[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/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);
}

static void end_swap_bio_write(struct bio *bio)
{
	__end_swap_bio_write(bio);
	bio_put(bio);
}

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

	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);
	} else {
		SetPageUptodate(page);
	}
	unlock_page(page);
}

static void end_swap_bio_read(struct bio *bio)
{
	__end_swap_bio_read(bio);
	bio_put(bio);
}

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;

	if (folio_free_swap(folio)) {
		folio_unlock(folio);
		return 0;
	}
	/*
	 * 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);
		return ret;
	}
	if (frontswap_store(&folio->page) == 0) {
		folio_start_writeback(folio);
		folio_unlock(folio);
		folio_end_writeback(folio);
		return 0;
	}
	__swap_writepage(&folio->page, wbc);
	return 0;
}

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 void 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;
	}
	bvec_set_page(&sio->bvec[sio->pages], page, thp_size(page), 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;
}

static void swap_writepage_bdev_sync(struct page *page,
		struct writeback_control *wbc, struct swap_info_struct *sis)
{
	struct bio_vec bv;
	struct bio bio;

	bio_init(&bio, sis->bdev, &bv, 1,
		 REQ_OP_WRITE | REQ_SWAP | wbc_to_write_flags(wbc));
	bio.bi_iter.bi_sector = swap_page_sector(page);
	__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_wait(&bio);
	__end_swap_bio_write(&bio);
}

static void swap_writepage_bdev_async(struct page *page,
		struct writeback_control *wbc, struct swap_info_struct *sis)
{
	struct bio *bio;

	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);
}

void __swap_writepage(struct page *page, struct writeback_control *wbc)
{
	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))
		swap_writepage_fs(page, wbc);
	else if (sis->flags & SWP_SYNCHRONOUS_IO)
		swap_writepage_bdev_sync(page, wbc, sis);
	else
		swap_writepage_bdev_async(page, wbc, sis);
}

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;
	}
	bvec_set_page(&sio->bvec[sio->pages], page, thp_size(page), 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;
}

static void swap_readpage_bdev_sync(struct page *page,
		struct swap_info_struct *sis)
{
	struct bio_vec bv;
	struct bio bio;

	bio_init(&bio, sis->bdev, &bv, 1, REQ_OP_READ);
	bio.bi_iter.bi_sector = swap_page_sector(page);
	__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.
	 */
	get_task_struct(current);
	count_vm_event(PSWPIN);
	submit_bio_wait(&bio);
	__end_swap_bio_read(&bio);
	put_task_struct(current);
}

static void swap_readpage_bdev_async(struct page *page,
		struct swap_info_struct *sis)
{
	struct bio *bio;

	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);
	count_vm_event(PSWPIN);
	submit_bio(bio);
}

void swap_readpage(struct page *page, bool synchronous, struct swap_iocb **plug)
{
	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);
	} else if (data_race(sis->flags & SWP_FS_OPS)) {
		swap_readpage_fs(page, plug);
	} else if (synchronous || (sis->flags & SWP_SYNCHRONOUS_IO)) {
		swap_readpage_bdev_sync(page, sis);
	} else {
		swap_readpage_bdev_async(page, sis);
	}

	if (workingset) {
		delayacct_thrashing_end(&in_thrashing);
		psi_memstall_leave(&pflags);
	}
	delayacct_swapin_end();
}

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>
	21	#include <linux/writeback.h>
38b5faf4	22	#include <linux/frontswap.h>
b430e9d1	23	#include <linux/blkdev.h>
93779069	24	#include <linux/psi.h>
e2e40f2c	25	#include <linux/uio.h>
b0ba2d0f	26	#include <linux/sched/task.h>
a3d5dc90	27	#include <linux/delayacct.h>
014bb1de	28	#include "swap.h"
1da177e4	29
3222d8c2	30	static void __end_swap_bio_write(struct bio *bio)
1da177e4	31	{
263663cd	32	struct page *page = bio_first_page_all(bio);
1da177e4	33
4e4cbee9	34	if (bio->bi_status) {
1da177e4	35	SetPageError(page);
6ddab3b9 PZ	36	/*
	37	* We failed to write the page out to swap-space.
	38	* Re-dirty the page in order to avoid it being reclaimed.
	39	* Also print a dire warning that things will go BAD (tm)
	40	* very quickly.
	41	*
575ced1c	42	* Also clear PG_reclaim to avoid folio_rotate_reclaimable()
6ddab3b9 PZ	43	*/
6ddab3b9 PZ	44	set_page_dirty(page);
25eaab43 GD	45	pr_alert_ratelimited("Write-error on swap-device (%u:%u:%llu)\n",
	46	MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
	47	(unsigned long long)bio->bi_iter.bi_sector);
6ddab3b9 PZ	48	ClearPageReclaim(page);
6ddab3b9 PZ	49	}
1da177e4	50	end_page_writeback(page);
3222d8c2 CH	51	}
	52
	53	static void end_swap_bio_write(struct bio *bio)
	54	{
	55	__end_swap_bio_write(bio);
1da177e4	56	bio_put(bio);
1da177e4 LT	57	}
1da177e4 LT	58
9b4e30bd	59	static void __end_swap_bio_read(struct bio *bio)
1da177e4	60	{
263663cd	61	struct page *page = bio_first_page_all(bio);
1da177e4	62
4e4cbee9	63	if (bio->bi_status) {
1da177e4 LT	64	SetPageError(page);
1da177e4 LT	65	ClearPageUptodate(page);
25eaab43 GD	66	pr_alert_ratelimited("Read-error on swap-device (%u:%u:%llu)\n",
	67	MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
	68	(unsigned long long)bio->bi_iter.bi_sector);
9b4e30bd CH	69	} else {
9b4e30bd CH	70	SetPageUptodate(page);
1da177e4 LT	71	}
1da177e4 LT	72	unlock_page(page);
9b4e30bd CH	73	}
	74
	75	static void end_swap_bio_read(struct bio *bio)
	76	{
	77	__end_swap_bio_read(bio);
1da177e4	78	bio_put(bio);
1da177e4 LT	79	}
1da177e4 LT	80
a509bc1a MG	81	int generic_swapfile_activate(struct swap_info_struct *sis,
	82	struct file *swap_file,
	83	sector_t *span)
	84	{
	85	struct address_space *mapping = swap_file->f_mapping;
	86	struct inode *inode = mapping->host;
	87	unsigned blocks_per_page;
	88	unsigned long page_no;
	89	unsigned blkbits;
	90	sector_t probe_block;
	91	sector_t last_block;
	92	sector_t lowest_block = -1;
	93	sector_t highest_block = 0;
	94	int nr_extents = 0;
	95	int ret;
	96
	97	blkbits = inode->i_blkbits;
	98	blocks_per_page = PAGE_SIZE >> blkbits;
	99
	100	/*
4efaceb1	101	* Map all the blocks into the extent tree. This code doesn't try
a509bc1a MG	102	* to be very smart.
	103	*/
	104	probe_block = 0;
	105	page_no = 0;
	106	last_block = i_size_read(inode) >> blkbits;
	107	while ((probe_block + blocks_per_page) <= last_block &&
	108	page_no < sis->max) {
	109	unsigned block_in_page;
	110	sector_t first_block;
	111
7e4411bf MP	112	cond_resched();
7e4411bf MP	113
30460e1e CM	114	first_block = probe_block;
	115	ret = bmap(inode, &first_block);
	116	if (ret \|\| !first_block)
a509bc1a MG	117	goto bad_bmap;
	118
	119	/*
	120	* It must be PAGE_SIZE aligned on-disk
	121	*/
	122	if (first_block & (blocks_per_page - 1)) {
	123	probe_block++;
	124	goto reprobe;
	125	}
	126
	127	for (block_in_page = 1; block_in_page < blocks_per_page;
	128	block_in_page++) {
	129	sector_t block;
	130
30460e1e CM	131	block = probe_block + block_in_page;
	132	ret = bmap(inode, &block);
	133	if (ret \|\| !block)
a509bc1a	134	goto bad_bmap;
30460e1e	135
a509bc1a MG	136	if (block != first_block + block_in_page) {
	137	/* Discontiguity */
	138	probe_block++;
	139	goto reprobe;
	140	}
	141	}
	142
	143	first_block >>= (PAGE_SHIFT - blkbits);
	144	if (page_no) { /* exclude the header page */
	145	if (first_block < lowest_block)
	146	lowest_block = first_block;
	147	if (first_block > highest_block)
	148	highest_block = first_block;
	149	}
	150
	151	/*
	152	* We found a PAGE_SIZE-length, PAGE_SIZE-aligned run of blocks
	153	*/
	154	ret = add_swap_extent(sis, page_no, 1, first_block);
	155	if (ret < 0)
	156	goto out;
	157	nr_extents += ret;
	158	page_no++;
	159	probe_block += blocks_per_page;
	160	reprobe:
	161	continue;
	162	}
	163	ret = nr_extents;
	164	*span = 1 + highest_block - lowest_block;
	165	if (page_no == 0)
	166	page_no = 1; /* force Empty message */
	167	sis->max = page_no;
	168	sis->pages = page_no - 1;
	169	sis->highest_bit = page_no - 1;
	170	out:
	171	return ret;
	172	bad_bmap:
1170532b	173	pr_err("swapon: swapfile has holes\n");
a509bc1a MG	174	ret = -EINVAL;
	175	goto out;
	176	}
	177
1da177e4 LT	178	/*
	179	* We may have stale swap cache pages in memory: notice
	180	* them here and get rid of the unnecessary final write.
	181	*/
	182	int swap_writepage(struct page page, struct writeback_control wbc)
	183	{
71fa1a53	184	struct folio *folio = page_folio(page);
e3e2762b	185	int ret;
1da177e4	186
71fa1a53 MWO	187	if (folio_free_swap(folio)) {
71fa1a53 MWO	188	folio_unlock(folio);
e3e2762b	189	return 0;
1da177e4	190	}
8a84802e SP	191	/*
	192	* Arch code may have to preserve more data than just the page
	193	* contents, e.g. memory tags.
	194	*/
71fa1a53	195	ret = arch_prepare_to_swap(&folio->page);
8a84802e	196	if (ret) {
71fa1a53 MWO	197	folio_mark_dirty(folio);
71fa1a53 MWO	198	folio_unlock(folio);
e3e2762b	199	return ret;
8a84802e	200	}
71fa1a53 MWO	201	if (frontswap_store(&folio->page) == 0) {
	202	folio_start_writeback(folio);
	203	folio_unlock(folio);
	204	folio_end_writeback(folio);
e3e2762b	205	return 0;
38b5faf4	206	}
e3e2762b CH	207	__swap_writepage(&folio->page, wbc);
e3e2762b CH	208	return 0;
2f772e6c SJ	209	}
2f772e6c SJ	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
e3e2762b	295	static void swap_writepage_fs(struct page page, struct writeback_control wbc)
7eadabc0	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	320	}
8976fa6d	321	bvec_set_page(&sio->bvec[sio->pages], page, thp_size(page), 0);
a1a0dfd5	322	sio->len += thp_size(page);
2282679f N	323	sio->pages += 1;
	324	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !wbc->swap_plug) {
	325	swap_write_unplug(sio);
	326	sio = NULL;
	327	}
	328	if (wbc->swap_plug)
	329	*wbc->swap_plug = sio;
7eadabc0 N	330	}
7eadabc0 N	331
3222d8c2	332	static void swap_writepage_bdev_sync(struct page *page,
05cda97e	333	struct writeback_control wbc, struct swap_info_struct sis)
2f772e6c	334	{
3222d8c2 CH	335	struct bio_vec bv;
3222d8c2 CH	336	struct bio bio;
62c230bc	337
3222d8c2 CH	338	bio_init(&bio, sis->bdev, &bv, 1,
	339	REQ_OP_WRITE \| REQ_SWAP \| wbc_to_write_flags(wbc));
	340	bio.bi_iter.bi_sector = swap_page_sector(page);
cb58bf91	341	__bio_add_page(&bio, page, thp_size(page), 0);
62c230bc	342
3222d8c2 CH	343	bio_associate_blkg_from_page(&bio, page);
	344	count_swpout_vm_event(page);
	345
	346	set_page_writeback(page);
	347	unlock_page(page);
	348
	349	submit_bio_wait(&bio);
	350	__end_swap_bio_write(&bio);
	351	}
	352
	353	static void swap_writepage_bdev_async(struct page *page,
	354	struct writeback_control wbc, struct swap_info_struct sis)
	355	{
	356	struct bio *bio;
dd6bd0d9	357
07888c66 CH	358	bio = bio_alloc(sis->bdev, 1,
	359	REQ_OP_WRITE \| REQ_SWAP \| wbc_to_write_flags(wbc),
	360	GFP_NOIO);
48d15436	361	bio->bi_iter.bi_sector = swap_page_sector(page);
cf1e3fe4	362	bio->bi_end_io = end_swap_bio_write;
cb58bf91	363	__bio_add_page(bio, page, thp_size(page), 0);
48d15436	364
6a7f6d86	365	bio_associate_blkg_from_page(bio, page);
225311a4	366	count_swpout_vm_event(page);
1da177e4 LT	367	set_page_writeback(page);
1da177e4 LT	368	unlock_page(page);
4e49ea4a	369	submit_bio(bio);
1da177e4	370	}
548d9782	371
05cda97e CH	372	void __swap_writepage(struct page page, struct writeback_control wbc)
	373	{
	374	struct swap_info_struct *sis = page_swap_info(page);
	375
	376	VM_BUG_ON_PAGE(!PageSwapCache(page), page);
	377	/*
	378	* ->flags can be updated non-atomicially (scan_swap_map_slots),
	379	* but that will never affect SWP_FS_OPS, so the data_race
	380	* is safe.
	381	*/
	382	if (data_race(sis->flags & SWP_FS_OPS))
	383	swap_writepage_fs(page, wbc);
3222d8c2 CH	384	else if (sis->flags & SWP_SYNCHRONOUS_IO)
3222d8c2 CH	385	swap_writepage_bdev_sync(page, wbc, sis);
05cda97e	386	else
3222d8c2	387	swap_writepage_bdev_async(page, wbc, sis);
1da177e4 LT	388	}
1da177e4 LT	389
2282679f N	390	void swap_write_unplug(struct swap_iocb *sio)
	391	{
	392	struct iov_iter from;
	393	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	394	int ret;
	395
de4eda9d	396	iov_iter_bvec(&from, ITER_SOURCE, sio->bvec, sio->pages, sio->len);
2282679f N	397	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	398	if (ret != -EIOCBQUEUED)
	399	sio_write_complete(&sio->iocb, ret);
	400	}
	401
e1209d3a N	402	static void sio_read_complete(struct kiocb *iocb, long ret)
	403	{
	404	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
5169b844	405	int p;
e1209d3a	406
a1a0dfd5	407	if (ret == sio->len) {
5169b844 N	408	for (p = 0; p < sio->pages; p++) {
	409	struct page *page = sio->bvec[p].bv_page;
	410
	411	SetPageUptodate(page);
	412	unlock_page(page);
	413	}
	414	count_vm_events(PSWPIN, sio->pages);
e1209d3a	415	} else {
5169b844 N	416	for (p = 0; p < sio->pages; p++) {
	417	struct page *page = sio->bvec[p].bv_page;
	418
	419	SetPageError(page);
	420	ClearPageUptodate(page);
	421	unlock_page(page);
	422	}
	423	pr_alert_ratelimited("Read-error on swap-device\n");
e1209d3a	424	}
e1209d3a N	425	mempool_free(sio, sio_pool);
	426	}
	427
5169b844 N	428	static void swap_readpage_fs(struct page *page,
5169b844 N	429	struct swap_iocb **plug)
e1209d3a N	430	{
e1209d3a N	431	struct swap_info_struct *sis = page_swap_info(page);
5169b844	432	struct swap_iocb *sio = NULL;
e1209d3a	433	loff_t pos = page_file_offset(page);
e1209d3a	434
5169b844 N	435	if (plug)
	436	sio = *plug;
	437	if (sio) {
	438	if (sio->iocb.ki_filp != sis->swap_file \|\|
a1a0dfd5	439	sio->iocb.ki_pos + sio->len != pos) {
5169b844 N	440	swap_read_unplug(sio);
	441	sio = NULL;
	442	}
	443	}
	444	if (!sio) {
	445	sio = mempool_alloc(sio_pool, GFP_KERNEL);
	446	init_sync_kiocb(&sio->iocb, sis->swap_file);
	447	sio->iocb.ki_pos = pos;
	448	sio->iocb.ki_complete = sio_read_complete;
	449	sio->pages = 0;
a1a0dfd5	450	sio->len = 0;
5169b844	451	}
8976fa6d	452	bvec_set_page(&sio->bvec[sio->pages], page, thp_size(page), 0);
a1a0dfd5	453	sio->len += thp_size(page);
5169b844 N	454	sio->pages += 1;
	455	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !plug) {
	456	swap_read_unplug(sio);
	457	sio = NULL;
	458	}
	459	if (plug)
	460	*plug = sio;
e1209d3a N	461	}
e1209d3a N	462
9b4e30bd	463	static void swap_readpage_bdev_sync(struct page *page,
14bd75f5	464	struct swap_info_struct *sis)
1da177e4	465	{
9b4e30bd CH	466	struct bio_vec bv;
9b4e30bd CH	467	struct bio bio;
62c230bc	468
9b4e30bd CH	469	bio_init(&bio, sis->bdev, &bv, 1, REQ_OP_READ);
9b4e30bd CH	470	bio.bi_iter.bi_sector = swap_page_sector(page);
cb58bf91	471	__bio_add_page(&bio, page, thp_size(page), 0);
b0ba2d0f TH	472	/*
	473	* Keep this task valid during swap readpage because the oom killer may
	474	* attempt to access it in the page fault retry time check.
	475	*/
9b4e30bd	476	get_task_struct(current);
f8891e5e	477	count_vm_event(PSWPIN);
9b4e30bd CH	478	submit_bio_wait(&bio);
	479	__end_swap_bio_read(&bio);
	480	put_task_struct(current);
	481	}
	482
	483	static void swap_readpage_bdev_async(struct page *page,
	484	struct swap_info_struct *sis)
1da177e4 LT	485	{
1da177e4 LT	486	struct bio *bio;
23955622	487
9b4e30bd CH	488	bio = bio_alloc(sis->bdev, 1, REQ_OP_READ, GFP_KERNEL);
	489	bio->bi_iter.bi_sector = swap_page_sector(page);
	490	bio->bi_end_io = end_swap_bio_read;
cb58bf91	491	__bio_add_page(bio, page, thp_size(page), 0);
9b4e30bd CH	492	count_vm_event(PSWPIN);
9b4e30bd CH	493	submit_bio(bio);
14bd75f5 CH	494	}
	495
	496	void swap_readpage(struct page page, bool synchronous, struct swap_iocb *plug)
	497	{
62c230bc	498	struct swap_info_struct *sis = page_swap_info(page);
d8c47cc7	499	bool workingset = PageWorkingset(page);
93779069	500	unsigned long pflags;
3a9bb7b1	501	bool in_thrashing;
1da177e4	502
0bcac06f	503	VM_BUG_ON_PAGE(!PageSwapCache(page) && !synchronous, page);
309381fe SL	504	VM_BUG_ON_PAGE(!PageLocked(page), page);
309381fe SL	505	VM_BUG_ON_PAGE(PageUptodate(page), page);
93779069 MK	506
93779069 MK	507	/*
3a9bb7b1 YY	508	* Count submission time as memory stall and delay. When the device
	509	* is congested, or the submitting cgroup IO-throttled, submission
	510	* can be a significant part of overall IO time.
93779069	511	*/
3a9bb7b1 YY	512	if (workingset) {
3a9bb7b1 YY	513	delayacct_thrashing_start(&in_thrashing);
d8c47cc7	514	psi_memstall_enter(&pflags);
3a9bb7b1	515	}
a3d5dc90	516	delayacct_swapin_start();
93779069	517
165c8aed	518	if (frontswap_load(page) == 0) {
38b5faf4 DM	519	SetPageUptodate(page);
38b5faf4 DM	520	unlock_page(page);
14bd75f5	521	} else if (data_race(sis->flags & SWP_FS_OPS)) {
5169b844	522	swap_readpage_fs(page, plug);
3222d8c2	523	} else if (synchronous \|\| (sis->flags & SWP_SYNCHRONOUS_IO)) {
9b4e30bd	524	swap_readpage_bdev_sync(page, sis);
14bd75f5	525	} else {
9b4e30bd	526	swap_readpage_bdev_async(page, sis);
23955622	527	}
23955622	528
3a9bb7b1 YY	529	if (workingset) {
3a9bb7b1 YY	530	delayacct_thrashing_end(&in_thrashing);
d8c47cc7	531	psi_memstall_leave(&pflags);
3a9bb7b1	532	}
a3d5dc90	533	delayacct_swapin_end();
1da177e4	534	}
62c230bc	535
5169b844	536	void __swap_read_unplug(struct swap_iocb *sio)
62c230bc	537	{
5169b844 N	538	struct iov_iter from;
	539	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	540	int ret;
cc30c5d6	541
de4eda9d	542	iov_iter_bvec(&from, ITER_DEST, sio->bvec, sio->pages, sio->len);
5169b844 N	543	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	544	if (ret != -EIOCBQUEUED)
	545	sio_read_complete(&sio->iocb, ret);
62c230bc	546	}