	int flags)
81781b02 GX	72	{
	73	erofs_blk_t nblocks, lastblk;
	74	u64 offset = map->m_la;
a5876e24	75	struct erofs_inode *vi = EROFS_I(inode);
8a765682	76	bool tailendpacking = (vi->datalayout == EROFS_INODE_FLAT_INLINE);
81781b02	77
fdf80a47	78	nblocks = DIV_ROUND_UP(inode->i_size, EROFS_BLKSIZ);
8a765682	79	lastblk = nblocks - tailendpacking;
81781b02	80
81781b02 GX	81	/* there is no hole in flatmode */
81781b02 GX	82	map->m_flags = EROFS_MAP_MAPPED;
81781b02 GX	83	if (offset < blknr_to_addr(lastblk)) {
	84	map->m_pa = blknr_to_addr(vi->raw_blkaddr) + map->m_la;
	85	map->m_plen = blknr_to_addr(lastblk) - offset;
8a765682	86	} else if (tailendpacking) {
81781b02 GX	87	/* 2 - inode inline B: inode, [xattrs], inline last blk... */
	88	struct erofs_sb_info *sbi = EROFS_SB(inode->i_sb);
	89
	90	map->m_pa = iloc(sbi, vi->nid) + vi->inode_isize +
	91	vi->xattr_isize + erofs_blkoff(map->m_la);
	92	map->m_plen = inode->i_size - offset;
	93
469407a3 GX	94	/* inline data should be located in the same meta block */
469407a3 GX	95	if (erofs_blkoff(map->m_pa) + map->m_plen > EROFS_BLKSIZ) {
4f761fa2 GX	96	erofs_err(inode->i_sb,
	97	"inline data cross block boundary @ nid %llu",
	98	vi->nid);
9141b60c	99	DBG_BUGON(1);
469407a3	100	return -EFSCORRUPTED;
9141b60c	101	}
81781b02 GX	102	map->m_flags \|= EROFS_MAP_META;
81781b02 GX	103	} else {
4f761fa2 GX	104	erofs_err(inode->i_sb,
	105	"internal error @ nid: %llu (size %llu), m_la 0x%llx",
	106	vi->nid, inode->i_size, map->m_la);
9141b60c	107	DBG_BUGON(1);
469407a3	108	return -EIO;
81781b02	109	}
469407a3	110	return 0;
81781b02 GX	111	}
81781b02 GX	112
c5aa903a GX	113	static int erofs_map_blocks(struct inode *inode,
	114	struct erofs_map_blocks *map, int flags)
	115	{
	116	struct super_block *sb = inode->i_sb;
	117	struct erofs_inode *vi = EROFS_I(inode);
	118	struct erofs_inode_chunk_index *idx;
fdf80a47	119	struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
c5aa903a GX	120	u64 chunknr;
	121	unsigned int unit;
	122	erofs_off_t pos;
fdf80a47	123	void *kaddr;
c5aa903a GX	124	int err = 0;
c5aa903a GX	125
469407a3	126	trace_erofs_map_blocks_enter(inode, map, flags);
dfeab2e9	127	map->m_deviceid = 0;
c5aa903a GX	128	if (map->m_la >= inode->i_size) {
	129	/* leave out-of-bound access unmapped */
	130	map->m_flags = 0;
	131	map->m_plen = 0;
	132	goto out;
	133	}
	134
469407a3 GX	135	if (vi->datalayout != EROFS_INODE_CHUNK_BASED) {
	136	err = erofs_map_blocks_flatmode(inode, map, flags);
	137	goto out;
	138	}
c5aa903a GX	139
	140	if (vi->chunkformat & EROFS_CHUNK_FORMAT_INDEXES)
	141	unit = sizeof(idx); / chunk index */
	142	else
	143	unit = EROFS_BLOCK_MAP_ENTRY_SIZE; /* block map */
	144
	145	chunknr = map->m_la >> vi->chunkbits;
	146	pos = ALIGN(iloc(EROFS_SB(sb), vi->nid) + vi->inode_isize +
	147	vi->xattr_isize, unit) + unit * chunknr;
	148
fdf80a47 GX	149	kaddr = erofs_read_metabuf(&buf, sb, erofs_blknr(pos), EROFS_KMAP);
	150	if (IS_ERR(kaddr)) {
	151	err = PTR_ERR(kaddr);
469407a3 GX	152	goto out;
469407a3 GX	153	}
c5aa903a GX	154	map->m_la = chunknr << vi->chunkbits;
	155	map->m_plen = min_t(erofs_off_t, 1UL << vi->chunkbits,
	156	roundup(inode->i_size - map->m_la, EROFS_BLKSIZ));
	157
	158	/* handle block map */
	159	if (!(vi->chunkformat & EROFS_CHUNK_FORMAT_INDEXES)) {
fdf80a47	160	__le32 *blkaddr = kaddr + erofs_blkoff(pos);
c5aa903a GX	161
	162	if (le32_to_cpu(*blkaddr) == EROFS_NULL_ADDR) {
	163	map->m_flags = 0;
	164	} else {
	165	map->m_pa = blknr_to_addr(le32_to_cpu(*blkaddr));
	166	map->m_flags = EROFS_MAP_MAPPED;
	167	}
	168	goto out_unlock;
	169	}
	170	/* parse chunk indexes */
fdf80a47	171	idx = kaddr + erofs_blkoff(pos);
c5aa903a GX	172	switch (le32_to_cpu(idx->blkaddr)) {
	173	case EROFS_NULL_ADDR:
	174	map->m_flags = 0;
	175	break;
	176	default:
dfeab2e9 GX	177	map->m_deviceid = le16_to_cpu(idx->device_id) &
dfeab2e9 GX	178	EROFS_SB(sb)->device_id_mask;
c5aa903a GX	179	map->m_pa = blknr_to_addr(le32_to_cpu(idx->blkaddr));
	180	map->m_flags = EROFS_MAP_MAPPED;
	181	break;
	182	}
	183	out_unlock:
fdf80a47	184	erofs_put_metabuf(&buf);
c5aa903a	185	out:
469407a3 GX	186	if (!err)
	187	map->m_llen = map->m_plen;
	188	trace_erofs_map_blocks_exit(inode, map, flags, 0);
c5aa903a GX	189	return err;
	190	}
	191
dfeab2e9 GX	192	int erofs_map_dev(struct super_block sb, struct erofs_map_dev map)
	193	{
	194	struct erofs_dev_context *devs = EROFS_SB(sb)->devs;
	195	struct erofs_device_info *dif;
	196	int id;
	197
	198	/* primary device by default */
	199	map->m_bdev = sb->s_bdev;
	200	map->m_daxdev = EROFS_SB(sb)->dax_dev;
de205114	201	map->m_dax_part_off = EROFS_SB(sb)->dax_part_off;
dfeab2e9 GX	202
	203	if (map->m_deviceid) {
	204	down_read(&devs->rwsem);
	205	dif = idr_find(&devs->tree, map->m_deviceid - 1);
	206	if (!dif) {
	207	up_read(&devs->rwsem);
	208	return -ENODEV;
	209	}
	210	map->m_bdev = dif->bdev;
	211	map->m_daxdev = dif->dax_dev;
de205114	212	map->m_dax_part_off = dif->dax_part_off;
dfeab2e9 GX	213	up_read(&devs->rwsem);
	214	} else if (devs->extra_devices) {
	215	down_read(&devs->rwsem);
	216	idr_for_each_entry(&devs->tree, dif, id) {
	217	erofs_off_t startoff, length;
	218
	219	if (!dif->mapped_blkaddr)
	220	continue;
	221	startoff = blknr_to_addr(dif->mapped_blkaddr);
	222	length = blknr_to_addr(dif->blocks);
	223
	224	if (map->m_pa >= startoff &&
	225	map->m_pa < startoff + length) {
	226	map->m_pa -= startoff;
	227	map->m_bdev = dif->bdev;
	228	map->m_daxdev = dif->dax_dev;
de205114	229	map->m_dax_part_off = dif->dax_part_off;
dfeab2e9 GX	230	break;
	231	}
	232	}
	233	up_read(&devs->rwsem);
	234	}
	235	return 0;
	236	}
	237
a08e67a0 HJ	238	static int erofs_iomap_begin(struct inode *inode, loff_t offset, loff_t length,
	239	unsigned int flags, struct iomap iomap, struct iomap srcmap)
	240	{
	241	int ret;
	242	struct erofs_map_blocks map;
dfeab2e9	243	struct erofs_map_dev mdev;
a08e67a0 HJ	244
	245	map.m_la = offset;
	246	map.m_llen = length;
	247
c5aa903a	248	ret = erofs_map_blocks(inode, &map, EROFS_GET_BLOCKS_RAW);
a08e67a0 HJ	249	if (ret < 0)
	250	return ret;
	251
dfeab2e9 GX	252	mdev = (struct erofs_map_dev) {
	253	.m_deviceid = map.m_deviceid,
	254	.m_pa = map.m_pa,
	255	};
	256	ret = erofs_map_dev(inode->i_sb, &mdev);
	257	if (ret)
	258	return ret;
	259
a08e67a0	260	iomap->offset = map.m_la;
e33f42b2	261	if (flags & IOMAP_DAX)
de205114	262	iomap->dax_dev = mdev.m_daxdev;
e33f42b2	263	else
de205114	264	iomap->bdev = mdev.m_bdev;
a08e67a0 HJ	265	iomap->length = map.m_llen;
a08e67a0 HJ	266	iomap->flags = 0;
771c994e	267	iomap->private = NULL;
a08e67a0 HJ	268
	269	if (!(map.m_flags & EROFS_MAP_MAPPED)) {
	270	iomap->type = IOMAP_HOLE;
	271	iomap->addr = IOMAP_NULL_ADDR;
	272	if (!iomap->length)
	273	iomap->length = length;
	274	return 0;
	275	}
	276
a08e67a0	277	if (map.m_flags & EROFS_MAP_META) {
fdf80a47 GX	278	void *ptr;
fdf80a47 GX	279	struct erofs_buf buf = __EROFS_BUF_INITIALIZER;
771c994e GX	280
771c994e GX	281	iomap->type = IOMAP_INLINE;
fdf80a47 GX	282	ptr = erofs_read_metabuf(&buf, inode->i_sb,
	283	erofs_blknr(mdev.m_pa), EROFS_KMAP);
	284	if (IS_ERR(ptr))
	285	return PTR_ERR(ptr);
	286	iomap->inline_data = ptr + erofs_blkoff(mdev.m_pa);
	287	iomap->private = buf.base;
771c994e GX	288	} else {
771c994e GX	289	iomap->type = IOMAP_MAPPED;
dfeab2e9	290	iomap->addr = mdev.m_pa;
e33f42b2 GX	291	if (flags & IOMAP_DAX)
e33f42b2 GX	292	iomap->addr += mdev.m_dax_part_off;
a08e67a0	293	}
a08e67a0 HJ	294	return 0;
	295	}
	296
771c994e GX	297	static int erofs_iomap_end(struct inode *inode, loff_t pos, loff_t length,
	298	ssize_t written, unsigned int flags, struct iomap *iomap)
	299	{
fdf80a47 GX	300	void *ptr = iomap->private;
	301
	302	if (ptr) {
	303	struct erofs_buf buf = {
	304	.page = kmap_to_page(ptr),
	305	.base = ptr,
	306	.kmap_type = EROFS_KMAP,
	307	};
771c994e	308
771c994e	309	DBG_BUGON(iomap->type != IOMAP_INLINE);
fdf80a47	310	erofs_put_metabuf(&buf);
771c994e GX	311	} else {
	312	DBG_BUGON(iomap->type == IOMAP_INLINE);
	313	}
	314	return written;
	315	}
	316
a08e67a0 HJ	317	static const struct iomap_ops erofs_iomap_ops = {
a08e67a0 HJ	318	.iomap_begin = erofs_iomap_begin,
771c994e	319	.iomap_end = erofs_iomap_end,
a08e67a0 HJ	320	};
a08e67a0 HJ	321
eadcd6b5 GX	322	int erofs_fiemap(struct inode inode, struct fiemap_extent_info fieinfo,
	323	u64 start, u64 len)
	324	{
	325	if (erofs_inode_is_data_compressed(EROFS_I(inode)->datalayout)) {
	326	#ifdef CONFIG_EROFS_FS_ZIP
	327	return iomap_fiemap(inode, fieinfo, start, len,
	328	&z_erofs_iomap_report_ops);
	329	#else
	330	return -EOPNOTSUPP;
	331	#endif
	332	}
	333	return iomap_fiemap(inode, fieinfo, start, len, &erofs_iomap_ops);
	334	}
	335
771c994e GX	336	/*
	337	* since we dont have write or truncate flows, so no inode
	338	* locking needs to be held at the moment.
	339	*/
	340	static int erofs_readpage(struct file file, struct page page)
	341	{
	342	return iomap_readpage(page, &erofs_iomap_ops);
	343	}
	344
	345	static void erofs_readahead(struct readahead_control *rac)
	346	{
	347	return iomap_readahead(rac, &erofs_iomap_ops);
	348	}
	349
	350	static sector_t erofs_bmap(struct address_space *mapping, sector_t block)
	351	{
	352	return iomap_bmap(mapping, block, &erofs_iomap_ops);
	353	}
	354
a08e67a0 HJ	355	static int erofs_prepare_dio(struct kiocb iocb, struct iov_iter to)
	356	{
	357	struct inode *inode = file_inode(iocb->ki_filp);
	358	loff_t align = iocb->ki_pos \| iov_iter_count(to) \|
	359	iov_iter_alignment(to);
	360	struct block_device *bdev = inode->i_sb->s_bdev;
	361	unsigned int blksize_mask;
	362
	363	if (bdev)
	364	blksize_mask = (1 << ilog2(bdev_logical_block_size(bdev))) - 1;
	365	else
	366	blksize_mask = (1 << inode->i_blkbits) - 1;
	367
	368	if (align & blksize_mask)
	369	return -EINVAL;
a08e67a0 HJ	370	return 0;
	371	}
	372
	373	static ssize_t erofs_file_read_iter(struct kiocb iocb, struct iov_iter to)
	374	{
	375	/* no need taking (shared) inode lock since it's a ro filesystem */
	376	if (!iov_iter_count(to))
	377	return 0;
	378
06252e9c GX	379	#ifdef CONFIG_FS_DAX
	380	if (IS_DAX(iocb->ki_filp->f_mapping->host))
	381	return dax_iomap_rw(iocb, to, &erofs_iomap_ops);
	382	#endif
a08e67a0 HJ	383	if (iocb->ki_flags & IOCB_DIRECT) {
	384	int err = erofs_prepare_dio(iocb, to);
	385
	386	if (!err)
	387	return iomap_dio_rw(iocb, to, &erofs_iomap_ops,
4fdccaa0	388	NULL, 0, 0);
a08e67a0 HJ	389	if (err < 0)
	390	return err;
	391	}
	392	return filemap_read(iocb, to, 0);
	393	}
	394
81781b02 GX	395	/* for uncompressed (aligned) files and raw access for other files */
81781b02 GX	396	const struct address_space_operations erofs_raw_access_aops = {
771c994e GX	397	.readpage = erofs_readpage,
771c994e GX	398	.readahead = erofs_readahead,
9da681e0	399	.bmap = erofs_bmap,
a08e67a0 HJ	400	.direct_IO = noop_direct_IO,
	401	};
	402
06252e9c GX	403	#ifdef CONFIG_FS_DAX
	404	static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
	405	enum page_entry_size pe_size)
	406	{
	407	return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
	408	}
	409
	410	static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
	411	{
	412	return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
	413	}
	414
	415	static const struct vm_operations_struct erofs_dax_vm_ops = {
	416	.fault = erofs_dax_fault,
	417	.huge_fault = erofs_dax_huge_fault,
	418	};
	419
	420	static int erofs_file_mmap(struct file file, struct vm_area_struct vma)
	421	{
	422	if (!IS_DAX(file_inode(file)))
	423	return generic_file_readonly_mmap(file, vma);
	424
	425	if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE))
	426	return -EINVAL;
	427
	428	vma->vm_ops = &erofs_dax_vm_ops;
	429	vma->vm_flags \|= VM_HUGEPAGE;
	430	return 0;
	431	}
	432	#else
	433	#define erofs_file_mmap generic_file_readonly_mmap
	434	#endif
	435
a08e67a0 HJ	436	const struct file_operations erofs_file_fops = {
	437	.llseek = generic_file_llseek,
	438	.read_iter = erofs_file_read_iter,
06252e9c	439	.mmap = erofs_file_mmap,
a08e67a0	440	.splice_read = generic_file_splice_read,
81781b02	441	};