[linux-block.git] / mm / page_io.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  linux/mm/page_io.c
 *
 *  Copyright (C) 1991, 1992, 1993, 1994  Linus Torvalds
 *
 *  Swap reorganised 29.12.95, 
 *  Asynchronous swapping added 30.12.95. Stephen Tweedie
 *  Removed race in async swapping. 14.4.1996. Bruno Haible
 *  Add swap of shared pages through the page cache. 20.2.1998. Stephen Tweedie
 *  Always use brw_page, life becomes simpler. 12 May 1998 Eric Biederman
 */

#include <linux/mm.h>
#include <linux/kernel_stat.h>
#include <linux/gfp.h>
#include <linux/pagemap.h>
#include <linux/swap.h>
#include <linux/bio.h>
#include <linux/swapops.h>
#include <linux/buffer_head.h>
#include <linux/writeback.h>
#include <linux/frontswap.h>
#include <linux/blkdev.h>
#include <linux/psi.h>
#include <linux/uio.h>
#include <linux/sched/task.h>
#include <linux/delayacct.h>
#include "swap.h"

static void end_swap_bio_write(struct bio *bio)
{
	struct page *page = bio_first_page_all(bio);

	if (bio->bi_status) {
		SetPageError(page);
		/*
		 * We failed to write the page out to swap-space.
		 * Re-dirty the page in order to avoid it being reclaimed.
		 * Also print a dire warning that things will go BAD (tm)
		 * very quickly.
		 *
		 * Also clear PG_reclaim to avoid folio_rotate_reclaimable()
		 */
		set_page_dirty(page);
		pr_alert_ratelimited("Write-error on swap-device (%u:%u:%llu)\n",
				     MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
				     (unsigned long long)bio->bi_iter.bi_sector);
		ClearPageReclaim(page);
	}
	end_page_writeback(page);
	bio_put(bio);
}

static void end_swap_bio_read(struct bio *bio)
{
	struct page *page = bio_first_page_all(bio);
	struct task_struct *waiter = bio->bi_private;

	if (bio->bi_status) {
		SetPageError(page);
		ClearPageUptodate(page);
		pr_alert_ratelimited("Read-error on swap-device (%u:%u:%llu)\n",
				     MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
				     (unsigned long long)bio->bi_iter.bi_sector);
		goto out;
	}

	SetPageUptodate(page);
out:
	unlock_page(page);
	WRITE_ONCE(bio->bi_private, NULL);
	bio_put(bio);
	if (waiter) {
		blk_wake_io_task(waiter);
		put_task_struct(waiter);
	}
}

int generic_swapfile_activate(struct swap_info_struct *sis,
				struct file *swap_file,
				sector_t *span)
{
	struct address_space *mapping = swap_file->f_mapping;
	struct inode *inode = mapping->host;
	unsigned blocks_per_page;
	unsigned long page_no;
	unsigned blkbits;
	sector_t probe_block;
	sector_t last_block;
	sector_t lowest_block = -1;
	sector_t highest_block = 0;
	int nr_extents = 0;
	int ret;

	blkbits = inode->i_blkbits;
	blocks_per_page = PAGE_SIZE >> blkbits;

	/*
	 * Map all the blocks into the extent tree.  This code doesn't try
	 * to be very smart.
	 */
	probe_block = 0;
	page_no = 0;
	last_block = i_size_read(inode) >> blkbits;
	while ((probe_block + blocks_per_page) <= last_block &&
			page_no < sis->max) {
		unsigned block_in_page;
		sector_t first_block;

		cond_resched();

		first_block = probe_block;
		ret = bmap(inode, &first_block);
		if (ret || !first_block)
			goto bad_bmap;

		/*
		 * It must be PAGE_SIZE aligned on-disk
		 */
		if (first_block & (blocks_per_page - 1)) {
			probe_block++;
			goto reprobe;
		}

		for (block_in_page = 1; block_in_page < blocks_per_page;
					block_in_page++) {
			sector_t block;

			block = probe_block + block_in_page;
			ret = bmap(inode, &block);
			if (ret || !block)
				goto bad_bmap;

			if (block != first_block + block_in_page) {
				/* Discontiguity */
				probe_block++;
				goto reprobe;
			}
		}

		first_block >>= (PAGE_SHIFT - blkbits);
		if (page_no) {	/* exclude the header page */
			if (first_block < lowest_block)
				lowest_block = first_block;
			if (first_block > highest_block)
				highest_block = first_block;
		}

		/*
		 * We found a PAGE_SIZE-length, PAGE_SIZE-aligned run of blocks
		 */
		ret = add_swap_extent(sis, page_no, 1, first_block);
		if (ret < 0)
			goto out;
		nr_extents += ret;
		page_no++;
		probe_block += blocks_per_page;
reprobe:
		continue;
	}
	ret = nr_extents;
	*span = 1 + highest_block - lowest_block;
	if (page_no == 0)
		page_no = 1;	/* force Empty message */
	sis->max = page_no;
	sis->pages = page_no - 1;
	sis->highest_bit = page_no - 1;
out:
	return ret;
bad_bmap:
	pr_err("swapon: swapfile has holes\n");
	ret = -EINVAL;
	goto out;
}

/*
 * We may have stale swap cache pages in memory: notice
 * them here and get rid of the unnecessary final write.
 */
int swap_writepage(struct page *page, struct writeback_control *wbc)
{
	struct folio *folio = page_folio(page);
	int ret = 0;

	if (folio_free_swap(folio)) {
		folio_unlock(folio);
		goto out;
	}
	/*
	 * Arch code may have to preserve more data than just the page
	 * contents, e.g. memory tags.
	 */
	ret = arch_prepare_to_swap(&folio->page);
	if (ret) {
		folio_mark_dirty(folio);
		folio_unlock(folio);
		goto out;
	}
	if (frontswap_store(&folio->page) == 0) {
		folio_start_writeback(folio);
		folio_unlock(folio);
		folio_end_writeback(folio);
		goto out;
	}
	ret = __swap_writepage(&folio->page, wbc);
out:
	return ret;
}

static inline void count_swpout_vm_event(struct page *page)
{
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
	if (unlikely(PageTransHuge(page)))
		count_vm_event(THP_SWPOUT);
#endif
	count_vm_events(PSWPOUT, thp_nr_pages(page));
}

#if defined(CONFIG_MEMCG) && defined(CONFIG_BLK_CGROUP)
static void bio_associate_blkg_from_page(struct bio *bio, struct page *page)
{
	struct cgroup_subsys_state *css;
	struct mem_cgroup *memcg;

	memcg = page_memcg(page);
	if (!memcg)
		return;

	rcu_read_lock();
	css = cgroup_e_css(memcg->css.cgroup, &io_cgrp_subsys);
	bio_associate_blkg_from_css(bio, css);
	rcu_read_unlock();
}
#else
#define bio_associate_blkg_from_page(bio, page)		do { } while (0)
#endif /* CONFIG_MEMCG && CONFIG_BLK_CGROUP */

struct swap_iocb {
	struct kiocb		iocb;
	struct bio_vec		bvec[SWAP_CLUSTER_MAX];
	int			pages;
	int			len;
};
static mempool_t *sio_pool;

int sio_pool_init(void)
{
	if (!sio_pool) {
		mempool_t *pool = mempool_create_kmalloc_pool(
			SWAP_CLUSTER_MAX, sizeof(struct swap_iocb));
		if (cmpxchg(&sio_pool, NULL, pool))
			mempool_destroy(pool);
	}
	if (!sio_pool)
		return -ENOMEM;
	return 0;
}

static void sio_write_complete(struct kiocb *iocb, long ret)
{
	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
	struct page *page = sio->bvec[0].bv_page;
	int p;

	if (ret != sio->len) {
		/*
		 * In the case of swap-over-nfs, this can be a
		 * temporary failure if the system has limited
		 * memory for allocating transmit buffers.
		 * Mark the page dirty and avoid
		 * folio_rotate_reclaimable but rate-limit the
		 * messages but do not flag PageError like
		 * the normal direct-to-bio case as it could
		 * be temporary.
		 */
		pr_err_ratelimited("Write error %ld on dio swapfile (%llu)\n",
				   ret, page_file_offset(page));
		for (p = 0; p < sio->pages; p++) {
			page = sio->bvec[p].bv_page;
			set_page_dirty(page);
			ClearPageReclaim(page);
		}
	} else {
		for (p = 0; p < sio->pages; p++)
			count_swpout_vm_event(sio->bvec[p].bv_page);
	}

	for (p = 0; p < sio->pages; p++)
		end_page_writeback(sio->bvec[p].bv_page);

	mempool_free(sio, sio_pool);
}

static int swap_writepage_fs(struct page *page, struct writeback_control *wbc)
{
	struct swap_iocb *sio = NULL;
	struct swap_info_struct *sis = page_swap_info(page);
	struct file *swap_file = sis->swap_file;
	loff_t pos = page_file_offset(page);

	set_page_writeback(page);
	unlock_page(page);
	if (wbc->swap_plug)
		sio = *wbc->swap_plug;
	if (sio) {
		if (sio->iocb.ki_filp != swap_file ||
		    sio->iocb.ki_pos + sio->len != pos) {
			swap_write_unplug(sio);
			sio = NULL;
		}
	}
	if (!sio) {
		sio = mempool_alloc(sio_pool, GFP_NOIO);
		init_sync_kiocb(&sio->iocb, swap_file);
		sio->iocb.ki_complete = sio_write_complete;
		sio->iocb.ki_pos = pos;
		sio->pages = 0;
		sio->len = 0;
	}
	sio->bvec[sio->pages].bv_page = page;
	sio->bvec[sio->pages].bv_len = thp_size(page);
	sio->bvec[sio->pages].bv_offset = 0;
	sio->len += thp_size(page);
	sio->pages += 1;
	if (sio->pages == ARRAY_SIZE(sio->bvec) || !wbc->swap_plug) {
		swap_write_unplug(sio);
		sio = NULL;
	}
	if (wbc->swap_plug)
		*wbc->swap_plug = sio;

	return 0;
}

int __swap_writepage(struct page *page, struct writeback_control *wbc)
{
	struct bio *bio;
	int ret;
	struct swap_info_struct *sis = page_swap_info(page);

	VM_BUG_ON_PAGE(!PageSwapCache(page), page);
	/*
	 * ->flags can be updated non-atomicially (scan_swap_map_slots),
	 * but that will never affect SWP_FS_OPS, so the data_race
	 * is safe.
	 */
	if (data_race(sis->flags & SWP_FS_OPS))
		return swap_writepage_fs(page, wbc);

	ret = bdev_write_page(sis->bdev, swap_page_sector(page), page, wbc);
	if (!ret) {
		count_swpout_vm_event(page);
		return 0;
	}

	bio = bio_alloc(sis->bdev, 1,
			REQ_OP_WRITE | REQ_SWAP | wbc_to_write_flags(wbc),
			GFP_NOIO);
	bio->bi_iter.bi_sector = swap_page_sector(page);
	bio->bi_end_io = end_swap_bio_write;
	bio_add_page(bio, page, thp_size(page), 0);

	bio_associate_blkg_from_page(bio, page);
	count_swpout_vm_event(page);
	set_page_writeback(page);
	unlock_page(page);
	submit_bio(bio);

	return 0;
}

void swap_write_unplug(struct swap_iocb *sio)
{
	struct iov_iter from;
	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	int ret;

	iov_iter_bvec(&from, ITER_SOURCE, sio->bvec, sio->pages, sio->len);
	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	if (ret != -EIOCBQUEUED)
		sio_write_complete(&sio->iocb, ret);
}

static void sio_read_complete(struct kiocb *iocb, long ret)
{
	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
	int p;

	if (ret == sio->len) {
		for (p = 0; p < sio->pages; p++) {
			struct page *page = sio->bvec[p].bv_page;

			SetPageUptodate(page);
			unlock_page(page);
		}
		count_vm_events(PSWPIN, sio->pages);
	} else {
		for (p = 0; p < sio->pages; p++) {
			struct page *page = sio->bvec[p].bv_page;

			SetPageError(page);
			ClearPageUptodate(page);
			unlock_page(page);
		}
		pr_alert_ratelimited("Read-error on swap-device\n");
	}
	mempool_free(sio, sio_pool);
}

static void swap_readpage_fs(struct page *page,
			     struct swap_iocb **plug)
{
	struct swap_info_struct *sis = page_swap_info(page);
	struct swap_iocb *sio = NULL;
	loff_t pos = page_file_offset(page);

	if (plug)
		sio = *plug;
	if (sio) {
		if (sio->iocb.ki_filp != sis->swap_file ||
		    sio->iocb.ki_pos + sio->len != pos) {
			swap_read_unplug(sio);
			sio = NULL;
		}
	}
	if (!sio) {
		sio = mempool_alloc(sio_pool, GFP_KERNEL);
		init_sync_kiocb(&sio->iocb, sis->swap_file);
		sio->iocb.ki_pos = pos;
		sio->iocb.ki_complete = sio_read_complete;
		sio->pages = 0;
		sio->len = 0;
	}
	sio->bvec[sio->pages].bv_page = page;
	sio->bvec[sio->pages].bv_len = thp_size(page);
	sio->bvec[sio->pages].bv_offset = 0;
	sio->len += thp_size(page);
	sio->pages += 1;
	if (sio->pages == ARRAY_SIZE(sio->bvec) || !plug) {
		swap_read_unplug(sio);
		sio = NULL;
	}
	if (plug)
		*plug = sio;
}

int swap_readpage(struct page *page, bool synchronous,
		  struct swap_iocb **plug)
{
	struct bio *bio;
	int ret = 0;
	struct swap_info_struct *sis = page_swap_info(page);
	bool workingset = PageWorkingset(page);
	unsigned long pflags;
	bool in_thrashing;

	VM_BUG_ON_PAGE(!PageSwapCache(page) && !synchronous, page);
	VM_BUG_ON_PAGE(!PageLocked(page), page);
	VM_BUG_ON_PAGE(PageUptodate(page), page);

	/*
	 * Count submission time as memory stall and delay. When the device
	 * is congested, or the submitting cgroup IO-throttled, submission
	 * can be a significant part of overall IO time.
	 */
	if (workingset) {
		delayacct_thrashing_start(&in_thrashing);
		psi_memstall_enter(&pflags);
	}
	delayacct_swapin_start();

	if (frontswap_load(page) == 0) {
		SetPageUptodate(page);
		unlock_page(page);
		goto out;
	}

	if (data_race(sis->flags & SWP_FS_OPS)) {
		swap_readpage_fs(page, plug);
		goto out;
	}

	if (sis->flags & SWP_SYNCHRONOUS_IO) {
		ret = bdev_read_page(sis->bdev, swap_page_sector(page), page);
		if (!ret) {
			count_vm_event(PSWPIN);
			goto out;
		}
	}

	ret = 0;
	bio = bio_alloc(sis->bdev, 1, REQ_OP_READ, GFP_KERNEL);
	bio->bi_iter.bi_sector = swap_page_sector(page);
	bio->bi_end_io = end_swap_bio_read;
	bio_add_page(bio, page, thp_size(page), 0);
	/*
	 * Keep this task valid during swap readpage because the oom killer may
	 * attempt to access it in the page fault retry time check.
	 */
	if (synchronous) {
		get_task_struct(current);
		bio->bi_private = current;
	}
	count_vm_event(PSWPIN);
	bio_get(bio);
	submit_bio(bio);
	while (synchronous) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!READ_ONCE(bio->bi_private))
			break;

		blk_io_schedule();
	}
	__set_current_state(TASK_RUNNING);
	bio_put(bio);

out:
	if (workingset) {
		delayacct_thrashing_end(&in_thrashing);
		psi_memstall_leave(&pflags);
	}
	delayacct_swapin_end();
	return ret;
}

void __swap_read_unplug(struct swap_iocb *sio)
{
	struct iov_iter from;
	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	int ret;

	iov_iter_bvec(&from, ITER_DEST, sio->bvec, sio->pages, sio->len);
	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	if (ret != -EIOCBQUEUED)
		sio_read_complete(&sio->iocb, ret);
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* linux/mm/page_io.c
	4	*
	5	* Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
	6	*
	7	* Swap reorganised 29.12.95,
	8	* Asynchronous swapping added 30.12.95. Stephen Tweedie
	9	* Removed race in async swapping. 14.4.1996. Bruno Haible
	10	* Add swap of shared pages through the page cache. 20.2.1998. Stephen Tweedie
	11	* Always use brw_page, life becomes simpler. 12 May 1998 Eric Biederman
	12	*/
	13
	14	#include <linux/mm.h>
	15	#include <linux/kernel_stat.h>
5a0e3ad6	16	#include <linux/gfp.h>
1da177e4 LT	17	#include <linux/pagemap.h>
	18	#include <linux/swap.h>
	19	#include <linux/bio.h>
	20	#include <linux/swapops.h>
62c230bc	21	#include <linux/buffer_head.h>
1da177e4	22	#include <linux/writeback.h>
38b5faf4	23	#include <linux/frontswap.h>
b430e9d1	24	#include <linux/blkdev.h>
93779069	25	#include <linux/psi.h>
e2e40f2c	26	#include <linux/uio.h>
b0ba2d0f	27	#include <linux/sched/task.h>
a3d5dc90	28	#include <linux/delayacct.h>
014bb1de	29	#include "swap.h"
1da177e4	30
cf1e3fe4	31	static void end_swap_bio_write(struct bio *bio)
1da177e4	32	{
263663cd	33	struct page *page = bio_first_page_all(bio);
1da177e4	34
4e4cbee9	35	if (bio->bi_status) {
1da177e4	36	SetPageError(page);
6ddab3b9 PZ	37	/*
	38	* We failed to write the page out to swap-space.
	39	* Re-dirty the page in order to avoid it being reclaimed.
	40	* Also print a dire warning that things will go BAD (tm)
	41	* very quickly.
	42	*
575ced1c	43	* Also clear PG_reclaim to avoid folio_rotate_reclaimable()
6ddab3b9 PZ	44	*/
6ddab3b9 PZ	45	set_page_dirty(page);
25eaab43 GD	46	pr_alert_ratelimited("Write-error on swap-device (%u:%u:%llu)\n",
	47	MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
	48	(unsigned long long)bio->bi_iter.bi_sector);
6ddab3b9 PZ	49	ClearPageReclaim(page);
6ddab3b9 PZ	50	}
1da177e4 LT	51	end_page_writeback(page);
1da177e4 LT	52	bio_put(bio);
1da177e4 LT	53	}
1da177e4 LT	54
4246a0b6	55	static void end_swap_bio_read(struct bio *bio)
1da177e4	56	{
263663cd	57	struct page *page = bio_first_page_all(bio);
23955622	58	struct task_struct *waiter = bio->bi_private;
1da177e4	59
4e4cbee9	60	if (bio->bi_status) {
1da177e4 LT	61	SetPageError(page);
1da177e4 LT	62	ClearPageUptodate(page);
25eaab43 GD	63	pr_alert_ratelimited("Read-error on swap-device (%u:%u:%llu)\n",
	64	MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
	65	(unsigned long long)bio->bi_iter.bi_sector);
b430e9d1	66	goto out;
1da177e4	67	}
b430e9d1 MK	68
b430e9d1 MK	69	SetPageUptodate(page);
b430e9d1	70	out:
1da177e4	71	unlock_page(page);
23955622	72	WRITE_ONCE(bio->bi_private, NULL);
1da177e4	73	bio_put(bio);
87518530 ON	74	if (waiter) {
	75	blk_wake_io_task(waiter);
	76	put_task_struct(waiter);
	77	}
1da177e4 LT	78	}
1da177e4 LT	79
a509bc1a MG	80	int generic_swapfile_activate(struct swap_info_struct *sis,
	81	struct file *swap_file,
	82	sector_t *span)
	83	{
	84	struct address_space *mapping = swap_file->f_mapping;
	85	struct inode *inode = mapping->host;
	86	unsigned blocks_per_page;
	87	unsigned long page_no;
	88	unsigned blkbits;
	89	sector_t probe_block;
	90	sector_t last_block;
	91	sector_t lowest_block = -1;
	92	sector_t highest_block = 0;
	93	int nr_extents = 0;
	94	int ret;
	95
	96	blkbits = inode->i_blkbits;
	97	blocks_per_page = PAGE_SIZE >> blkbits;
	98
	99	/*
4efaceb1	100	* Map all the blocks into the extent tree. This code doesn't try
a509bc1a MG	101	* to be very smart.
	102	*/
	103	probe_block = 0;
	104	page_no = 0;
	105	last_block = i_size_read(inode) >> blkbits;
	106	while ((probe_block + blocks_per_page) <= last_block &&
	107	page_no < sis->max) {
	108	unsigned block_in_page;
	109	sector_t first_block;
	110
7e4411bf MP	111	cond_resched();
7e4411bf MP	112
30460e1e CM	113	first_block = probe_block;
	114	ret = bmap(inode, &first_block);
	115	if (ret \|\| !first_block)
a509bc1a MG	116	goto bad_bmap;
	117
	118	/*
	119	* It must be PAGE_SIZE aligned on-disk
	120	*/
	121	if (first_block & (blocks_per_page - 1)) {
	122	probe_block++;
	123	goto reprobe;
	124	}
	125
	126	for (block_in_page = 1; block_in_page < blocks_per_page;
	127	block_in_page++) {
	128	sector_t block;
	129
30460e1e CM	130	block = probe_block + block_in_page;
	131	ret = bmap(inode, &block);
	132	if (ret \|\| !block)
a509bc1a	133	goto bad_bmap;
30460e1e	134
a509bc1a MG	135	if (block != first_block + block_in_page) {
	136	/* Discontiguity */
	137	probe_block++;
	138	goto reprobe;
	139	}
	140	}
	141
	142	first_block >>= (PAGE_SHIFT - blkbits);
	143	if (page_no) { /* exclude the header page */
	144	if (first_block < lowest_block)
	145	lowest_block = first_block;
	146	if (first_block > highest_block)
	147	highest_block = first_block;
	148	}
	149
	150	/*
	151	* We found a PAGE_SIZE-length, PAGE_SIZE-aligned run of blocks
	152	*/
	153	ret = add_swap_extent(sis, page_no, 1, first_block);
	154	if (ret < 0)
	155	goto out;
	156	nr_extents += ret;
	157	page_no++;
	158	probe_block += blocks_per_page;
	159	reprobe:
	160	continue;
	161	}
	162	ret = nr_extents;
	163	*span = 1 + highest_block - lowest_block;
	164	if (page_no == 0)
	165	page_no = 1; /* force Empty message */
	166	sis->max = page_no;
	167	sis->pages = page_no - 1;
	168	sis->highest_bit = page_no - 1;
	169	out:
	170	return ret;
	171	bad_bmap:
1170532b	172	pr_err("swapon: swapfile has holes\n");
a509bc1a MG	173	ret = -EINVAL;
	174	goto out;
	175	}
	176
1da177e4 LT	177	/*
	178	* We may have stale swap cache pages in memory: notice
	179	* them here and get rid of the unnecessary final write.
	180	*/
	181	int swap_writepage(struct page page, struct writeback_control wbc)
	182	{
71fa1a53	183	struct folio *folio = page_folio(page);
2f772e6c	184	int ret = 0;
1da177e4	185
71fa1a53 MWO	186	if (folio_free_swap(folio)) {
71fa1a53 MWO	187	folio_unlock(folio);
1da177e4 LT	188	goto out;
1da177e4 LT	189	}
8a84802e SP	190	/*
	191	* Arch code may have to preserve more data than just the page
	192	* contents, e.g. memory tags.
	193	*/
71fa1a53	194	ret = arch_prepare_to_swap(&folio->page);
8a84802e	195	if (ret) {
71fa1a53 MWO	196	folio_mark_dirty(folio);
71fa1a53 MWO	197	folio_unlock(folio);
8a84802e SP	198	goto out;
8a84802e SP	199	}
71fa1a53 MWO	200	if (frontswap_store(&folio->page) == 0) {
	201	folio_start_writeback(folio);
	202	folio_unlock(folio);
	203	folio_end_writeback(folio);
38b5faf4 DM	204	goto out;
38b5faf4 DM	205	}
71fa1a53	206	ret = __swap_writepage(&folio->page, wbc);
2f772e6c SJ	207	out:
	208	return ret;
	209	}
	210
225311a4 HY	211	static inline void count_swpout_vm_event(struct page *page)
	212	{
	213	#ifdef CONFIG_TRANSPARENT_HUGEPAGE
	214	if (unlikely(PageTransHuge(page)))
	215	count_vm_event(THP_SWPOUT);
	216	#endif
6c357848	217	count_vm_events(PSWPOUT, thp_nr_pages(page));
225311a4 HY	218	}
225311a4 HY	219
a18b9b15 CH	220	#if defined(CONFIG_MEMCG) && defined(CONFIG_BLK_CGROUP)
	221	static void bio_associate_blkg_from_page(struct bio bio, struct page page)
	222	{
	223	struct cgroup_subsys_state *css;
bcfe06bf	224	struct mem_cgroup *memcg;
a18b9b15	225
bcfe06bf RG	226	memcg = page_memcg(page);
bcfe06bf RG	227	if (!memcg)
a18b9b15 CH	228	return;
	229
	230	rcu_read_lock();
bcfe06bf	231	css = cgroup_e_css(memcg->css.cgroup, &io_cgrp_subsys);
a18b9b15 CH	232	bio_associate_blkg_from_css(bio, css);
	233	rcu_read_unlock();
	234	}
	235	#else
	236	#define bio_associate_blkg_from_page(bio, page) do { } while (0)
	237	#endif /* CONFIG_MEMCG && CONFIG_BLK_CGROUP */
	238
e1209d3a N	239	struct swap_iocb {
e1209d3a N	240	struct kiocb iocb;
5169b844 N	241	struct bio_vec bvec[SWAP_CLUSTER_MAX];
5169b844 N	242	int pages;
a1a0dfd5	243	int len;
e1209d3a N	244	};
	245	static mempool_t *sio_pool;
	246
	247	int sio_pool_init(void)
2f772e6c	248	{
e1209d3a N	249	if (!sio_pool) {
	250	mempool_t *pool = mempool_create_kmalloc_pool(
	251	SWAP_CLUSTER_MAX, sizeof(struct swap_iocb));
	252	if (cmpxchg(&sio_pool, NULL, pool))
	253	mempool_destroy(pool);
	254	}
	255	if (!sio_pool)
	256	return -ENOMEM;
	257	return 0;
	258	}
62c230bc	259
7eadabc0 N	260	static void sio_write_complete(struct kiocb *iocb, long ret)
	261	{
	262	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
5169b844	263	struct page *page = sio->bvec[0].bv_page;
2282679f	264	int p;
62c230bc	265
a1a0dfd5	266	if (ret != sio->len) {
7eadabc0 N	267	/*
	268	* In the case of swap-over-nfs, this can be a
	269	* temporary failure if the system has limited
	270	* memory for allocating transmit buffers.
	271	* Mark the page dirty and avoid
	272	* folio_rotate_reclaimable but rate-limit the
	273	* messages but do not flag PageError like
	274	* the normal direct-to-bio case as it could
	275	* be temporary.
	276	*/
7eadabc0 N	277	pr_err_ratelimited("Write error %ld on dio swapfile (%llu)\n",
7eadabc0 N	278	ret, page_file_offset(page));
2282679f N	279	for (p = 0; p < sio->pages; p++) {
2282679f N	280	page = sio->bvec[p].bv_page;
2d30d31e	281	set_page_dirty(page);
0cdc444a	282	ClearPageReclaim(page);
62c230bc	283	}
6341a446 N	284	} else {
	285	for (p = 0; p < sio->pages; p++)
	286	count_swpout_vm_event(sio->bvec[p].bv_page);
62c230bc MG	287	}
62c230bc MG	288
2282679f N	289	for (p = 0; p < sio->pages; p++)
	290	end_page_writeback(sio->bvec[p].bv_page);
	291
7eadabc0 N	292	mempool_free(sio, sio_pool);
	293	}
	294
	295	static int swap_writepage_fs(struct page page, struct writeback_control wbc)
	296	{
2282679f	297	struct swap_iocb *sio = NULL;
7eadabc0 N	298	struct swap_info_struct *sis = page_swap_info(page);
7eadabc0 N	299	struct file *swap_file = sis->swap_file;
2282679f	300	loff_t pos = page_file_offset(page);
7eadabc0 N	301
	302	set_page_writeback(page);
	303	unlock_page(page);
2282679f N	304	if (wbc->swap_plug)
	305	sio = *wbc->swap_plug;
	306	if (sio) {
	307	if (sio->iocb.ki_filp != swap_file \|\|
a1a0dfd5	308	sio->iocb.ki_pos + sio->len != pos) {
2282679f N	309	swap_write_unplug(sio);
	310	sio = NULL;
	311	}
	312	}
	313	if (!sio) {
	314	sio = mempool_alloc(sio_pool, GFP_NOIO);
	315	init_sync_kiocb(&sio->iocb, swap_file);
	316	sio->iocb.ki_complete = sio_write_complete;
	317	sio->iocb.ki_pos = pos;
	318	sio->pages = 0;
a1a0dfd5	319	sio->len = 0;
2282679f N	320	}
2282679f N	321	sio->bvec[sio->pages].bv_page = page;
a1a0dfd5	322	sio->bvec[sio->pages].bv_len = thp_size(page);
2282679f	323	sio->bvec[sio->pages].bv_offset = 0;
a1a0dfd5	324	sio->len += thp_size(page);
2282679f N	325	sio->pages += 1;
	326	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !wbc->swap_plug) {
	327	swap_write_unplug(sio);
	328	sio = NULL;
	329	}
	330	if (wbc->swap_plug)
	331	*wbc->swap_plug = sio;
	332
	333	return 0;
7eadabc0 N	334	}
7eadabc0 N	335
cf1e3fe4	336	int __swap_writepage(struct page page, struct writeback_control wbc)
2f772e6c SJ	337	{
2f772e6c SJ	338	struct bio *bio;
4e49ea4a	339	int ret;
2f772e6c	340	struct swap_info_struct *sis = page_swap_info(page);
62c230bc	341
cc30c5d6	342	VM_BUG_ON_PAGE(!PageSwapCache(page), page);
7eadabc0 N	343	/*
	344	* ->flags can be updated non-atomicially (scan_swap_map_slots),
	345	* but that will never affect SWP_FS_OPS, so the data_race
	346	* is safe.
	347	*/
	348	if (data_race(sis->flags & SWP_FS_OPS))
	349	return swap_writepage_fs(page, wbc);
62c230bc	350
dd6bd0d9 MW	351	ret = bdev_write_page(sis->bdev, swap_page_sector(page), page, wbc);
dd6bd0d9 MW	352	if (!ret) {
225311a4	353	count_swpout_vm_event(page);
dd6bd0d9 MW	354	return 0;
	355	}
	356
07888c66 CH	357	bio = bio_alloc(sis->bdev, 1,
	358	REQ_OP_WRITE \| REQ_SWAP \| wbc_to_write_flags(wbc),
	359	GFP_NOIO);
48d15436	360	bio->bi_iter.bi_sector = swap_page_sector(page);
cf1e3fe4	361	bio->bi_end_io = end_swap_bio_write;
48d15436 CH	362	bio_add_page(bio, page, thp_size(page), 0);
48d15436 CH	363
6a7f6d86	364	bio_associate_blkg_from_page(bio, page);
225311a4	365	count_swpout_vm_event(page);
1da177e4 LT	366	set_page_writeback(page);
1da177e4 LT	367	unlock_page(page);
4e49ea4a	368	submit_bio(bio);
548d9782 ML	369
548d9782 ML	370	return 0;
1da177e4 LT	371	}
1da177e4 LT	372
2282679f N	373	void swap_write_unplug(struct swap_iocb *sio)
	374	{
	375	struct iov_iter from;
	376	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	377	int ret;
	378
de4eda9d	379	iov_iter_bvec(&from, ITER_SOURCE, sio->bvec, sio->pages, sio->len);
2282679f N	380	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	381	if (ret != -EIOCBQUEUED)
	382	sio_write_complete(&sio->iocb, ret);
	383	}
	384
e1209d3a N	385	static void sio_read_complete(struct kiocb *iocb, long ret)
	386	{
	387	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
5169b844	388	int p;
e1209d3a	389
a1a0dfd5	390	if (ret == sio->len) {
5169b844 N	391	for (p = 0; p < sio->pages; p++) {
	392	struct page *page = sio->bvec[p].bv_page;
	393
	394	SetPageUptodate(page);
	395	unlock_page(page);
	396	}
	397	count_vm_events(PSWPIN, sio->pages);
e1209d3a	398	} else {
5169b844 N	399	for (p = 0; p < sio->pages; p++) {
	400	struct page *page = sio->bvec[p].bv_page;
	401
	402	SetPageError(page);
	403	ClearPageUptodate(page);
	404	unlock_page(page);
	405	}
	406	pr_alert_ratelimited("Read-error on swap-device\n");
e1209d3a	407	}
e1209d3a N	408	mempool_free(sio, sio_pool);
	409	}
	410
5169b844 N	411	static void swap_readpage_fs(struct page *page,
5169b844 N	412	struct swap_iocb **plug)
e1209d3a N	413	{
e1209d3a N	414	struct swap_info_struct *sis = page_swap_info(page);
5169b844	415	struct swap_iocb *sio = NULL;
e1209d3a	416	loff_t pos = page_file_offset(page);
e1209d3a	417
5169b844 N	418	if (plug)
	419	sio = *plug;
	420	if (sio) {
	421	if (sio->iocb.ki_filp != sis->swap_file \|\|
a1a0dfd5	422	sio->iocb.ki_pos + sio->len != pos) {
5169b844 N	423	swap_read_unplug(sio);
	424	sio = NULL;
	425	}
	426	}
	427	if (!sio) {
	428	sio = mempool_alloc(sio_pool, GFP_KERNEL);
	429	init_sync_kiocb(&sio->iocb, sis->swap_file);
	430	sio->iocb.ki_pos = pos;
	431	sio->iocb.ki_complete = sio_read_complete;
	432	sio->pages = 0;
a1a0dfd5	433	sio->len = 0;
5169b844 N	434	}
5169b844 N	435	sio->bvec[sio->pages].bv_page = page;
a1a0dfd5	436	sio->bvec[sio->pages].bv_len = thp_size(page);
5169b844	437	sio->bvec[sio->pages].bv_offset = 0;
a1a0dfd5	438	sio->len += thp_size(page);
5169b844 N	439	sio->pages += 1;
	440	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !plug) {
	441	swap_read_unplug(sio);
	442	sio = NULL;
	443	}
	444	if (plug)
	445	*plug = sio;
e1209d3a N	446	}
e1209d3a N	447
5169b844 N	448	int swap_readpage(struct page *page, bool synchronous,
5169b844 N	449	struct swap_iocb **plug)
1da177e4 LT	450	{
	451	struct bio *bio;
	452	int ret = 0;
62c230bc	453	struct swap_info_struct *sis = page_swap_info(page);
d8c47cc7	454	bool workingset = PageWorkingset(page);
93779069	455	unsigned long pflags;
3a9bb7b1	456	bool in_thrashing;
1da177e4	457
0bcac06f	458	VM_BUG_ON_PAGE(!PageSwapCache(page) && !synchronous, page);
309381fe SL	459	VM_BUG_ON_PAGE(!PageLocked(page), page);
309381fe SL	460	VM_BUG_ON_PAGE(PageUptodate(page), page);
93779069 MK	461
93779069 MK	462	/*
3a9bb7b1 YY	463	* Count submission time as memory stall and delay. When the device
	464	* is congested, or the submitting cgroup IO-throttled, submission
	465	* can be a significant part of overall IO time.
93779069	466	*/
3a9bb7b1 YY	467	if (workingset) {
3a9bb7b1 YY	468	delayacct_thrashing_start(&in_thrashing);
d8c47cc7	469	psi_memstall_enter(&pflags);
3a9bb7b1	470	}
a3d5dc90	471	delayacct_swapin_start();
93779069	472
165c8aed	473	if (frontswap_load(page) == 0) {
38b5faf4 DM	474	SetPageUptodate(page);
	475	unlock_page(page);
	476	goto out;
	477	}
62c230bc	478
32646315	479	if (data_race(sis->flags & SWP_FS_OPS)) {
5169b844	480	swap_readpage_fs(page, plug);
93779069	481	goto out;
62c230bc MG	482	}
62c230bc MG	483
5115db10 CH	484	if (sis->flags & SWP_SYNCHRONOUS_IO) {
	485	ret = bdev_read_page(sis->bdev, swap_page_sector(page), page);
	486	if (!ret) {
5115db10 CH	487	count_vm_event(PSWPIN);
	488	goto out;
	489	}
dd6bd0d9 MW	490	}
	491
	492	ret = 0;
07888c66	493	bio = bio_alloc(sis->bdev, 1, REQ_OP_READ, GFP_KERNEL);
48d15436 CH	494	bio->bi_iter.bi_sector = swap_page_sector(page);
	495	bio->bi_end_io = end_swap_bio_read;
	496	bio_add_page(bio, page, thp_size(page), 0);
b0ba2d0f TH	497	/*
	498	* Keep this task valid during swap readpage because the oom killer may
	499	* attempt to access it in the page fault retry time check.
	500	*/
87518530	501	if (synchronous) {
87518530 ON	502	get_task_struct(current);
	503	bio->bi_private = current;
	504	}
f8891e5e	505	count_vm_event(PSWPIN);
23955622	506	bio_get(bio);
3e08773c	507	submit_bio(bio);
0bcac06f	508	while (synchronous) {
1ac5cd49	509	set_current_state(TASK_UNINTERRUPTIBLE);
23955622 SL	510	if (!READ_ONCE(bio->bi_private))
	511	break;
	512
9650b453	513	blk_io_schedule();
23955622 SL	514	}
	515	__set_current_state(TASK_RUNNING);
	516	bio_put(bio);
	517
1da177e4	518	out:
3a9bb7b1 YY	519	if (workingset) {
3a9bb7b1 YY	520	delayacct_thrashing_end(&in_thrashing);
d8c47cc7	521	psi_memstall_leave(&pflags);
3a9bb7b1	522	}
a3d5dc90	523	delayacct_swapin_end();
1da177e4 LT	524	return ret;
1da177e4 LT	525	}
62c230bc	526
5169b844	527	void __swap_read_unplug(struct swap_iocb *sio)
62c230bc	528	{
5169b844 N	529	struct iov_iter from;
	530	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	531	int ret;
cc30c5d6	532
de4eda9d	533	iov_iter_bvec(&from, ITER_DEST, sio->bvec, sio->pages, sio->len);
5169b844 N	534	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	535	if (ret != -EIOCBQUEUED)
	536	sio_read_complete(&sio->iocb, ret);
62c230bc	537	}