[thirdparty/kernel/linux.git] / fs / ext4 / readpage.c

// SPDX-License-Identifier: GPL-2.0
/*
 * linux/fs/ext4/readpage.c
 *
 * Copyright (C) 2002, Linus Torvalds.
 * Copyright (C) 2015, Google, Inc.
 *
 * This was originally taken from fs/mpage.c
 *
 * The ext4_mpage_readpages() function here is intended to
 * replace mpage_readahead() in the general case, not just for
 * encrypted files.  It has some limitations (see below), where it
 * will fall back to read_block_full_page(), but these limitations
 * should only be hit when page_size != block_size.
 *
 * This will allow us to attach a callback function to support ext4
 * encryption.
 *
 * If anything unusual happens, such as:
 *
 * - encountering a page which has buffers
 * - encountering a page which has a non-hole after a hole
 * - encountering a page with non-contiguous blocks
 *
 * then this code just gives up and calls the buffer_head-based read function.
 * It does handle a page which has holes at the end - that is a common case:
 * the end-of-file on blocksize < PAGE_SIZE setups.
 *
 */

#include <linux/kernel.h>
#include <linux/export.h>
#include <linux/mm.h>
#include <linux/kdev_t.h>
#include <linux/gfp.h>
#include <linux/bio.h>
#include <linux/fs.h>
#include <linux/buffer_head.h>
#include <linux/blkdev.h>
#include <linux/highmem.h>
#include <linux/prefetch.h>
#include <linux/mpage.h>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
#include <linux/pagevec.h>

#include "ext4.h"

#define NUM_PREALLOC_POST_READ_CTXS	128

static struct kmem_cache *bio_post_read_ctx_cache;
static mempool_t *bio_post_read_ctx_pool;

/* postprocessing steps for read bios */
enum bio_post_read_step {
	STEP_INITIAL = 0,
	STEP_DECRYPT,
	STEP_VERITY,
	STEP_MAX,
};

struct bio_post_read_ctx {
	struct bio *bio;
	struct work_struct work;
	unsigned int cur_step;
	unsigned int enabled_steps;
};

static void __read_end_io(struct bio *bio)
{
	struct folio_iter fi;

	bio_for_each_folio_all(fi, bio)
		folio_end_read(fi.folio, bio->bi_status == 0);
	if (bio->bi_private)
		mempool_free(bio->bi_private, bio_post_read_ctx_pool);
	bio_put(bio);
}

static void bio_post_read_processing(struct bio_post_read_ctx *ctx);

static void decrypt_work(struct work_struct *work)
{
	struct bio_post_read_ctx *ctx =
		container_of(work, struct bio_post_read_ctx, work);
	struct bio *bio = ctx->bio;

	if (fscrypt_decrypt_bio(bio))
		bio_post_read_processing(ctx);
	else
		__read_end_io(bio);
}

static void verity_work(struct work_struct *work)
{
	struct bio_post_read_ctx *ctx =
		container_of(work, struct bio_post_read_ctx, work);
	struct bio *bio = ctx->bio;

	/*
	 * fsverity_verify_bio() may call readahead() again, and although verity
	 * will be disabled for that, decryption may still be needed, causing
	 * another bio_post_read_ctx to be allocated.  So to guarantee that
	 * mempool_alloc() never deadlocks we must free the current ctx first.
	 * This is safe because verity is the last post-read step.
	 */
	BUILD_BUG_ON(STEP_VERITY + 1 != STEP_MAX);
	mempool_free(ctx, bio_post_read_ctx_pool);
	bio->bi_private = NULL;

	fsverity_verify_bio(bio);

	__read_end_io(bio);
}

static void bio_post_read_processing(struct bio_post_read_ctx *ctx)
{
	/*
	 * We use different work queues for decryption and for verity because
	 * verity may require reading metadata pages that need decryption, and
	 * we shouldn't recurse to the same workqueue.
	 */
	switch (++ctx->cur_step) {
	case STEP_DECRYPT:
		if (ctx->enabled_steps & (1 << STEP_DECRYPT)) {
			INIT_WORK(&ctx->work, decrypt_work);
			fscrypt_enqueue_decrypt_work(&ctx->work);
			return;
		}
		ctx->cur_step++;
		fallthrough;
	case STEP_VERITY:
		if (ctx->enabled_steps & (1 << STEP_VERITY)) {
			INIT_WORK(&ctx->work, verity_work);
			fsverity_enqueue_verify_work(&ctx->work);
			return;
		}
		ctx->cur_step++;
		fallthrough;
	default:
		__read_end_io(ctx->bio);
	}
}

static bool bio_post_read_required(struct bio *bio)
{
	return bio->bi_private && !bio->bi_status;
}

/*
 * I/O completion handler for multipage BIOs.
 *
 * The mpage code never puts partial pages into a BIO (except for end-of-file).
 * If a page does not map to a contiguous run of blocks then it simply falls
 * back to block_read_full_folio().
 *
 * Why is this?  If a page's completion depends on a number of different BIOs
 * which can complete in any order (or at the same time) then determining the
 * status of that page is hard.  See end_buffer_async_read() for the details.
 * There is no point in duplicating all that complexity.
 */
static void mpage_end_io(struct bio *bio)
{
	if (bio_post_read_required(bio)) {
		struct bio_post_read_ctx *ctx = bio->bi_private;

		ctx->cur_step = STEP_INITIAL;
		bio_post_read_processing(ctx);
		return;
	}
	__read_end_io(bio);
}

static inline bool ext4_need_verity(const struct inode *inode, pgoff_t idx)
{
	return fsverity_active(inode) &&
	       idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
}

static void ext4_set_bio_post_read_ctx(struct bio *bio,
				       const struct inode *inode,
				       pgoff_t first_idx)
{
	unsigned int post_read_steps = 0;

	if (fscrypt_inode_uses_fs_layer_crypto(inode))
		post_read_steps |= 1 << STEP_DECRYPT;

	if (ext4_need_verity(inode, first_idx))
		post_read_steps |= 1 << STEP_VERITY;

	if (post_read_steps) {
		/* Due to the mempool, this never fails. */
		struct bio_post_read_ctx *ctx =
			mempool_alloc(bio_post_read_ctx_pool, GFP_NOFS);

		ctx->bio = bio;
		ctx->enabled_steps = post_read_steps;
		bio->bi_private = ctx;
	}
}

static inline loff_t ext4_readpage_limit(struct inode *inode)
{
	if (IS_ENABLED(CONFIG_FS_VERITY) && IS_VERITY(inode))
		return inode->i_sb->s_maxbytes;

	return i_size_read(inode);
}

int ext4_mpage_readpages(struct inode *inode,
		struct readahead_control *rac, struct folio *folio)
{
	struct bio *bio = NULL;
	sector_t last_block_in_bio = 0;

	const unsigned blkbits = inode->i_blkbits;
	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
	const unsigned blocksize = 1 << blkbits;
	sector_t next_block;
	sector_t block_in_file;
	sector_t last_block;
	sector_t last_block_in_file;
	sector_t first_block;
	unsigned page_block;
	struct block_device *bdev = inode->i_sb->s_bdev;
	int length;
	unsigned relative_block = 0;
	struct ext4_map_blocks map;
	unsigned int nr_pages, folio_pages;

	map.m_pblk = 0;
	map.m_lblk = 0;
	map.m_len = 0;
	map.m_flags = 0;

	nr_pages = rac ? readahead_count(rac) : folio_nr_pages(folio);
	for (; nr_pages; nr_pages -= folio_pages) {
		int fully_mapped = 1;
		unsigned int first_hole;
		unsigned int blocks_per_folio;

		if (rac)
			folio = readahead_folio(rac);

		folio_pages = folio_nr_pages(folio);
		prefetchw(&folio->flags);

		if (folio_buffers(folio))
			goto confused;

		blocks_per_folio = folio_size(folio) >> blkbits;
		first_hole = blocks_per_folio;
		block_in_file = next_block =
			(sector_t)folio->index << (PAGE_SHIFT - blkbits);
		last_block = block_in_file + nr_pages * blocks_per_page;
		last_block_in_file = (ext4_readpage_limit(inode) +
				      blocksize - 1) >> blkbits;
		if (last_block > last_block_in_file)
			last_block = last_block_in_file;
		page_block = 0;

		/*
		 * Map blocks using the previous result first.
		 */
		if ((map.m_flags & EXT4_MAP_MAPPED) &&
		    block_in_file > map.m_lblk &&
		    block_in_file < (map.m_lblk + map.m_len)) {
			unsigned map_offset = block_in_file - map.m_lblk;
			unsigned last = map.m_len - map_offset;

			first_block = map.m_pblk + map_offset;
			for (relative_block = 0; ; relative_block++) {
				if (relative_block == last) {
					/* needed? */
					map.m_flags &= ~EXT4_MAP_MAPPED;
					break;
				}
				if (page_block == blocks_per_folio)
					break;
				page_block++;
				block_in_file++;
			}
		}

		/*
		 * Then do more ext4_map_blocks() calls until we are
		 * done with this folio.
		 */
		while (page_block < blocks_per_folio) {
			if (block_in_file < last_block) {
				map.m_lblk = block_in_file;
				map.m_len = last_block - block_in_file;

				if (ext4_map_blocks(NULL, inode, &map, 0) < 0) {
				set_error_page:
					folio_zero_segment(folio, 0,
							  folio_size(folio));
					folio_unlock(folio);
					goto next_page;
				}
			}
			if ((map.m_flags & EXT4_MAP_MAPPED) == 0) {
				fully_mapped = 0;
				if (first_hole == blocks_per_folio)
					first_hole = page_block;
				page_block++;
				block_in_file++;
				continue;
			}
			if (first_hole != blocks_per_folio)
				goto confused;		/* hole -> non-hole */

			/* Contiguous blocks? */
			if (!page_block)
				first_block = map.m_pblk;
			else if (first_block + page_block != map.m_pblk)
				goto confused;
			for (relative_block = 0; ; relative_block++) {
				if (relative_block == map.m_len) {
					/* needed? */
					map.m_flags &= ~EXT4_MAP_MAPPED;
					break;
				} else if (page_block == blocks_per_folio)
					break;
				page_block++;
				block_in_file++;
			}
		}
		if (first_hole != blocks_per_folio) {
			folio_zero_segment(folio, first_hole << blkbits,
					  folio_size(folio));
			if (first_hole == 0) {
				if (ext4_need_verity(inode, folio->index) &&
				    !fsverity_verify_folio(folio))
					goto set_error_page;
				folio_end_read(folio, true);
				continue;
			}
		} else if (fully_mapped) {
			folio_set_mappedtodisk(folio);
		}

		/*
		 * This folio will go to BIO.  Do we need to send this
		 * BIO off first?
		 */
		if (bio && (last_block_in_bio != first_block - 1 ||
			    !fscrypt_mergeable_bio(bio, inode, next_block))) {
		submit_and_realloc:
			submit_bio(bio);
			bio = NULL;
		}
		if (bio == NULL) {
			/*
			 * bio_alloc will _always_ be able to allocate a bio if
			 * __GFP_DIRECT_RECLAIM is set, see bio_alloc_bioset().
			 */
			bio = bio_alloc(bdev, bio_max_segs(nr_pages),
					REQ_OP_READ, GFP_KERNEL);
			fscrypt_set_bio_crypt_ctx(bio, inode, next_block,
						  GFP_KERNEL);
			ext4_set_bio_post_read_ctx(bio, inode, folio->index);
			bio->bi_iter.bi_sector = first_block << (blkbits - 9);
			bio->bi_end_io = mpage_end_io;
			if (rac)
				bio->bi_opf |= REQ_RAHEAD;
		}

		length = first_hole << blkbits;
		if (!bio_add_folio(bio, folio, length, 0))
			goto submit_and_realloc;

		if (((map.m_flags & EXT4_MAP_BOUNDARY) &&
		     (relative_block == map.m_len)) ||
		    (first_hole != blocks_per_folio)) {
			submit_bio(bio);
			bio = NULL;
		} else
			last_block_in_bio = first_block + blocks_per_folio - 1;
		continue;
	confused:
		if (bio) {
			submit_bio(bio);
			bio = NULL;
		}
		if (!folio_test_uptodate(folio))
			block_read_full_folio(folio, ext4_get_block);
		else
			folio_unlock(folio);
next_page:
		; /* A label shall be followed by a statement until C23 */
	}
	if (bio)
		submit_bio(bio);
	return 0;
}

int __init ext4_init_post_read_processing(void)
{
	bio_post_read_ctx_cache = KMEM_CACHE(bio_post_read_ctx, SLAB_RECLAIM_ACCOUNT);

	if (!bio_post_read_ctx_cache)
		goto fail;
	bio_post_read_ctx_pool =
		mempool_create_slab_pool(NUM_PREALLOC_POST_READ_CTXS,
					 bio_post_read_ctx_cache);
	if (!bio_post_read_ctx_pool)
		goto fail_free_cache;
	return 0;

fail_free_cache:
	kmem_cache_destroy(bio_post_read_ctx_cache);
fail:
	return -ENOMEM;
}

void ext4_exit_post_read_processing(void)
{
	mempool_destroy(bio_post_read_ctx_pool);
	kmem_cache_destroy(bio_post_read_ctx_cache);
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
f64e02fe TT	2	/*
	3	* linux/fs/ext4/readpage.c
	4	*
	5	* Copyright (C) 2002, Linus Torvalds.
	6	* Copyright (C) 2015, Google, Inc.
	7	*
	8	* This was originally taken from fs/mpage.c
	9	*
6311f91f MWO	10	* The ext4_mpage_readpages() function here is intended to
6311f91f MWO	11	* replace mpage_readahead() in the general case, not just for
f64e02fe TT	12	* encrypted files. It has some limitations (see below), where it
	13	* will fall back to read_block_full_page(), but these limitations
	14	* should only be hit when page_size != block_size.
	15	*
	16	* This will allow us to attach a callback function to support ext4
	17	* encryption.
	18	*
	19	* If anything unusual happens, such as:
	20	*
	21	* - encountering a page which has buffers
	22	* - encountering a page which has a non-hole after a hole
	23	* - encountering a page with non-contiguous blocks
	24	*
	25	* then this code just gives up and calls the buffer_head-based read function.
	26	* It does handle a page which has holes at the end - that is a common case:
ea1754a0	27	* the end-of-file on blocksize < PAGE_SIZE setups.
f64e02fe TT	28	*
	29	*/
	30
	31	#include <linux/kernel.h>
	32	#include <linux/export.h>
	33	#include <linux/mm.h>
	34	#include <linux/kdev_t.h>
	35	#include <linux/gfp.h>
	36	#include <linux/bio.h>
	37	#include <linux/fs.h>
	38	#include <linux/buffer_head.h>
	39	#include <linux/blkdev.h>
	40	#include <linux/highmem.h>
	41	#include <linux/prefetch.h>
	42	#include <linux/mpage.h>
	43	#include <linux/writeback.h>
	44	#include <linux/backing-dev.h>
	45	#include <linux/pagevec.h>
f64e02fe TT	46
	47	#include "ext4.h"
	48
22cfe4b4 EB	49	#define NUM_PREALLOC_POST_READ_CTXS 128
	50
	51	static struct kmem_cache *bio_post_read_ctx_cache;
	52	static mempool_t *bio_post_read_ctx_pool;
	53
	54	/* postprocessing steps for read bios */
	55	enum bio_post_read_step {
	56	STEP_INITIAL = 0,
	57	STEP_DECRYPT,
	58	STEP_VERITY,
68e45330	59	STEP_MAX,
22cfe4b4 EB	60	};
	61
	62	struct bio_post_read_ctx {
	63	struct bio *bio;
	64	struct work_struct work;
	65	unsigned int cur_step;
	66	unsigned int enabled_steps;
	67	};
	68
	69	static void __read_end_io(struct bio *bio)
c9c7429c	70	{
f2b229a8	71	struct folio_iter fi;
22cfe4b4	72
f8174a11 MWO	73	bio_for_each_folio_all(fi, bio)
f8174a11 MWO	74	folio_end_read(fi.folio, bio->bi_status == 0);
22cfe4b4 EB	75	if (bio->bi_private)
	76	mempool_free(bio->bi_private, bio_post_read_ctx_pool);
	77	bio_put(bio);
	78	}
	79
	80	static void bio_post_read_processing(struct bio_post_read_ctx *ctx);
	81
	82	static void decrypt_work(struct work_struct *work)
	83	{
	84	struct bio_post_read_ctx *ctx =
	85	container_of(work, struct bio_post_read_ctx, work);
14db0b3c	86	struct bio *bio = ctx->bio;
22cfe4b4	87
14db0b3c EB	88	if (fscrypt_decrypt_bio(bio))
	89	bio_post_read_processing(ctx);
	90	else
	91	__read_end_io(bio);
22cfe4b4 EB	92	}
	93
	94	static void verity_work(struct work_struct *work)
	95	{
	96	struct bio_post_read_ctx *ctx =
	97	container_of(work, struct bio_post_read_ctx, work);
68e45330	98	struct bio *bio = ctx->bio;
22cfe4b4	99
68e45330	100	/*
704528d8	101	* fsverity_verify_bio() may call readahead() again, and although verity
68e45330 EB	102	* will be disabled for that, decryption may still be needed, causing
	103	* another bio_post_read_ctx to be allocated. So to guarantee that
	104	* mempool_alloc() never deadlocks we must free the current ctx first.
	105	* This is safe because verity is the last post-read step.
	106	*/
	107	BUILD_BUG_ON(STEP_VERITY + 1 != STEP_MAX);
	108	mempool_free(ctx, bio_post_read_ctx_pool);
	109	bio->bi_private = NULL;
	110
	111	fsverity_verify_bio(bio);
	112
	113	__read_end_io(bio);
22cfe4b4 EB	114	}
	115
	116	static void bio_post_read_processing(struct bio_post_read_ctx *ctx)
	117	{
	118	/*
	119	* We use different work queues for decryption and for verity because
	120	* verity may require reading metadata pages that need decryption, and
	121	* we shouldn't recurse to the same workqueue.
	122	*/
	123	switch (++ctx->cur_step) {
	124	case STEP_DECRYPT:
	125	if (ctx->enabled_steps & (1 << STEP_DECRYPT)) {
	126	INIT_WORK(&ctx->work, decrypt_work);
	127	fscrypt_enqueue_decrypt_work(&ctx->work);
	128	return;
	129	}
	130	ctx->cur_step++;
70d7ced2	131	fallthrough;
22cfe4b4 EB	132	case STEP_VERITY:
	133	if (ctx->enabled_steps & (1 << STEP_VERITY)) {
	134	INIT_WORK(&ctx->work, verity_work);
	135	fsverity_enqueue_verify_work(&ctx->work);
	136	return;
	137	}
	138	ctx->cur_step++;
70d7ced2	139	fallthrough;
22cfe4b4 EB	140	default:
	141	__read_end_io(ctx->bio);
	142	}
	143	}
	144
	145	static bool bio_post_read_required(struct bio *bio)
	146	{
	147	return bio->bi_private && !bio->bi_status;
c9c7429c MH	148	}
c9c7429c MH	149
f64e02fe TT	150	/*
	151	* I/O completion handler for multipage BIOs.
	152	*
	153	* The mpage code never puts partial pages into a BIO (except for end-of-file).
	154	* If a page does not map to a contiguous run of blocks then it simply falls
2c69e205	155	* back to block_read_full_folio().
f64e02fe TT	156	*
	157	* Why is this? If a page's completion depends on a number of different BIOs
	158	* which can complete in any order (or at the same time) then determining the
	159	* status of that page is hard. See end_buffer_async_read() for the details.
	160	* There is no point in duplicating all that complexity.
	161	*/
4246a0b6	162	static void mpage_end_io(struct bio *bio)
f64e02fe	163	{
22cfe4b4 EB	164	if (bio_post_read_required(bio)) {
22cfe4b4 EB	165	struct bio_post_read_ctx *ctx = bio->bi_private;
f64e02fe	166
22cfe4b4 EB	167	ctx->cur_step = STEP_INITIAL;
	168	bio_post_read_processing(ctx);
	169	return;
c9c7429c	170	}
22cfe4b4 EB	171	__read_end_io(bio);
22cfe4b4 EB	172	}
f64e02fe	173
22cfe4b4 EB	174	static inline bool ext4_need_verity(const struct inode *inode, pgoff_t idx)
	175	{
	176	return fsverity_active(inode) &&
	177	idx < DIV_ROUND_UP(inode->i_size, PAGE_SIZE);
	178	}
	179
fd5fe253 EB	180	static void ext4_set_bio_post_read_ctx(struct bio *bio,
	181	const struct inode *inode,
	182	pgoff_t first_idx)
22cfe4b4 EB	183	{
22cfe4b4 EB	184	unsigned int post_read_steps = 0;
22cfe4b4	185
4f74d15f	186	if (fscrypt_inode_uses_fs_layer_crypto(inode))
22cfe4b4 EB	187	post_read_steps \|= 1 << STEP_DECRYPT;
	188
	189	if (ext4_need_verity(inode, first_idx))
	190	post_read_steps \|= 1 << STEP_VERITY;
	191
	192	if (post_read_steps) {
fd5fe253 EB	193	/* Due to the mempool, this never fails. */
	194	struct bio_post_read_ctx *ctx =
	195	mempool_alloc(bio_post_read_ctx_pool, GFP_NOFS);
	196
22cfe4b4 EB	197	ctx->bio = bio;
	198	ctx->enabled_steps = post_read_steps;
	199	bio->bi_private = ctx;
f64e02fe	200	}
22cfe4b4	201	}
f64e02fe	202
22cfe4b4 EB	203	static inline loff_t ext4_readpage_limit(struct inode *inode)
22cfe4b4 EB	204	{
5e122148	205	if (IS_ENABLED(CONFIG_FS_VERITY) && IS_VERITY(inode))
22cfe4b4 EB	206	return inode->i_sb->s_maxbytes;
	207
	208	return i_size_read(inode);
f64e02fe TT	209	}
f64e02fe TT	210
a07f624b	211	int ext4_mpage_readpages(struct inode *inode,
c0be8e6f	212	struct readahead_control rac, struct folio folio)
f64e02fe TT	213	{
f64e02fe TT	214	struct bio *bio = NULL;
f64e02fe TT	215	sector_t last_block_in_bio = 0;
f64e02fe TT	216
f64e02fe	217	const unsigned blkbits = inode->i_blkbits;
09cbfeaf	218	const unsigned blocks_per_page = PAGE_SIZE >> blkbits;
f64e02fe	219	const unsigned blocksize = 1 << blkbits;
4f74d15f	220	sector_t next_block;
f64e02fe TT	221	sector_t block_in_file;
	222	sector_t last_block;
	223	sector_t last_block_in_file;
e37c9e17	224	sector_t first_block;
f64e02fe TT	225	unsigned page_block;
	226	struct block_device *bdev = inode->i_sb->s_bdev;
	227	int length;
	228	unsigned relative_block = 0;
	229	struct ext4_map_blocks map;
fdbd0df9	230	unsigned int nr_pages, folio_pages;
f64e02fe TT	231
	232	map.m_pblk = 0;
	233	map.m_lblk = 0;
	234	map.m_len = 0;
	235	map.m_flags = 0;
	236
fdbd0df9 ZY	237	nr_pages = rac ? readahead_count(rac) : folio_nr_pages(folio);
fdbd0df9 ZY	238	for (; nr_pages; nr_pages -= folio_pages) {
f64e02fe	239	int fully_mapped = 1;
fdbd0df9 ZY	240	unsigned int first_hole;
fdbd0df9 ZY	241	unsigned int blocks_per_folio;
f64e02fe	242
c0be8e6f MW	243	if (rac)
c0be8e6f MW	244	folio = readahead_folio(rac);
fdbd0df9 ZY	245
fdbd0df9 ZY	246	folio_pages = folio_nr_pages(folio);
c0be8e6f	247	prefetchw(&folio->flags);
f64e02fe	248
c0be8e6f	249	if (folio_buffers(folio))
f64e02fe TT	250	goto confused;
f64e02fe TT	251
fdbd0df9 ZY	252	blocks_per_folio = folio_size(folio) >> blkbits;
fdbd0df9 ZY	253	first_hole = blocks_per_folio;
4f74d15f	254	block_in_file = next_block =
c0be8e6f	255	(sector_t)folio->index << (PAGE_SHIFT - blkbits);
f64e02fe	256	last_block = block_in_file + nr_pages * blocks_per_page;
22cfe4b4 EB	257	last_block_in_file = (ext4_readpage_limit(inode) +
22cfe4b4 EB	258	blocksize - 1) >> blkbits;
f64e02fe TT	259	if (last_block > last_block_in_file)
	260	last_block = last_block_in_file;
	261	page_block = 0;
	262
	263	/*
	264	* Map blocks using the previous result first.
	265	*/
	266	if ((map.m_flags & EXT4_MAP_MAPPED) &&
	267	block_in_file > map.m_lblk &&
	268	block_in_file < (map.m_lblk + map.m_len)) {
	269	unsigned map_offset = block_in_file - map.m_lblk;
	270	unsigned last = map.m_len - map_offset;
	271
e37c9e17	272	first_block = map.m_pblk + map_offset;
f64e02fe TT	273	for (relative_block = 0; ; relative_block++) {
	274	if (relative_block == last) {
	275	/* needed? */
	276	map.m_flags &= ~EXT4_MAP_MAPPED;
	277	break;
	278	}
fdbd0df9	279	if (page_block == blocks_per_folio)
f64e02fe	280	break;
f64e02fe TT	281	page_block++;
	282	block_in_file++;
	283	}
	284	}
	285
	286	/*
	287	* Then do more ext4_map_blocks() calls until we are
c0be8e6f	288	* done with this folio.
f64e02fe	289	*/
fdbd0df9	290	while (page_block < blocks_per_folio) {
f64e02fe TT	291	if (block_in_file < last_block) {
	292	map.m_lblk = block_in_file;
	293	map.m_len = last_block - block_in_file;
	294
	295	if (ext4_map_blocks(NULL, inode, &map, 0) < 0) {
	296	set_error_page:
c0be8e6f MW	297	folio_zero_segment(folio, 0,
	298	folio_size(folio));
	299	folio_unlock(folio);
f64e02fe TT	300	goto next_page;
	301	}
	302	}
	303	if ((map.m_flags & EXT4_MAP_MAPPED) == 0) {
	304	fully_mapped = 0;
fdbd0df9	305	if (first_hole == blocks_per_folio)
f64e02fe TT	306	first_hole = page_block;
	307	page_block++;
	308	block_in_file++;
	309	continue;
	310	}
fdbd0df9	311	if (first_hole != blocks_per_folio)
f64e02fe TT	312	goto confused; /* hole -> non-hole */
	313
	314	/* Contiguous blocks? */
e37c9e17 MWO	315	if (!page_block)
	316	first_block = map.m_pblk;
	317	else if (first_block + page_block != map.m_pblk)
f64e02fe TT	318	goto confused;
	319	for (relative_block = 0; ; relative_block++) {
	320	if (relative_block == map.m_len) {
	321	/* needed? */
	322	map.m_flags &= ~EXT4_MAP_MAPPED;
	323	break;
fdbd0df9	324	} else if (page_block == blocks_per_folio)
f64e02fe	325	break;
f64e02fe TT	326	page_block++;
	327	block_in_file++;
	328	}
	329	}
fdbd0df9	330	if (first_hole != blocks_per_folio) {
c0be8e6f MW	331	folio_zero_segment(folio, first_hole << blkbits,
c0be8e6f MW	332	folio_size(folio));
f64e02fe	333	if (first_hole == 0) {
c0be8e6f	334	if (ext4_need_verity(inode, folio->index) &&
0dea40aa	335	!fsverity_verify_folio(folio))
22cfe4b4	336	goto set_error_page;
f8174a11	337	folio_end_read(folio, true);
c0be8e6f	338	continue;
f64e02fe TT	339	}
f64e02fe TT	340	} else if (fully_mapped) {
c0be8e6f	341	folio_set_mappedtodisk(folio);
f64e02fe	342	}
f64e02fe TT	343
f64e02fe TT	344	/*
c0be8e6f	345	* This folio will go to BIO. Do we need to send this
f64e02fe TT	346	* BIO off first?
f64e02fe TT	347	*/
e37c9e17	348	if (bio && (last_block_in_bio != first_block - 1 \|\|
4f74d15f	349	!fscrypt_mergeable_bio(bio, inode, next_block))) {
f64e02fe	350	submit_and_realloc:
4e49ea4a	351	submit_bio(bio);
f64e02fe TT	352	bio = NULL;
	353	}
	354	if (bio == NULL) {
5500221e GX	355	/*
	356	* bio_alloc will _always_ be able to allocate a bio if
	357	* __GFP_DIRECT_RECLAIM is set, see bio_alloc_bioset().
	358	*/
07888c66 CH	359	bio = bio_alloc(bdev, bio_max_segs(nr_pages),
07888c66 CH	360	REQ_OP_READ, GFP_KERNEL);
4f74d15f EB	361	fscrypt_set_bio_crypt_ctx(bio, inode, next_block,
4f74d15f EB	362	GFP_KERNEL);
c0be8e6f	363	ext4_set_bio_post_read_ctx(bio, inode, folio->index);
e37c9e17	364	bio->bi_iter.bi_sector = first_block << (blkbits - 9);
f64e02fe	365	bio->bi_end_io = mpage_end_io;
07888c66 CH	366	if (rac)
07888c66 CH	367	bio->bi_opf \|= REQ_RAHEAD;
f64e02fe TT	368	}
	369
	370	length = first_hole << blkbits;
c0be8e6f	371	if (!bio_add_folio(bio, folio, length, 0))
f64e02fe TT	372	goto submit_and_realloc;
	373
	374	if (((map.m_flags & EXT4_MAP_BOUNDARY) &&
	375	(relative_block == map.m_len)) \|\|
fdbd0df9	376	(first_hole != blocks_per_folio)) {
4e49ea4a	377	submit_bio(bio);
f64e02fe TT	378	bio = NULL;
f64e02fe TT	379	} else
fdbd0df9	380	last_block_in_bio = first_block + blocks_per_folio - 1;
c0be8e6f	381	continue;
f64e02fe TT	382	confused:
f64e02fe TT	383	if (bio) {
4e49ea4a	384	submit_bio(bio);
f64e02fe TT	385	bio = NULL;
f64e02fe TT	386	}
c0be8e6f MW	387	if (!folio_test_uptodate(folio))
c0be8e6f MW	388	block_read_full_folio(folio, ext4_get_block);
f64e02fe	389	else
c0be8e6f MW	390	folio_unlock(folio);
	391	next_page:
	392	; /* A label shall be followed by a statement until C23 */
f64e02fe	393	}
f64e02fe	394	if (bio)
4e49ea4a	395	submit_bio(bio);
f64e02fe TT	396	return 0;
f64e02fe TT	397	}
22cfe4b4 EB	398
	399	int __init ext4_init_post_read_processing(void)
	400	{
060f7739 JS	401	bio_post_read_ctx_cache = KMEM_CACHE(bio_post_read_ctx, SLAB_RECLAIM_ACCOUNT);
060f7739 JS	402
22cfe4b4 EB	403	if (!bio_post_read_ctx_cache)
	404	goto fail;
	405	bio_post_read_ctx_pool =
	406	mempool_create_slab_pool(NUM_PREALLOC_POST_READ_CTXS,
	407	bio_post_read_ctx_cache);
	408	if (!bio_post_read_ctx_pool)
	409	goto fail_free_cache;
	410	return 0;
	411
	412	fail_free_cache:
	413	kmem_cache_destroy(bio_post_read_ctx_cache);
	414	fail:
	415	return -ENOMEM;
	416	}
	417
	418	void ext4_exit_post_read_processing(void)
	419	{
	420	mempool_destroy(bio_post_read_ctx_pool);
	421	kmem_cache_destroy(bio_post_read_ctx_cache);
	422	}