[thirdparty/linux.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);
	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 = 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;
}

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

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

	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;

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

	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) {
		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
cf1e3fe4	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 LT	50	end_page_writeback(page);
1da177e4 LT	51	bio_put(bio);
1da177e4 LT	52	}
1da177e4 LT	53
9b4e30bd	54	static void __end_swap_bio_read(struct bio *bio)
1da177e4	55	{
263663cd	56	struct page *page = bio_first_page_all(bio);
1da177e4	57
4e4cbee9	58	if (bio->bi_status) {
1da177e4 LT	59	SetPageError(page);
1da177e4 LT	60	ClearPageUptodate(page);
25eaab43 GD	61	pr_alert_ratelimited("Read-error on swap-device (%u:%u:%llu)\n",
	62	MAJOR(bio_dev(bio)), MINOR(bio_dev(bio)),
	63	(unsigned long long)bio->bi_iter.bi_sector);
9b4e30bd CH	64	} else {
9b4e30bd CH	65	SetPageUptodate(page);
1da177e4 LT	66	}
1da177e4 LT	67	unlock_page(page);
9b4e30bd CH	68	}
	69
	70	static void end_swap_bio_read(struct bio *bio)
	71	{
	72	__end_swap_bio_read(bio);
1da177e4	73	bio_put(bio);
1da177e4 LT	74	}
1da177e4 LT	75
a509bc1a MG	76	int generic_swapfile_activate(struct swap_info_struct *sis,
	77	struct file *swap_file,
	78	sector_t *span)
	79	{
	80	struct address_space *mapping = swap_file->f_mapping;
	81	struct inode *inode = mapping->host;
	82	unsigned blocks_per_page;
	83	unsigned long page_no;
	84	unsigned blkbits;
	85	sector_t probe_block;
	86	sector_t last_block;
	87	sector_t lowest_block = -1;
	88	sector_t highest_block = 0;
	89	int nr_extents = 0;
	90	int ret;
	91
	92	blkbits = inode->i_blkbits;
	93	blocks_per_page = PAGE_SIZE >> blkbits;
	94
	95	/*
4efaceb1	96	* Map all the blocks into the extent tree. This code doesn't try
a509bc1a MG	97	* to be very smart.
	98	*/
	99	probe_block = 0;
	100	page_no = 0;
	101	last_block = i_size_read(inode) >> blkbits;
	102	while ((probe_block + blocks_per_page) <= last_block &&
	103	page_no < sis->max) {
	104	unsigned block_in_page;
	105	sector_t first_block;
	106
7e4411bf MP	107	cond_resched();
7e4411bf MP	108
30460e1e CM	109	first_block = probe_block;
	110	ret = bmap(inode, &first_block);
	111	if (ret \|\| !first_block)
a509bc1a MG	112	goto bad_bmap;
	113
	114	/*
	115	* It must be PAGE_SIZE aligned on-disk
	116	*/
	117	if (first_block & (blocks_per_page - 1)) {
	118	probe_block++;
	119	goto reprobe;
	120	}
	121
	122	for (block_in_page = 1; block_in_page < blocks_per_page;
	123	block_in_page++) {
	124	sector_t block;
	125
30460e1e CM	126	block = probe_block + block_in_page;
	127	ret = bmap(inode, &block);
	128	if (ret \|\| !block)
a509bc1a	129	goto bad_bmap;
30460e1e	130
a509bc1a MG	131	if (block != first_block + block_in_page) {
	132	/* Discontiguity */
	133	probe_block++;
	134	goto reprobe;
	135	}
	136	}
	137
	138	first_block >>= (PAGE_SHIFT - blkbits);
	139	if (page_no) { /* exclude the header page */
	140	if (first_block < lowest_block)
	141	lowest_block = first_block;
	142	if (first_block > highest_block)
	143	highest_block = first_block;
	144	}
	145
	146	/*
	147	* We found a PAGE_SIZE-length, PAGE_SIZE-aligned run of blocks
	148	*/
	149	ret = add_swap_extent(sis, page_no, 1, first_block);
	150	if (ret < 0)
	151	goto out;
	152	nr_extents += ret;
	153	page_no++;
	154	probe_block += blocks_per_page;
	155	reprobe:
	156	continue;
	157	}
	158	ret = nr_extents;
	159	*span = 1 + highest_block - lowest_block;
	160	if (page_no == 0)
	161	page_no = 1; /* force Empty message */
	162	sis->max = page_no;
	163	sis->pages = page_no - 1;
	164	sis->highest_bit = page_no - 1;
	165	out:
	166	return ret;
	167	bad_bmap:
1170532b	168	pr_err("swapon: swapfile has holes\n");
a509bc1a MG	169	ret = -EINVAL;
	170	goto out;
	171	}
	172
1da177e4 LT	173	/*
	174	* We may have stale swap cache pages in memory: notice
	175	* them here and get rid of the unnecessary final write.
	176	*/
	177	int swap_writepage(struct page page, struct writeback_control wbc)
	178	{
71fa1a53	179	struct folio *folio = page_folio(page);
2f772e6c	180	int ret = 0;
1da177e4	181
71fa1a53 MWO	182	if (folio_free_swap(folio)) {
71fa1a53 MWO	183	folio_unlock(folio);
1da177e4 LT	184	goto out;
1da177e4 LT	185	}
8a84802e SP	186	/*
	187	* Arch code may have to preserve more data than just the page
	188	* contents, e.g. memory tags.
	189	*/
71fa1a53	190	ret = arch_prepare_to_swap(&folio->page);
8a84802e	191	if (ret) {
71fa1a53 MWO	192	folio_mark_dirty(folio);
71fa1a53 MWO	193	folio_unlock(folio);
8a84802e SP	194	goto out;
8a84802e SP	195	}
71fa1a53 MWO	196	if (frontswap_store(&folio->page) == 0) {
	197	folio_start_writeback(folio);
	198	folio_unlock(folio);
	199	folio_end_writeback(folio);
38b5faf4 DM	200	goto out;
38b5faf4 DM	201	}
71fa1a53	202	ret = __swap_writepage(&folio->page, wbc);
2f772e6c SJ	203	out:
	204	return ret;
	205	}
	206
225311a4 HY	207	static inline void count_swpout_vm_event(struct page *page)
	208	{
	209	#ifdef CONFIG_TRANSPARENT_HUGEPAGE
	210	if (unlikely(PageTransHuge(page)))
	211	count_vm_event(THP_SWPOUT);
	212	#endif
6c357848	213	count_vm_events(PSWPOUT, thp_nr_pages(page));
225311a4 HY	214	}
225311a4 HY	215
a18b9b15 CH	216	#if defined(CONFIG_MEMCG) && defined(CONFIG_BLK_CGROUP)
	217	static void bio_associate_blkg_from_page(struct bio bio, struct page page)
	218	{
	219	struct cgroup_subsys_state *css;
bcfe06bf	220	struct mem_cgroup *memcg;
a18b9b15	221
bcfe06bf RG	222	memcg = page_memcg(page);
bcfe06bf RG	223	if (!memcg)
a18b9b15 CH	224	return;
	225
	226	rcu_read_lock();
bcfe06bf	227	css = cgroup_e_css(memcg->css.cgroup, &io_cgrp_subsys);
a18b9b15 CH	228	bio_associate_blkg_from_css(bio, css);
	229	rcu_read_unlock();
	230	}
	231	#else
	232	#define bio_associate_blkg_from_page(bio, page) do { } while (0)
	233	#endif /* CONFIG_MEMCG && CONFIG_BLK_CGROUP */
	234
e1209d3a N	235	struct swap_iocb {
e1209d3a N	236	struct kiocb iocb;
5169b844 N	237	struct bio_vec bvec[SWAP_CLUSTER_MAX];
5169b844 N	238	int pages;
a1a0dfd5	239	int len;
e1209d3a N	240	};
	241	static mempool_t *sio_pool;
	242
	243	int sio_pool_init(void)
2f772e6c	244	{
e1209d3a N	245	if (!sio_pool) {
	246	mempool_t *pool = mempool_create_kmalloc_pool(
	247	SWAP_CLUSTER_MAX, sizeof(struct swap_iocb));
	248	if (cmpxchg(&sio_pool, NULL, pool))
	249	mempool_destroy(pool);
	250	}
	251	if (!sio_pool)
	252	return -ENOMEM;
	253	return 0;
	254	}
62c230bc	255
7eadabc0 N	256	static void sio_write_complete(struct kiocb *iocb, long ret)
	257	{
	258	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
5169b844	259	struct page *page = sio->bvec[0].bv_page;
2282679f	260	int p;
62c230bc	261
a1a0dfd5	262	if (ret != sio->len) {
7eadabc0 N	263	/*
	264	* In the case of swap-over-nfs, this can be a
	265	* temporary failure if the system has limited
	266	* memory for allocating transmit buffers.
	267	* Mark the page dirty and avoid
	268	* folio_rotate_reclaimable but rate-limit the
	269	* messages but do not flag PageError like
	270	* the normal direct-to-bio case as it could
	271	* be temporary.
	272	*/
7eadabc0 N	273	pr_err_ratelimited("Write error %ld on dio swapfile (%llu)\n",
7eadabc0 N	274	ret, page_file_offset(page));
2282679f N	275	for (p = 0; p < sio->pages; p++) {
2282679f N	276	page = sio->bvec[p].bv_page;
2d30d31e	277	set_page_dirty(page);
0cdc444a	278	ClearPageReclaim(page);
62c230bc	279	}
6341a446 N	280	} else {
	281	for (p = 0; p < sio->pages; p++)
	282	count_swpout_vm_event(sio->bvec[p].bv_page);
62c230bc MG	283	}
62c230bc MG	284
2282679f N	285	for (p = 0; p < sio->pages; p++)
	286	end_page_writeback(sio->bvec[p].bv_page);
	287
7eadabc0 N	288	mempool_free(sio, sio_pool);
	289	}
	290
	291	static int swap_writepage_fs(struct page page, struct writeback_control wbc)
	292	{
2282679f	293	struct swap_iocb *sio = NULL;
7eadabc0 N	294	struct swap_info_struct *sis = page_swap_info(page);
7eadabc0 N	295	struct file *swap_file = sis->swap_file;
2282679f	296	loff_t pos = page_file_offset(page);
7eadabc0 N	297
	298	set_page_writeback(page);
	299	unlock_page(page);
2282679f N	300	if (wbc->swap_plug)
	301	sio = *wbc->swap_plug;
	302	if (sio) {
	303	if (sio->iocb.ki_filp != swap_file \|\|
a1a0dfd5	304	sio->iocb.ki_pos + sio->len != pos) {
2282679f N	305	swap_write_unplug(sio);
	306	sio = NULL;
	307	}
	308	}
	309	if (!sio) {
	310	sio = mempool_alloc(sio_pool, GFP_NOIO);
	311	init_sync_kiocb(&sio->iocb, swap_file);
	312	sio->iocb.ki_complete = sio_write_complete;
	313	sio->iocb.ki_pos = pos;
	314	sio->pages = 0;
a1a0dfd5	315	sio->len = 0;
2282679f N	316	}
2282679f N	317	sio->bvec[sio->pages].bv_page = page;
a1a0dfd5	318	sio->bvec[sio->pages].bv_len = thp_size(page);
2282679f	319	sio->bvec[sio->pages].bv_offset = 0;
a1a0dfd5	320	sio->len += thp_size(page);
2282679f N	321	sio->pages += 1;
	322	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !wbc->swap_plug) {
	323	swap_write_unplug(sio);
	324	sio = NULL;
	325	}
	326	if (wbc->swap_plug)
	327	*wbc->swap_plug = sio;
	328
	329	return 0;
7eadabc0 N	330	}
7eadabc0 N	331
cf1e3fe4	332	int __swap_writepage(struct page page, struct writeback_control wbc)
2f772e6c SJ	333	{
2f772e6c SJ	334	struct bio *bio;
4e49ea4a	335	int ret;
2f772e6c	336	struct swap_info_struct *sis = page_swap_info(page);
62c230bc	337
cc30c5d6	338	VM_BUG_ON_PAGE(!PageSwapCache(page), page);
7eadabc0 N	339	/*
	340	* ->flags can be updated non-atomicially (scan_swap_map_slots),
	341	* but that will never affect SWP_FS_OPS, so the data_race
	342	* is safe.
	343	*/
	344	if (data_race(sis->flags & SWP_FS_OPS))
	345	return swap_writepage_fs(page, wbc);
62c230bc	346
dd6bd0d9 MW	347	ret = bdev_write_page(sis->bdev, swap_page_sector(page), page, wbc);
dd6bd0d9 MW	348	if (!ret) {
225311a4	349	count_swpout_vm_event(page);
dd6bd0d9 MW	350	return 0;
	351	}
	352
07888c66 CH	353	bio = bio_alloc(sis->bdev, 1,
	354	REQ_OP_WRITE \| REQ_SWAP \| wbc_to_write_flags(wbc),
	355	GFP_NOIO);
48d15436	356	bio->bi_iter.bi_sector = swap_page_sector(page);
cf1e3fe4	357	bio->bi_end_io = end_swap_bio_write;
48d15436 CH	358	bio_add_page(bio, page, thp_size(page), 0);
48d15436 CH	359
6a7f6d86	360	bio_associate_blkg_from_page(bio, page);
225311a4	361	count_swpout_vm_event(page);
1da177e4 LT	362	set_page_writeback(page);
1da177e4 LT	363	unlock_page(page);
4e49ea4a	364	submit_bio(bio);
548d9782 ML	365
548d9782 ML	366	return 0;
1da177e4 LT	367	}
1da177e4 LT	368
2282679f N	369	void swap_write_unplug(struct swap_iocb *sio)
	370	{
	371	struct iov_iter from;
	372	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	373	int ret;
	374
de4eda9d	375	iov_iter_bvec(&from, ITER_SOURCE, sio->bvec, sio->pages, sio->len);
2282679f N	376	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	377	if (ret != -EIOCBQUEUED)
	378	sio_write_complete(&sio->iocb, ret);
	379	}
	380
e1209d3a N	381	static void sio_read_complete(struct kiocb *iocb, long ret)
	382	{
	383	struct swap_iocb *sio = container_of(iocb, struct swap_iocb, iocb);
5169b844	384	int p;
e1209d3a	385
a1a0dfd5	386	if (ret == sio->len) {
5169b844 N	387	for (p = 0; p < sio->pages; p++) {
	388	struct page *page = sio->bvec[p].bv_page;
	389
	390	SetPageUptodate(page);
	391	unlock_page(page);
	392	}
	393	count_vm_events(PSWPIN, sio->pages);
e1209d3a	394	} else {
5169b844 N	395	for (p = 0; p < sio->pages; p++) {
	396	struct page *page = sio->bvec[p].bv_page;
	397
	398	SetPageError(page);
	399	ClearPageUptodate(page);
	400	unlock_page(page);
	401	}
	402	pr_alert_ratelimited("Read-error on swap-device\n");
e1209d3a	403	}
e1209d3a N	404	mempool_free(sio, sio_pool);
	405	}
	406
5169b844 N	407	static void swap_readpage_fs(struct page *page,
5169b844 N	408	struct swap_iocb **plug)
e1209d3a N	409	{
e1209d3a N	410	struct swap_info_struct *sis = page_swap_info(page);
5169b844	411	struct swap_iocb *sio = NULL;
e1209d3a	412	loff_t pos = page_file_offset(page);
e1209d3a	413
5169b844 N	414	if (plug)
	415	sio = *plug;
	416	if (sio) {
	417	if (sio->iocb.ki_filp != sis->swap_file \|\|
a1a0dfd5	418	sio->iocb.ki_pos + sio->len != pos) {
5169b844 N	419	swap_read_unplug(sio);
	420	sio = NULL;
	421	}
	422	}
	423	if (!sio) {
	424	sio = mempool_alloc(sio_pool, GFP_KERNEL);
	425	init_sync_kiocb(&sio->iocb, sis->swap_file);
	426	sio->iocb.ki_pos = pos;
	427	sio->iocb.ki_complete = sio_read_complete;
	428	sio->pages = 0;
a1a0dfd5	429	sio->len = 0;
5169b844 N	430	}
5169b844 N	431	sio->bvec[sio->pages].bv_page = page;
a1a0dfd5	432	sio->bvec[sio->pages].bv_len = thp_size(page);
5169b844	433	sio->bvec[sio->pages].bv_offset = 0;
a1a0dfd5	434	sio->len += thp_size(page);
5169b844 N	435	sio->pages += 1;
	436	if (sio->pages == ARRAY_SIZE(sio->bvec) \|\| !plug) {
	437	swap_read_unplug(sio);
	438	sio = NULL;
	439	}
	440	if (plug)
	441	*plug = sio;
e1209d3a N	442	}
e1209d3a N	443
9b4e30bd	444	static void swap_readpage_bdev_sync(struct page *page,
14bd75f5	445	struct swap_info_struct *sis)
1da177e4	446	{
9b4e30bd CH	447	struct bio_vec bv;
9b4e30bd CH	448	struct bio bio;
62c230bc	449
a8c1408f CH	450	if ((sis->flags & SWP_SYNCHRONOUS_IO) &&
	451	!bdev_read_page(sis->bdev, swap_page_sector(page), page)) {
	452	count_vm_event(PSWPIN);
14bd75f5	453	return;
dd6bd0d9 MW	454	}
dd6bd0d9 MW	455
9b4e30bd CH	456	bio_init(&bio, sis->bdev, &bv, 1, REQ_OP_READ);
	457	bio.bi_iter.bi_sector = swap_page_sector(page);
	458	bio_add_page(&bio, page, thp_size(page), 0);
b0ba2d0f TH	459	/*
	460	* Keep this task valid during swap readpage because the oom killer may
	461	* attempt to access it in the page fault retry time check.
	462	*/
9b4e30bd	463	get_task_struct(current);
f8891e5e	464	count_vm_event(PSWPIN);
9b4e30bd CH	465	submit_bio_wait(&bio);
	466	__end_swap_bio_read(&bio);
	467	put_task_struct(current);
	468	}
	469
	470	static void swap_readpage_bdev_async(struct page *page,
	471	struct swap_info_struct *sis)
	472	{
	473	struct bio *bio;
23955622	474
9b4e30bd CH	475	if ((sis->flags & SWP_SYNCHRONOUS_IO) &&
	476	!bdev_read_page(sis->bdev, swap_page_sector(page), page)) {
	477	count_vm_event(PSWPIN);
	478	return;
23955622	479	}
9b4e30bd CH	480
	481	bio = bio_alloc(sis->bdev, 1, REQ_OP_READ, GFP_KERNEL);
	482	bio->bi_iter.bi_sector = swap_page_sector(page);
	483	bio->bi_end_io = end_swap_bio_read;
	484	bio_add_page(bio, page, thp_size(page), 0);
	485	count_vm_event(PSWPIN);
	486	submit_bio(bio);
14bd75f5 CH	487	}
	488
	489	void swap_readpage(struct page page, bool synchronous, struct swap_iocb *plug)
	490	{
	491	struct swap_info_struct *sis = page_swap_info(page);
	492	bool workingset = PageWorkingset(page);
	493	unsigned long pflags;
	494	bool in_thrashing;
	495
	496	VM_BUG_ON_PAGE(!PageSwapCache(page) && !synchronous, page);
	497	VM_BUG_ON_PAGE(!PageLocked(page), page);
	498	VM_BUG_ON_PAGE(PageUptodate(page), page);
	499
	500	/*
	501	* Count submission time as memory stall and delay. When the device
	502	* is congested, or the submitting cgroup IO-throttled, submission
	503	* can be a significant part of overall IO time.
	504	*/
	505	if (workingset) {
	506	delayacct_thrashing_start(&in_thrashing);
	507	psi_memstall_enter(&pflags);
	508	}
	509	delayacct_swapin_start();
	510
	511	if (frontswap_load(page) == 0) {
	512	SetPageUptodate(page);
	513	unlock_page(page);
	514	} else if (data_race(sis->flags & SWP_FS_OPS)) {
	515	swap_readpage_fs(page, plug);
9b4e30bd CH	516	} else if (synchronous) {
9b4e30bd CH	517	swap_readpage_bdev_sync(page, sis);
14bd75f5	518	} else {
9b4e30bd	519	swap_readpage_bdev_async(page, sis);
14bd75f5	520	}
23955622	521
3a9bb7b1 YY	522	if (workingset) {
3a9bb7b1 YY	523	delayacct_thrashing_end(&in_thrashing);
d8c47cc7	524	psi_memstall_leave(&pflags);
3a9bb7b1	525	}
a3d5dc90	526	delayacct_swapin_end();
1da177e4	527	}
62c230bc	528
5169b844	529	void __swap_read_unplug(struct swap_iocb *sio)
62c230bc	530	{
5169b844 N	531	struct iov_iter from;
	532	struct address_space *mapping = sio->iocb.ki_filp->f_mapping;
	533	int ret;
cc30c5d6	534
de4eda9d	535	iov_iter_bvec(&from, ITER_DEST, sio->bvec, sio->pages, sio->len);
5169b844 N	536	ret = mapping->a_ops->swap_rw(&sio->iocb, &from);
	537	if (ret != -EIOCBQUEUED)
	538	sio_read_complete(&sio->iocb, ret);
62c230bc	539	}