^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * linux/mm/swapfile.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * Swap reorganised 29.12.95, Stephen Tweedie
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include <linux/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include <linux/sched/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include <linux/sched/task.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include <linux/hugetlb.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include <linux/mman.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include <linux/slab.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <linux/kernel_stat.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <linux/swap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <linux/vmalloc.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <linux/pagemap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/namei.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <linux/shmem_fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <linux/blkdev.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include <linux/random.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include <linux/writeback.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include <linux/proc_fs.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include <linux/seq_file.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include <linux/init.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include <linux/ksm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include <linux/rmap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include <linux/security.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include <linux/backing-dev.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include <linux/mutex.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) #include <linux/capability.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #include <linux/syscalls.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) #include <linux/memcontrol.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) #include <linux/poll.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) #include <linux/oom.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) #include <linux/frontswap.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) #include <linux/swapfile.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) #include <linux/export.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) #include <linux/swap_slots.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) #include <linux/sort.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) #include <asm/tlbflush.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) #include <linux/swapops.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) #include <linux/swap_cgroup.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) #include <trace/hooks/mm.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) static bool swap_count_continued(struct swap_info_struct *, pgoff_t,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) unsigned char);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) static void free_swap_count_continuations(struct swap_info_struct *);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) static sector_t map_swap_entry(swp_entry_t, struct block_device**);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) DEFINE_SPINLOCK(swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) static unsigned int nr_swapfiles;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) atomic_long_t nr_swap_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) * Some modules use swappable objects and may try to swap them out under
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) * memory pressure (via the shrinker). Before doing so, they may wish to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * check to see if any swap space is available.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) EXPORT_SYMBOL_GPL(nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) /* protected with swap_lock. reading in vm_swap_full() doesn't need lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) long total_swap_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) static int least_priority = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) static const char Bad_file[] = "Bad swap file entry ";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) static const char Unused_file[] = "Unused swap file entry ";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) static const char Bad_offset[] = "Bad swap offset entry ";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) static const char Unused_offset[] = "Unused swap offset entry ";
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * all active swap_info_structs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * protected with swap_lock, and ordered by priority.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) PLIST_HEAD(swap_active_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) * all available (active, not full) swap_info_structs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) * protected with swap_avail_lock, ordered by priority.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) * This is used by get_swap_page() instead of swap_active_head
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) * because swap_active_head includes all swap_info_structs,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) * but get_swap_page() doesn't need to look at full ones.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) * This uses its own lock instead of swap_lock because when a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * swap_info_struct changes between not-full/full, it needs to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) * add/remove itself to/from this list, but the swap_info_struct->lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * is held and the locking order requires swap_lock to be taken
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) * before any swap_info_struct->lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) static struct plist_head *swap_avail_heads;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) static DEFINE_SPINLOCK(swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) struct swap_info_struct *swap_info[MAX_SWAPFILES];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) static DEFINE_MUTEX(swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) static DECLARE_WAIT_QUEUE_HEAD(proc_poll_wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) /* Activity counter to indicate that a swapon or swapoff has occurred */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) static atomic_t proc_poll_event = ATOMIC_INIT(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) atomic_t nr_rotate_swap = ATOMIC_INIT(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) struct swap_info_struct *swap_type_to_swap_info(int type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) if (type >= READ_ONCE(nr_swapfiles))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) smp_rmb(); /* Pairs with smp_wmb in alloc_swap_info. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) return READ_ONCE(swap_info[type]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) EXPORT_SYMBOL_GPL(swap_type_to_swap_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) static inline unsigned char swap_count(unsigned char ent)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) return ent & ~SWAP_HAS_CACHE; /* may include COUNT_CONTINUED flag */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) /* Reclaim the swap entry anyway if possible */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) #define TTRS_ANYWAY 0x1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) * Reclaim the swap entry if there are no more mappings of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) * corresponding page
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) #define TTRS_UNMAPPED 0x2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) /* Reclaim the swap entry if swap is getting full*/
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) #define TTRS_FULL 0x4
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) /* returns 1 if swap entry is freed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) static int __try_to_reclaim_swap(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) unsigned long offset, unsigned long flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) swp_entry_t entry = swp_entry(si->type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) page = find_get_page(swap_address_space(entry), offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) if (!page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) * When this function is called from scan_swap_map_slots() and it's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) * called by vmscan.c at reclaiming pages. So, we hold a lock on a page,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) * here. We have to use trylock for avoiding deadlock. This is a special
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) * case and you should use try_to_free_swap() with explicit lock_page()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) * in usual operations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) if (trylock_page(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) if ((flags & TTRS_ANYWAY) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) ((flags & TTRS_UNMAPPED) && !page_mapped(page)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) ((flags & TTRS_FULL) && mem_cgroup_swap_full(page)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) ret = try_to_free_swap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) static inline struct swap_extent *first_se(struct swap_info_struct *sis)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) struct rb_node *rb = rb_first(&sis->swap_extent_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) return rb_entry(rb, struct swap_extent, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) static inline struct swap_extent *next_se(struct swap_extent *se)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) struct rb_node *rb = rb_next(&se->rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) return rb ? rb_entry(rb, struct swap_extent, rb_node) : NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) * swapon tell device that all the old swap contents can be discarded,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) * to allow the swap device to optimize its wear-levelling.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) static int discard_swap(struct swap_info_struct *si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) struct swap_extent *se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) sector_t start_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) sector_t nr_blocks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) int err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) /* Do not discard the swap header page! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) se = first_se(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) start_block = (se->start_block + 1) << (PAGE_SHIFT - 9);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) nr_blocks = ((sector_t)se->nr_pages - 1) << (PAGE_SHIFT - 9);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) if (nr_blocks) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) err = blkdev_issue_discard(si->bdev, start_block,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) nr_blocks, GFP_KERNEL, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) for (se = next_se(se); se; se = next_se(se)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) start_block = se->start_block << (PAGE_SHIFT - 9);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) nr_blocks = (sector_t)se->nr_pages << (PAGE_SHIFT - 9);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) err = blkdev_issue_discard(si->bdev, start_block,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) nr_blocks, GFP_KERNEL, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) return err; /* That will often be -EOPNOTSUPP */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) static struct swap_extent *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) offset_to_swap_extent(struct swap_info_struct *sis, unsigned long offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) struct swap_extent *se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) struct rb_node *rb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) rb = sis->swap_extent_root.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) while (rb) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) se = rb_entry(rb, struct swap_extent, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) if (offset < se->start_page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) rb = rb->rb_left;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) else if (offset >= se->start_page + se->nr_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) rb = rb->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) return se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) /* It *must* be present */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) BUG();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) sector_t swap_page_sector(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) struct swap_info_struct *sis = page_swap_info(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) struct swap_extent *se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) sector_t sector;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) offset = __page_file_index(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) se = offset_to_swap_extent(sis, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) sector = se->start_block + (offset - se->start_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) return sector << (PAGE_SHIFT - 9);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) * swap allocation tell device that a cluster of swap can now be discarded,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) * to allow the swap device to optimize its wear-levelling.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) static void discard_swap_cluster(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) pgoff_t start_page, pgoff_t nr_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) struct swap_extent *se = offset_to_swap_extent(si, start_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) while (nr_pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) pgoff_t offset = start_page - se->start_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) sector_t start_block = se->start_block + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) sector_t nr_blocks = se->nr_pages - offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) if (nr_blocks > nr_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) nr_blocks = nr_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) start_page += nr_blocks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) nr_pages -= nr_blocks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) start_block <<= PAGE_SHIFT - 9;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) nr_blocks <<= PAGE_SHIFT - 9;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) if (blkdev_issue_discard(si->bdev, start_block,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) nr_blocks, GFP_NOIO, 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) se = next_se(se);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) #ifdef CONFIG_THP_SWAP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) #define SWAPFILE_CLUSTER HPAGE_PMD_NR
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) #define swap_entry_size(size) (size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) #define SWAPFILE_CLUSTER 256
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) * Define swap_entry_size() as constant to let compiler to optimize
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) * out some code if !CONFIG_THP_SWAP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) #define swap_entry_size(size) 1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) #define LATENCY_LIMIT 256
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) static inline void cluster_set_flag(struct swap_cluster_info *info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) unsigned int flag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) info->flags = flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) static inline unsigned int cluster_count(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) return info->data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) static inline void cluster_set_count(struct swap_cluster_info *info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) unsigned int c)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) info->data = c;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) static inline void cluster_set_count_flag(struct swap_cluster_info *info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) unsigned int c, unsigned int f)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) info->flags = f;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) info->data = c;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) static inline unsigned int cluster_next(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) return info->data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) static inline void cluster_set_next(struct swap_cluster_info *info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) unsigned int n)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) info->data = n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) static inline void cluster_set_next_flag(struct swap_cluster_info *info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) unsigned int n, unsigned int f)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) info->flags = f;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) info->data = n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) static inline bool cluster_is_free(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) return info->flags & CLUSTER_FLAG_FREE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) static inline bool cluster_is_null(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) return info->flags & CLUSTER_FLAG_NEXT_NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) static inline void cluster_set_null(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) info->flags = CLUSTER_FLAG_NEXT_NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) info->data = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) static inline bool cluster_is_huge(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) if (IS_ENABLED(CONFIG_THP_SWAP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) return info->flags & CLUSTER_FLAG_HUGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) static inline void cluster_clear_huge(struct swap_cluster_info *info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) info->flags &= ~CLUSTER_FLAG_HUGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) static inline struct swap_cluster_info *lock_cluster(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) unsigned long offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) ci = si->cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) if (ci) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) ci += offset / SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) spin_lock(&ci->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) return ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) static inline void unlock_cluster(struct swap_cluster_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) if (ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) spin_unlock(&ci->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) * Determine the locking method in use for this device. Return
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) * swap_cluster_info if SSD-style cluster-based locking is in place.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) static inline struct swap_cluster_info *lock_cluster_or_swap_info(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) struct swap_info_struct *si, unsigned long offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) /* Try to use fine-grained SSD-style locking if available: */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) /* Otherwise, fall back to traditional, coarse locking: */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) if (!ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) return ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) static inline void unlock_cluster_or_swap_info(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) struct swap_cluster_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) if (ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) static inline bool cluster_list_empty(struct swap_cluster_list *list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) return cluster_is_null(&list->head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) static inline unsigned int cluster_list_first(struct swap_cluster_list *list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) return cluster_next(&list->head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) static void cluster_list_init(struct swap_cluster_list *list)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) cluster_set_null(&list->head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) cluster_set_null(&list->tail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) static void cluster_list_add_tail(struct swap_cluster_list *list,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) struct swap_cluster_info *ci,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) unsigned int idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) if (cluster_list_empty(list)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) cluster_set_next_flag(&list->head, idx, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) cluster_set_next_flag(&list->tail, idx, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) struct swap_cluster_info *ci_tail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) unsigned int tail = cluster_next(&list->tail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) * Nested cluster lock, but both cluster locks are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) * only acquired when we held swap_info_struct->lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) ci_tail = ci + tail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) spin_lock_nested(&ci_tail->lock, SINGLE_DEPTH_NESTING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) cluster_set_next(ci_tail, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) spin_unlock(&ci_tail->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) cluster_set_next_flag(&list->tail, idx, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) static unsigned int cluster_list_del_first(struct swap_cluster_list *list,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) struct swap_cluster_info *ci)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) unsigned int idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) idx = cluster_next(&list->head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) if (cluster_next(&list->tail) == idx) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) cluster_set_null(&list->head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) cluster_set_null(&list->tail);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) cluster_set_next_flag(&list->head,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) cluster_next(&ci[idx]), 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) return idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) /* Add a cluster to discard list and schedule it to do discard */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) static void swap_cluster_schedule_discard(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) unsigned int idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) * If scan_swap_map() can't find a free cluster, it will check
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) * si->swap_map directly. To make sure the discarding cluster isn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) * taken by scan_swap_map(), mark the swap entries bad (occupied). It
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) * will be cleared after discard
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) memset(si->swap_map + idx * SWAPFILE_CLUSTER,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) SWAP_MAP_BAD, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) cluster_list_add_tail(&si->discard_clusters, si->cluster_info, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) schedule_work(&si->discard_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) static void __free_cluster(struct swap_info_struct *si, unsigned long idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) struct swap_cluster_info *ci = si->cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) cluster_set_flag(ci + idx, CLUSTER_FLAG_FREE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) cluster_list_add_tail(&si->free_clusters, ci, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) * Doing discard actually. After a cluster discard is finished, the cluster
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) * will be added to free cluster list. caller should hold si->lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) static void swap_do_scheduled_discard(struct swap_info_struct *si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) struct swap_cluster_info *info, *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) unsigned int idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) info = si->cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) while (!cluster_list_empty(&si->discard_clusters)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) idx = cluster_list_del_first(&si->discard_clusters, info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) discard_swap_cluster(si, idx * SWAPFILE_CLUSTER,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) ci = lock_cluster(si, idx * SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) __free_cluster(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) memset(si->swap_map + idx * SWAPFILE_CLUSTER,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) 0, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) static void swap_discard_work(struct work_struct *work)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) si = container_of(work, struct swap_info_struct, discard_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) swap_do_scheduled_discard(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) static void alloc_cluster(struct swap_info_struct *si, unsigned long idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) struct swap_cluster_info *ci = si->cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) VM_BUG_ON(cluster_list_first(&si->free_clusters) != idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) cluster_list_del_first(&si->free_clusters, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) cluster_set_count_flag(ci + idx, 0, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) static void free_cluster(struct swap_info_struct *si, unsigned long idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) struct swap_cluster_info *ci = si->cluster_info + idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) VM_BUG_ON(cluster_count(ci) != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) * If the swap is discardable, prepare discard the cluster
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) * instead of free it immediately. The cluster will be freed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) * after discard.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) if ((si->flags & (SWP_WRITEOK | SWP_PAGE_DISCARD)) ==
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) (SWP_WRITEOK | SWP_PAGE_DISCARD)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) swap_cluster_schedule_discard(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) __free_cluster(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) * The cluster corresponding to page_nr will be used. The cluster will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) * removed from free cluster list and its usage counter will be increased.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) static void inc_cluster_info_page(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) struct swap_cluster_info *cluster_info, unsigned long page_nr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) unsigned long idx = page_nr / SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) if (!cluster_info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) if (cluster_is_free(&cluster_info[idx]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) alloc_cluster(p, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) VM_BUG_ON(cluster_count(&cluster_info[idx]) >= SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) cluster_set_count(&cluster_info[idx],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) cluster_count(&cluster_info[idx]) + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) * The cluster corresponding to page_nr decreases one usage. If the usage
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) * counter becomes 0, which means no page in the cluster is in using, we can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) * optionally discard the cluster and add it to free cluster list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) static void dec_cluster_info_page(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) struct swap_cluster_info *cluster_info, unsigned long page_nr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) unsigned long idx = page_nr / SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) if (!cluster_info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) VM_BUG_ON(cluster_count(&cluster_info[idx]) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) cluster_set_count(&cluster_info[idx],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) cluster_count(&cluster_info[idx]) - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) if (cluster_count(&cluster_info[idx]) == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) free_cluster(p, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) * It's possible scan_swap_map() uses a free cluster in the middle of free
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) * cluster list. Avoiding such abuse to avoid list corruption.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) static bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) scan_swap_map_ssd_cluster_conflict(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) unsigned long offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) struct percpu_cluster *percpu_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) bool conflict;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) offset /= SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) conflict = !cluster_list_empty(&si->free_clusters) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) offset != cluster_list_first(&si->free_clusters) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) cluster_is_free(&si->cluster_info[offset]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) if (!conflict)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) percpu_cluster = this_cpu_ptr(si->percpu_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) cluster_set_null(&percpu_cluster->index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) * Try to get a swap entry from current cpu's swap entry pool (a cluster). This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) * might involve allocating a new cluster for current CPU too.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) static bool scan_swap_map_try_ssd_cluster(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) unsigned long *offset, unsigned long *scan_base)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) struct percpu_cluster *cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) unsigned long tmp, max;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) new_cluster:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) cluster = this_cpu_ptr(si->percpu_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) if (cluster_is_null(&cluster->index)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) if (!cluster_list_empty(&si->free_clusters)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) cluster->index = si->free_clusters.head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) cluster->next = cluster_next(&cluster->index) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) } else if (!cluster_list_empty(&si->discard_clusters)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) * we don't have free cluster but have some clusters in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) * discarding, do discard now and reclaim them, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) * reread cluster_next_cpu since we dropped si->lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) swap_do_scheduled_discard(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) *scan_base = this_cpu_read(*si->cluster_next_cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) *offset = *scan_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) goto new_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) * Other CPUs can use our cluster if they can't find a free cluster,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) * check if there is still free entry in the cluster
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) tmp = cluster->next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) max = min_t(unsigned long, si->max,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) (cluster_next(&cluster->index) + 1) * SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) if (tmp < max) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) ci = lock_cluster(si, tmp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) while (tmp < max) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) if (!si->swap_map[tmp])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) tmp++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) if (tmp >= max) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) cluster_set_null(&cluster->index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) goto new_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) cluster->next = tmp + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) *offset = tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) *scan_base = tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) static void __del_from_avail_list(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) int nid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) for_each_node(nid)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) plist_del(&p->avail_lists[nid], &swap_avail_heads[nid]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) static void del_from_avail_list(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) __del_from_avail_list(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) static void swap_range_alloc(struct swap_info_struct *si, unsigned long offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) unsigned int nr_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) unsigned int end = offset + nr_entries - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) if (offset == si->lowest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) si->lowest_bit += nr_entries;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) if (end == si->highest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) WRITE_ONCE(si->highest_bit, si->highest_bit - nr_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) si->inuse_pages += nr_entries;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) if (si->inuse_pages == si->pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) si->lowest_bit = si->max;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) si->highest_bit = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) del_from_avail_list(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) static void add_to_avail_list(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) int nid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) for_each_node(nid) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) WARN_ON(!plist_node_empty(&p->avail_lists[nid]));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) plist_add(&p->avail_lists[nid], &swap_avail_heads[nid]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) static void swap_range_free(struct swap_info_struct *si, unsigned long offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) unsigned int nr_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) unsigned long begin = offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) unsigned long end = offset + nr_entries - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) void (*swap_slot_free_notify)(struct block_device *, unsigned long);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) if (offset < si->lowest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) si->lowest_bit = offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) if (end > si->highest_bit) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) bool was_full = !si->highest_bit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) WRITE_ONCE(si->highest_bit, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) if (was_full && (si->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) add_to_avail_list(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) trace_android_vh_account_swap_pages(si, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) if (!skip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) atomic_long_add(nr_entries, &nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) si->inuse_pages -= nr_entries;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) if (si->flags & SWP_BLKDEV)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) swap_slot_free_notify =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) si->bdev->bd_disk->fops->swap_slot_free_notify;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) swap_slot_free_notify = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) while (offset <= end) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) arch_swap_invalidate_page(si->type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) frontswap_invalidate_page(si->type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) if (swap_slot_free_notify)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) swap_slot_free_notify(si->bdev, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) offset++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) clear_shadow_from_swap_cache(si->type, begin, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) static void set_cluster_next(struct swap_info_struct *si, unsigned long next)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) unsigned long prev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) if (!(si->flags & SWP_SOLIDSTATE)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) si->cluster_next = next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) prev = this_cpu_read(*si->cluster_next_cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) * Cross the swap address space size aligned trunk, choose
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) * another trunk randomly to avoid lock contention on swap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) * address space if possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) if ((prev >> SWAP_ADDRESS_SPACE_SHIFT) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) (next >> SWAP_ADDRESS_SPACE_SHIFT)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) /* No free swap slots available */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) if (si->highest_bit <= si->lowest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) next = si->lowest_bit +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) prandom_u32_max(si->highest_bit - si->lowest_bit + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) next = ALIGN_DOWN(next, SWAP_ADDRESS_SPACE_PAGES);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) next = max_t(unsigned int, next, si->lowest_bit);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) this_cpu_write(*si->cluster_next_cpu, next);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) int scan_swap_map_slots(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) unsigned char usage, int nr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) swp_entry_t slots[])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) unsigned long scan_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) unsigned long last_in_cluster = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) int latency_ration = LATENCY_LIMIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) int n_ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) bool scanned_many = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) * We try to cluster swap pages by allocating them sequentially
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) * in swap. Once we've allocated SWAPFILE_CLUSTER pages this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) * way, however, we resort to first-free allocation, starting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) * a new cluster. This prevents us from scattering swap pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) * all over the entire swap partition, so that we reduce
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) * overall disk seek times between swap pages. -- sct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) * But we do now try to find an empty cluster. -Andrea
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) * And we let swap pages go all over an SSD partition. Hugh
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) si->flags += SWP_SCANNING;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) * Use percpu scan base for SSD to reduce lock contention on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) * cluster and swap cache. For HDD, sequential access is more
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) * important.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) if (si->flags & SWP_SOLIDSTATE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) scan_base = this_cpu_read(*si->cluster_next_cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) scan_base = si->cluster_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) offset = scan_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) /* SSD algorithm */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) if (si->cluster_info) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) if (!scan_swap_map_try_ssd_cluster(si, &offset, &scan_base))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) goto scan;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) } else if (unlikely(!si->cluster_nr--)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) if (si->pages - si->inuse_pages < SWAPFILE_CLUSTER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) si->cluster_nr = SWAPFILE_CLUSTER - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) * If seek is expensive, start searching for new cluster from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) * start of partition, to minimize the span of allocated swap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) * If seek is cheap, that is the SWP_SOLIDSTATE si->cluster_info
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) * case, just handled by scan_swap_map_try_ssd_cluster() above.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) scan_base = offset = si->lowest_bit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) last_in_cluster = offset + SWAPFILE_CLUSTER - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) /* Locate the first empty (unaligned) cluster */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) for (; last_in_cluster <= si->highest_bit; offset++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) if (si->swap_map[offset])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) last_in_cluster = offset + SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) else if (offset == last_in_cluster) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) offset -= SWAPFILE_CLUSTER - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) si->cluster_next = offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) si->cluster_nr = SWAPFILE_CLUSTER - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) if (unlikely(--latency_ration < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) latency_ration = LATENCY_LIMIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) offset = scan_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) si->cluster_nr = SWAPFILE_CLUSTER - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) checks:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) if (si->cluster_info) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) while (scan_swap_map_ssd_cluster_conflict(si, offset)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) /* take a break if we already got some slots */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) if (n_ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) if (!scan_swap_map_try_ssd_cluster(si, &offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) &scan_base))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) goto scan;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) if (!(si->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) goto no_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) if (!si->highest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) goto no_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) if (offset > si->highest_bit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) scan_base = offset = si->lowest_bit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) /* reuse swap entry of cache-only swap if not busy. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) int swap_was_freed;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) swap_was_freed = __try_to_reclaim_swap(si, offset, TTRS_ANYWAY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) /* entry was freed successfully, try to use this again */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) if (swap_was_freed)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) goto scan; /* check next one */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) if (si->swap_map[offset]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) if (!n_ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) goto scan;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) WRITE_ONCE(si->swap_map[offset], usage);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) inc_cluster_info_page(si, si->cluster_info, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) swap_range_alloc(si, offset, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) slots[n_ret++] = swp_entry(si->type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) /* got enough slots or reach max slots? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) if ((n_ret == nr) || (offset >= si->highest_bit))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) /* search for next available slot */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) /* time to take a break? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) if (unlikely(--latency_ration < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) if (n_ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) goto done;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912) latency_ration = LATENCY_LIMIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) /* try to get more slots in cluster */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) if (si->cluster_info) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) if (scan_swap_map_try_ssd_cluster(si, &offset, &scan_base))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) } else if (si->cluster_nr && !si->swap_map[++offset]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) /* non-ssd case, still more slots in cluster? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) --si->cluster_nr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) * Even if there's no free clusters available (fragmented),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) * try to scan a little more quickly with lock held unless we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) * have scanned too many slots already.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) if (!scanned_many) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) unsigned long scan_limit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) if (offset < scan_base)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) scan_limit = scan_base;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) scan_limit = si->highest_bit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) for (; offset <= scan_limit && --latency_ration > 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) offset++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) if (!si->swap_map[offset])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) done:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) set_cluster_next(si, offset + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) si->flags -= SWP_SCANNING;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) return n_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) scan:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) while (++offset <= READ_ONCE(si->highest_bit)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) if (data_race(!si->swap_map[offset])) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) if (vm_swap_full() &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) READ_ONCE(si->swap_map[offset]) == SWAP_HAS_CACHE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) if (unlikely(--latency_ration < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) latency_ration = LATENCY_LIMIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) scanned_many = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) offset = si->lowest_bit;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) while (offset < scan_base) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) if (data_race(!si->swap_map[offset])) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) if (vm_swap_full() &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) READ_ONCE(si->swap_map[offset]) == SWAP_HAS_CACHE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) goto checks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) if (unlikely(--latency_ration < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) latency_ration = LATENCY_LIMIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) scanned_many = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) offset++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) no_page:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) si->flags -= SWP_SCANNING;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) return n_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) EXPORT_SYMBOL_GPL(scan_swap_map_slots);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) int swap_alloc_cluster(struct swap_info_struct *si, swp_entry_t *slot)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) unsigned long idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) unsigned long offset, i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) unsigned char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) * Should not even be attempting cluster allocations when huge
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) * page swap is disabled. Warn and fail the allocation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) if (!IS_ENABLED(CONFIG_THP_SWAP)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) VM_WARN_ON_ONCE(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) if (cluster_list_empty(&si->free_clusters))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) idx = cluster_list_first(&si->free_clusters);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) offset = idx * SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) alloc_cluster(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) cluster_set_count_flag(ci, SWAPFILE_CLUSTER, CLUSTER_FLAG_HUGE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) map = si->swap_map + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) for (i = 0; i < SWAPFILE_CLUSTER; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) map[i] = SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) swap_range_alloc(si, offset, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) *slot = swp_entry(si->type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) EXPORT_SYMBOL_GPL(swap_alloc_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) static void swap_free_cluster(struct swap_info_struct *si, unsigned long idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) unsigned long offset = idx * SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) memset(si->swap_map + offset, 0, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) cluster_set_count_flag(ci, 0, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) free_cluster(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) swap_range_free(si, offset, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) static unsigned long scan_swap_map(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) unsigned char usage)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) int n_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) n_ret = scan_swap_map_slots(si, usage, 1, &entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) if (n_ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) return swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) int get_swap_pages(int n_goal, swp_entry_t swp_entries[], int entry_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059) unsigned long size = swap_entry_size(entry_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) struct swap_info_struct *si, *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061) long avail_pgs;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) int n_ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) int node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065) /* Only single cluster request supported */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) WARN_ON_ONCE(n_goal > 1 && size == SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070) avail_pgs = atomic_long_read(&nr_swap_pages) / size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) if (avail_pgs <= 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073) goto noswap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076) n_goal = min3((long)n_goal, (long)SWAP_BATCH, avail_pgs);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) atomic_long_sub(n_goal * size, &nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) start_over:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) node = numa_node_id();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) plist_for_each_entry_safe(si, next, &swap_avail_heads[node], avail_lists[node]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083) /* requeue si to after same-priority siblings */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) plist_requeue(&si->avail_lists[node], &swap_avail_heads[node]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) if (!si->highest_bit || !(si->flags & SWP_WRITEOK)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) if (plist_node_empty(&si->avail_lists[node])) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) goto nextsi;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) WARN(!si->highest_bit,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) "swap_info %d in list but !highest_bit\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) si->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096) WARN(!(si->flags & SWP_WRITEOK),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097) "swap_info %d in list but !SWP_WRITEOK\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) si->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) __del_from_avail_list(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) goto nextsi;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) if (size == SWAPFILE_CLUSTER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) if (si->flags & SWP_BLKDEV)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) n_ret = swap_alloc_cluster(si, swp_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107) n_ret = scan_swap_map_slots(si, SWAP_HAS_CACHE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108) n_goal, swp_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) if (n_ret || size == SWAPFILE_CLUSTER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) goto check_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) pr_debug("scan_swap_map of si %d failed to find offset\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) si->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) nextsi:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) * if we got here, it's likely that si was almost full before,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) * and since scan_swap_map() can drop the si->lock, multiple
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) * callers probably all tried to get a page from the same si
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121) * and it filled up before we could get one; or, the si filled
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) * up between us dropping swap_avail_lock and taking si->lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) * Since we dropped the swap_avail_lock, the swap_avail_head
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) * list may have been modified; so if next is still in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) * swap_avail_head list then try it, otherwise start over
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) * if we have not gotten any slots.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) if (plist_node_empty(&next->avail_lists[node]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129) goto start_over;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134) check_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) if (n_ret < n_goal)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) atomic_long_add((long)(n_goal - n_ret) * size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) &nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) noswap:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) return n_ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) /* The only caller of this function is now suspend routine */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) swp_entry_t get_swap_page_of_type(int type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145) struct swap_info_struct *si = swap_type_to_swap_info(type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) if (!si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) goto fail;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) if (si->flags & SWP_WRITEOK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) /* This is called for allocating swap entry, not cache */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155) offset = scan_swap_map(si, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) if (offset) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) trace_android_vh_account_swap_pages(si, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) if (!skip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) atomic_long_dec(&nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) return swp_entry(type, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) fail:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166) return (swp_entry_t) {0};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) static struct swap_info_struct *__swap_info_get(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) if (!entry.val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) p = swp_swap_info(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) if (!p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178) goto bad_nofile;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) if (data_race(!(p->flags & SWP_USED)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) goto bad_device;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) if (offset >= p->max)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) goto bad_offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) return p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) bad_offset:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) pr_err("swap_info_get: %s%08lx\n", Bad_offset, entry.val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) bad_device:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) pr_err("swap_info_get: %s%08lx\n", Unused_file, entry.val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) bad_nofile:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) pr_err("swap_info_get: %s%08lx\n", Bad_file, entry.val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) static struct swap_info_struct *_swap_info_get(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) p = __swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) if (!p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) if (data_race(!p->swap_map[swp_offset(entry)]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206) goto bad_free;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207) return p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209) bad_free:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) pr_err("swap_info_get: %s%08lx\n", Unused_offset, entry.val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215) static struct swap_info_struct *swap_info_get(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) if (p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222) return p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1223) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1225) static struct swap_info_struct *swap_info_get_cont(swp_entry_t entry,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226) struct swap_info_struct *q)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) if (p != q) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) if (q != NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234) spin_unlock(&q->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235) if (p != NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238) return p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241) static unsigned char __swap_entry_free_locked(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) unsigned long offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243) unsigned char usage)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246) unsigned char has_cache;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) count = p->swap_map[offset];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250) has_cache = count & SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251) count &= ~SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) if (usage == SWAP_HAS_CACHE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254) VM_BUG_ON(!has_cache);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) has_cache = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) } else if (count == SWAP_MAP_SHMEM) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258) * Or we could insist on shmem.c using a special
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259) * swap_shmem_free() and free_shmem_swap_and_cache()...
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262) } else if ((count & ~COUNT_CONTINUED) <= SWAP_MAP_MAX) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263) if (count == COUNT_CONTINUED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) if (swap_count_continued(p, offset, count))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265) count = SWAP_MAP_MAX | COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267) count = SWAP_MAP_MAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) count--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) usage = count | has_cache;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273) if (usage)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) WRITE_ONCE(p->swap_map[offset], usage);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) WRITE_ONCE(p->swap_map[offset], SWAP_HAS_CACHE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278) return usage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) * Check whether swap entry is valid in the swap device. If so,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283) * return pointer to swap_info_struct, and keep the swap entry valid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) * via preventing the swap device from being swapoff, until
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285) * put_swap_device() is called. Otherwise return NULL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) * The entirety of the RCU read critical section must come before the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288) * return from or after the call to synchronize_rcu() in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) * enable_swap_info() or swapoff(). So if "si->flags & SWP_VALID" is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) * true, the si->map, si->cluster_info, etc. must be valid in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291) * critical section.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) * Notice that swapoff or swapoff+swapon can still happen before the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294) * rcu_read_lock() in get_swap_device() or after the rcu_read_unlock()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295) * in put_swap_device() if there isn't any other way to prevent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296) * swapoff, such as page lock, page table lock, etc. The caller must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297) * be prepared for that. For example, the following situation is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) * possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300) * CPU1 CPU2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301) * do_swap_page()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302) * ... swapoff+swapon
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) * __read_swap_cache_async()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304) * swapcache_prepare()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305) * __swap_duplicate()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306) * // check swap_map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) * // verify PTE not changed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) * In __swap_duplicate(), the swap_map need to be checked before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310) * changing partly because the specified swap entry may be for another
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) * swap device which has been swapoff. And in do_swap_page(), after
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) * the page is read from the swap device, the PTE is verified not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313) * changed with the page table locked to check whether the swap device
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314) * has been swapoff or swapoff+swapon.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316) struct swap_info_struct *get_swap_device(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1321) if (!entry.val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1322) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323) si = swp_swap_info(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324) if (!si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325) goto bad_nofile;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328) if (data_race(!(si->flags & SWP_VALID)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) goto unlock_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) if (offset >= si->max)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332) goto unlock_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) return si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335) bad_nofile:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336) pr_err("%s: %s%08lx\n", __func__, Bad_file, entry.val);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) unlock_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344) static unsigned char __swap_entry_free(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348) unsigned long offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) unsigned char usage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) ci = lock_cluster_or_swap_info(p, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1352) usage = __swap_entry_free_locked(p, offset, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1353) unlock_cluster_or_swap_info(p, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1354) if (!usage)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355) free_swap_slot(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) return usage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360) static void swap_entry_free(struct swap_info_struct *p, swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) unsigned long offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366) ci = lock_cluster(p, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) count = p->swap_map[offset];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368) VM_BUG_ON(count != SWAP_HAS_CACHE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369) p->swap_map[offset] = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370) dec_cluster_info_page(p, p->cluster_info, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373) mem_cgroup_uncharge_swap(entry, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) swap_range_free(p, offset, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378) * Caller has made sure that the swap device corresponding to entry
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379) * is still around or has not been recycled.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381) void swap_free(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) if (p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) __swap_entry_free(p, entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) * Called after dropping swapcache to decrease refcnt to swap entries.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) void put_swap_page(struct page *page, swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395) unsigned long offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) unsigned long idx = offset / SWAPFILE_CLUSTER;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) unsigned char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1400) unsigned int i, free_entries = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1401) unsigned char val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1402) int size = swap_entry_size(thp_nr_pages(page));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1403)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1404) si = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1405) if (!si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1406) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1408) ci = lock_cluster_or_swap_info(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1409) if (size == SWAPFILE_CLUSTER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1410) VM_BUG_ON(!cluster_is_huge(ci));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1411) map = si->swap_map + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1412) for (i = 0; i < SWAPFILE_CLUSTER; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1413) val = map[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1414) VM_BUG_ON(!(val & SWAP_HAS_CACHE));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1415) if (val == SWAP_HAS_CACHE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1416) free_entries++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1417) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1418) cluster_clear_huge(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1419) if (free_entries == SWAPFILE_CLUSTER) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1420) unlock_cluster_or_swap_info(si, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1421) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1422) mem_cgroup_uncharge_swap(entry, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1423) swap_free_cluster(si, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1424) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1425) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1426) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1427) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1428) for (i = 0; i < size; i++, entry.val++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1429) if (!__swap_entry_free_locked(si, offset + i, SWAP_HAS_CACHE)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1430) unlock_cluster_or_swap_info(si, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1431) free_swap_slot(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1432) if (i == size - 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1433) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1434) lock_cluster_or_swap_info(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1435) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1436) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1437) unlock_cluster_or_swap_info(si, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1438) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1439)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1440) #ifdef CONFIG_THP_SWAP
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1441) int split_swap_cluster(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1442) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1443) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1444) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1445) unsigned long offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1446)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1447) si = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1448) if (!si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1449) return -EBUSY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1450) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1451) cluster_clear_huge(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1452) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1453) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1454) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1455) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1456)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1457) static int swp_entry_cmp(const void *ent1, const void *ent2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1458) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1459) const swp_entry_t *e1 = ent1, *e2 = ent2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1460)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1461) return (int)swp_type(*e1) - (int)swp_type(*e2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1462) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1463)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1464) void swapcache_free_entries(swp_entry_t *entries, int n)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1465) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1466) struct swap_info_struct *p, *prev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1467) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1468)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1469) if (n <= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1470) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1471)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1472) prev = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1473) p = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1474)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1475) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1476) * Sort swap entries by swap device, so each lock is only taken once.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1477) * nr_swapfiles isn't absolutely correct, but the overhead of sort() is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1478) * so low that it isn't necessary to optimize further.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1479) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1480) if (nr_swapfiles > 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1481) sort(entries, n, sizeof(entries[0]), swp_entry_cmp, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1482) for (i = 0; i < n; ++i) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1483) p = swap_info_get_cont(entries[i], prev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1484) if (p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1485) swap_entry_free(p, entries[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1486) prev = p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1487) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1488) if (p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1489) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1490) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1491) EXPORT_SYMBOL_GPL(swapcache_free_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1492)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1493) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1494) * How many references to page are currently swapped out?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1495) * This does not give an exact answer when swap count is continued,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1496) * but does include the high COUNT_CONTINUED flag to allow for that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1497) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1498) int page_swapcount(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1499) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1500) int count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1501) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1502) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1503) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1504) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1505)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1506) entry.val = page_private(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1507) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1508) if (p) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1509) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1510) ci = lock_cluster_or_swap_info(p, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1511) count = swap_count(p->swap_map[offset]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1512) unlock_cluster_or_swap_info(p, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1513) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1514) return count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1515) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1516)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1517) int __swap_count(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1518) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1519) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1520) pgoff_t offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1521) int count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1522)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1523) si = get_swap_device(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1524) if (si) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1525) count = swap_count(si->swap_map[offset]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1526) put_swap_device(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1527) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1528) return count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1529) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1530)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1531) static int swap_swapcount(struct swap_info_struct *si, swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1532) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1533) int count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1534) pgoff_t offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1535) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1537) ci = lock_cluster_or_swap_info(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1538) count = swap_count(si->swap_map[offset]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1539) unlock_cluster_or_swap_info(si, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1540) return count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1541) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1542)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1543) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1544) * How many references to @entry are currently swapped out?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1545) * This does not give an exact answer when swap count is continued,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1546) * but does include the high COUNT_CONTINUED flag to allow for that.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1547) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1548) int __swp_swapcount(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1549) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1550) int count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1551) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1553) si = get_swap_device(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1554) if (si) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1555) count = swap_swapcount(si, entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1556) put_swap_device(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1557) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1558) return count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1559) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1560)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1561) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1562) * How many references to @entry are currently swapped out?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1563) * This considers COUNT_CONTINUED so it returns exact answer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1564) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1565) int swp_swapcount(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1566) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1567) int count, tmp_count, n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1568) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1569) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1570) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1571) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1572) unsigned char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1573)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1574) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1575) if (!p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1576) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1577)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1578) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1580) ci = lock_cluster_or_swap_info(p, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1581)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1582) count = swap_count(p->swap_map[offset]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1583) if (!(count & COUNT_CONTINUED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1584) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1586) count &= ~COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1587) n = SWAP_MAP_MAX + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1588)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1589) page = vmalloc_to_page(p->swap_map + offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1590) offset &= ~PAGE_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1591) VM_BUG_ON(page_private(page) != SWP_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1592)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1593) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1594) page = list_next_entry(page, lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1595) map = kmap_atomic(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1596) tmp_count = map[offset];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1597) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1598)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1599) count += (tmp_count & ~COUNT_CONTINUED) * n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1600) n *= (SWAP_CONT_MAX + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1601) } while (tmp_count & COUNT_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1602) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1603) unlock_cluster_or_swap_info(p, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1604) return count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1605) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1607) static bool swap_page_trans_huge_swapped(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1608) swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1609) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1610) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1611) unsigned char *map = si->swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1612) unsigned long roffset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1613) unsigned long offset = round_down(roffset, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1614) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1615) bool ret = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1616)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1617) ci = lock_cluster_or_swap_info(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1618) if (!ci || !cluster_is_huge(ci)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1619) if (swap_count(map[roffset]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1620) ret = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1621) goto unlock_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1622) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1623) for (i = 0; i < SWAPFILE_CLUSTER; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1624) if (swap_count(map[offset + i])) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1625) ret = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1626) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1627) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1628) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1629) unlock_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1630) unlock_cluster_or_swap_info(si, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1631) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1632) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1633)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1634) static bool page_swapped(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1635) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1636) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1637) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1638)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1639) if (!IS_ENABLED(CONFIG_THP_SWAP) || likely(!PageTransCompound(page)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1640) return page_swapcount(page) != 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1641)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1642) page = compound_head(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1643) entry.val = page_private(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1644) si = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1645) if (si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1646) return swap_page_trans_huge_swapped(si, entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1647) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1648) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1649)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1650) static int page_trans_huge_map_swapcount(struct page *page, int *total_mapcount,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1651) int *total_swapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1652) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1653) int i, map_swapcount, _total_mapcount, _total_swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1654) unsigned long offset = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1655) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1656) struct swap_cluster_info *ci = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1657) unsigned char *map = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1658) int mapcount, swapcount = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1659)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1660) /* hugetlbfs shouldn't call it */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1661) VM_BUG_ON_PAGE(PageHuge(page), page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1662)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1663) if (!IS_ENABLED(CONFIG_THP_SWAP) || likely(!PageTransCompound(page))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1664) mapcount = page_trans_huge_mapcount(page, total_mapcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1665) if (PageSwapCache(page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1666) swapcount = page_swapcount(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1667) if (total_swapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1668) *total_swapcount = swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1669) return mapcount + swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1670) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1671)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1672) page = compound_head(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1673)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1674) _total_mapcount = _total_swapcount = map_swapcount = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1675) if (PageSwapCache(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1676) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1677)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1678) entry.val = page_private(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1679) si = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1680) if (si) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1681) map = si->swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1682) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1683) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1684) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1685) if (map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1686) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1687) for (i = 0; i < HPAGE_PMD_NR; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1688) mapcount = atomic_read(&page[i]._mapcount) + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1689) _total_mapcount += mapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1690) if (map) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1691) swapcount = swap_count(map[offset + i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1692) _total_swapcount += swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1693) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1694) map_swapcount = max(map_swapcount, mapcount + swapcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1695) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1696) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1697) if (PageDoubleMap(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1698) map_swapcount -= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1699) _total_mapcount -= HPAGE_PMD_NR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1700) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1701) mapcount = compound_mapcount(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1702) map_swapcount += mapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1703) _total_mapcount += mapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1704) if (total_mapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1705) *total_mapcount = _total_mapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1706) if (total_swapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1707) *total_swapcount = _total_swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1708)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1709) return map_swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1710) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1711)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1712) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1713) * We can write to an anon page without COW if there are no other references
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1714) * to it. And as a side-effect, free up its swap: because the old content
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1715) * on disk will never be read, and seeking back there to write new content
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1716) * later would only waste time away from clustering.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1717) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1718) * NOTE: total_map_swapcount should not be relied upon by the caller if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1719) * reuse_swap_page() returns false, but it may be always overwritten
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1720) * (see the other implementation for CONFIG_SWAP=n).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1721) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1722) bool reuse_swap_page(struct page *page, int *total_map_swapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1723) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1724) int count, total_mapcount, total_swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1725)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1726) VM_BUG_ON_PAGE(!PageLocked(page), page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1727) if (unlikely(PageKsm(page)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1728) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1729) count = page_trans_huge_map_swapcount(page, &total_mapcount,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1730) &total_swapcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1731) if (total_map_swapcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1732) *total_map_swapcount = total_mapcount + total_swapcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1733) if (count == 1 && PageSwapCache(page) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1734) (likely(!PageTransCompound(page)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1735) /* The remaining swap count will be freed soon */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1736) total_swapcount == page_swapcount(page))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1737) if (!PageWriteback(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1738) page = compound_head(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1739) delete_from_swap_cache(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1740) SetPageDirty(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1741) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1742) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1743) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1744)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1745) entry.val = page_private(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1746) p = swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1747) if (p->flags & SWP_STABLE_WRITES) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1748) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1749) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1750) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1751) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1752) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1753) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1754)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1755) return count <= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1756) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1757)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1758) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1759) * If swap is getting full, or if there are no more mappings of this page,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1760) * then try_to_free_swap is called to free its swap space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1761) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1762) int try_to_free_swap(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1763) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1764) VM_BUG_ON_PAGE(!PageLocked(page), page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1765)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1766) if (!PageSwapCache(page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1767) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1768) if (PageWriteback(page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1769) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1770) if (page_swapped(page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1771) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1772)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1773) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1774) * Once hibernation has begun to create its image of memory,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1775) * there's a danger that one of the calls to try_to_free_swap()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1776) * - most probably a call from __try_to_reclaim_swap() while
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1777) * hibernation is allocating its own swap pages for the image,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1778) * but conceivably even a call from memory reclaim - will free
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1779) * the swap from a page which has already been recorded in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1780) * image as a clean swapcache page, and then reuse its swap for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1781) * another page of the image. On waking from hibernation, the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1782) * original page might be freed under memory pressure, then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1783) * later read back in from swap, now with the wrong data.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1784) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1785) * Hibernation suspends storage while it is writing the image
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1786) * to disk so check that here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1787) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1788) if (pm_suspended_storage())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1789) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1791) page = compound_head(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1792) delete_from_swap_cache(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1793) SetPageDirty(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1794) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1795) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1796)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1797) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1798) * Free the swap entry like above, but also try to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1799) * free the page cache entry if it is the last user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1800) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1801) int free_swap_and_cache(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1802) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1803) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1804) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1805)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1806) if (non_swap_entry(entry))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1807) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1808)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1809) p = _swap_info_get(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1810) if (p) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1811) count = __swap_entry_free(p, entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1812) if (count == SWAP_HAS_CACHE &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1813) !swap_page_trans_huge_swapped(p, entry))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1814) __try_to_reclaim_swap(p, swp_offset(entry),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1815) TTRS_UNMAPPED | TTRS_FULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1816) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1817) return p != NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1818) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1819)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1820) #ifdef CONFIG_HIBERNATION
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1821) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1822) * Find the swap type that corresponds to given device (if any).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1823) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1824) * @offset - number of the PAGE_SIZE-sized block of the device, starting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1825) * from 0, in which the swap header is expected to be located.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1826) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1827) * This is needed for the suspend to disk (aka swsusp).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1828) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1829) int swap_type_of(dev_t device, sector_t offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1830) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1831) int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1832)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1833) if (!device)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1834) return -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1835)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1836) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1837) for (type = 0; type < nr_swapfiles; type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1838) struct swap_info_struct *sis = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1839)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1840) if (!(sis->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1841) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1842)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1843) if (device == sis->bdev->bd_dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1844) struct swap_extent *se = first_se(sis);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1845)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1846) if (se->start_block == offset) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1847) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1848) return type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1849) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1850) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1851) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1852) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1853) return -ENODEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1854) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1855)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1856) int find_first_swap(dev_t *device)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1857) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1858) int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1859)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1860) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1861) for (type = 0; type < nr_swapfiles; type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1862) struct swap_info_struct *sis = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1863)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1864) if (!(sis->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1865) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1866) *device = sis->bdev->bd_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1867) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1868) return type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1869) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1870) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1871) return -ENODEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1872) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1873)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1874) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1875) * Get the (PAGE_SIZE) block corresponding to given offset on the swapdev
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1876) * corresponding to given index in swap_info (swap type).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1877) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1878) sector_t swapdev_block(int type, pgoff_t offset)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1879) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1880) struct block_device *bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1881) struct swap_info_struct *si = swap_type_to_swap_info(type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1882)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1883) if (!si || !(si->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1884) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1885) return map_swap_entry(swp_entry(type, offset), &bdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1886) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1887)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1888) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1889) * Return either the total number of swap pages of given type, or the number
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1890) * of free pages of that type (depending on @free)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1891) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1892) * This is needed for software suspend
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1893) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1894) unsigned int count_swap_pages(int type, int free)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1895) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1896) unsigned int n = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1897)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1898) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1899) if ((unsigned int)type < nr_swapfiles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1900) struct swap_info_struct *sis = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1901)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1902) spin_lock(&sis->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1903) if (sis->flags & SWP_WRITEOK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1904) n = sis->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1905) if (free)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1906) n -= sis->inuse_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1907) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1908) spin_unlock(&sis->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1909) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1910) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1911) return n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1912) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1913) #endif /* CONFIG_HIBERNATION */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1914)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1915) static inline int pte_same_as_swp(pte_t pte, pte_t swp_pte)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1916) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1917) return pte_same(pte_swp_clear_flags(pte), swp_pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1918) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1919)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1920) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1921) * No need to decide whether this PTE shares the swap entry with others,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1922) * just let do_wp_page work it out if a write is requested later - to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1923) * force COW, vm_page_prot omits write permission from any private vma.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1924) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1925) static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1926) unsigned long addr, swp_entry_t entry, struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1927) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1928) struct page *swapcache;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1929) spinlock_t *ptl;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1930) pte_t *pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1931) int ret = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1932)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1933) swapcache = page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1934) page = ksm_might_need_to_copy(page, vma, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1935) if (unlikely(!page))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1936) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1937)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1938) pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1939) if (unlikely(!pte_same_as_swp(*pte, swp_entry_to_pte(entry)))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1940) ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1941) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1942) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1943)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1944) dec_mm_counter(vma->vm_mm, MM_SWAPENTS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1945) inc_mm_counter(vma->vm_mm, MM_ANONPAGES);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1946) get_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1947) set_pte_at(vma->vm_mm, addr, pte,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1948) pte_mkold(mk_pte(page, vma->vm_page_prot)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1949) if (page == swapcache) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1950) page_add_anon_rmap(page, vma, addr, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1951) } else { /* ksm created a completely new copy */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1952) page_add_new_anon_rmap(page, vma, addr, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1953) lru_cache_add_inactive_or_unevictable(page, vma);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1954) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1955) swap_free(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1956) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1957) pte_unmap_unlock(pte, ptl);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1958) if (page != swapcache) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1959) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1960) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1961) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1962) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1963) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1964)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1965) static int unuse_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1966) unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1967) unsigned int type, bool frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1968) unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1969) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1970) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1971) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1972) pte_t *pte;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1973) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1974) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1975) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1976) volatile unsigned char *swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1977)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1978) si = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1979) pte = pte_offset_map(pmd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1980) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1981) if (!is_swap_pte(*pte))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1982) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1983)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1984) entry = pte_to_swp_entry(*pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1985) if (swp_type(entry) != type)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1986) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1987)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1988) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1989) if (frontswap && !frontswap_test(si, offset))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1990) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1991)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1992) pte_unmap(pte);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1993) swap_map = &si->swap_map[offset];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1994) page = lookup_swap_cache(entry, vma, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1995) if (!page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1996) struct vm_fault vmf = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1997) .vma = vma,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1998) .address = addr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1999) .pmd = pmd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2000) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2001)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2002) page = swapin_readahead(entry, GFP_HIGHUSER_MOVABLE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2003) &vmf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2004) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2005) if (!page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2006) if (*swap_map == 0 || *swap_map == SWAP_MAP_BAD)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2007) goto try_next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2008) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2009) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2010)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2011) lock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2012) wait_on_page_writeback(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2013) ret = unuse_pte(vma, pmd, addr, entry, page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2014) if (ret < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2015) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2016) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2017) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2018) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2019)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2020) try_to_free_swap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2021) trace_android_vh_unuse_swap_page(si, page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2022) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2023) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2024)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2025) if (*fs_pages_to_unuse && !--(*fs_pages_to_unuse)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2026) ret = FRONTSWAP_PAGES_UNUSED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2027) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2028) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2029) try_next:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2030) pte = pte_offset_map(pmd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2031) } while (pte++, addr += PAGE_SIZE, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2032) pte_unmap(pte - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2033)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2034) ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2035) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2036) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2037) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2038)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2039) static inline int unuse_pmd_range(struct vm_area_struct *vma, pud_t *pud,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2040) unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2041) unsigned int type, bool frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2042) unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2043) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2044) pmd_t *pmd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2045) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2046) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2047)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2048) pmd = pmd_offset(pud, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2049) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2050) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2051) next = pmd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2052) if (pmd_none_or_trans_huge_or_clear_bad(pmd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2053) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2054) ret = unuse_pte_range(vma, pmd, addr, next, type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2055) frontswap, fs_pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2056) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2057) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2058) } while (pmd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2059) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2060) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2061)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2062) static inline int unuse_pud_range(struct vm_area_struct *vma, p4d_t *p4d,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2063) unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2064) unsigned int type, bool frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2065) unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2066) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2067) pud_t *pud;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2068) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2069) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2070)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2071) pud = pud_offset(p4d, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2072) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2073) next = pud_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2074) if (pud_none_or_clear_bad(pud))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2075) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2076) ret = unuse_pmd_range(vma, pud, addr, next, type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2077) frontswap, fs_pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2078) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2079) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2080) } while (pud++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2081) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2082) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2084) static inline int unuse_p4d_range(struct vm_area_struct *vma, pgd_t *pgd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2085) unsigned long addr, unsigned long end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2086) unsigned int type, bool frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2087) unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2088) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2089) p4d_t *p4d;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2090) unsigned long next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2091) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2092)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2093) p4d = p4d_offset(pgd, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2094) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2095) next = p4d_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2096) if (p4d_none_or_clear_bad(p4d))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2097) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2098) ret = unuse_pud_range(vma, p4d, addr, next, type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2099) frontswap, fs_pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2100) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2101) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2102) } while (p4d++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2103) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2104) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2105)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2106) static int unuse_vma(struct vm_area_struct *vma, unsigned int type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2107) bool frontswap, unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2108) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2109) pgd_t *pgd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2110) unsigned long addr, end, next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2111) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2113) addr = vma->vm_start;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2114) end = vma->vm_end;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2115)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2116) pgd = pgd_offset(vma->vm_mm, addr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2117) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2118) next = pgd_addr_end(addr, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2119) if (pgd_none_or_clear_bad(pgd))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2120) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2121) ret = unuse_p4d_range(vma, pgd, addr, next, type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2122) frontswap, fs_pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2123) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2124) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2125) } while (pgd++, addr = next, addr != end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2126) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2127) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2129) static int unuse_mm(struct mm_struct *mm, unsigned int type,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2130) bool frontswap, unsigned long *fs_pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2131) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2132) struct vm_area_struct *vma;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2133) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2135) mmap_read_lock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2136) for (vma = mm->mmap; vma; vma = vma->vm_next) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2137) if (vma->anon_vma) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2138) ret = unuse_vma(vma, type, frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2139) fs_pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2140) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2141) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2142) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2143) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2144) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2145) mmap_read_unlock(mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2146) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2147) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2149) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2150) * Scan swap_map (or frontswap_map if frontswap parameter is true)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2151) * from current position to next entry still in use. Return 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2152) * if there are no inuse entries after prev till end of the map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2153) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2154) static unsigned int find_next_to_unuse(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2155) unsigned int prev, bool frontswap)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2156) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2157) unsigned int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2158) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2159)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2160) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2161) * No need for swap_lock here: we're just looking
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2162) * for whether an entry is in use, not modifying it; false
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2163) * hits are okay, and sys_swapoff() has already prevented new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2164) * allocations from this area (while holding swap_lock).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2165) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2166) for (i = prev + 1; i < si->max; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2167) count = READ_ONCE(si->swap_map[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2168) if (count && swap_count(count) != SWAP_MAP_BAD)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2169) if (!frontswap || frontswap_test(si, i))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2170) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2171) if ((i % LATENCY_LIMIT) == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2172) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2173) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2175) if (i == si->max)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2176) i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2178) return i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2179) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2180)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2181) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2182) * If the boolean frontswap is true, only unuse pages_to_unuse pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2183) * pages_to_unuse==0 means all pages; ignored if frontswap is false
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2184) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2185) int try_to_unuse(unsigned int type, bool frontswap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2186) unsigned long pages_to_unuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2187) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2188) struct mm_struct *prev_mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2189) struct mm_struct *mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2190) struct list_head *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2191) int retval = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2192) struct swap_info_struct *si = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2193) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2194) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2195) unsigned int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2197) if (!READ_ONCE(si->inuse_pages))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2198) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2200) if (!frontswap)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2201) pages_to_unuse = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2202)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2203) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2204) retval = shmem_unuse(type, frontswap, &pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2205) if (retval)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2206) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2208) prev_mm = &init_mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2209) mmget(prev_mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2211) spin_lock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2212) p = &init_mm.mmlist;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2213) while (READ_ONCE(si->inuse_pages) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2214) !signal_pending(current) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2215) (p = p->next) != &init_mm.mmlist) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2216)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2217) mm = list_entry(p, struct mm_struct, mmlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2218) if (!mmget_not_zero(mm))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2219) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2220) spin_unlock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2221) mmput(prev_mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2222) prev_mm = mm;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2223) retval = unuse_mm(mm, type, frontswap, &pages_to_unuse);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2224)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2225) if (retval) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2226) mmput(prev_mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2227) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2228) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2229)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2230) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2231) * Make sure that we aren't completely killing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2232) * interactive performance.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2233) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2234) cond_resched();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2235) spin_lock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2236) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2237) spin_unlock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2238)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2239) mmput(prev_mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2240)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2241) i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2242) while (READ_ONCE(si->inuse_pages) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2243) !signal_pending(current) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2244) (i = find_next_to_unuse(si, i, frontswap)) != 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2245)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2246) entry = swp_entry(type, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2247) page = find_get_page(swap_address_space(entry), i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2248) if (!page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2249) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2251) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2252) * It is conceivable that a racing task removed this page from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2253) * swap cache just before we acquired the page lock. The page
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2254) * might even be back in swap cache on another swap area. But
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2255) * that is okay, try_to_free_swap() only removes stale pages.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2256) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2257) lock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2258) wait_on_page_writeback(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2259) try_to_free_swap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2260) trace_android_vh_unuse_swap_page(si, page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2261) unlock_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2262) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2264) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2265) * For frontswap, we just need to unuse pages_to_unuse, if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2266) * it was specified. Need not check frontswap again here as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2267) * we already zeroed out pages_to_unuse if not frontswap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2268) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2269) if (pages_to_unuse && --pages_to_unuse == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2270) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2271) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2272)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2273) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2274) * Lets check again to see if there are still swap entries in the map.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2275) * If yes, we would need to do retry the unuse logic again.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2276) * Under global memory pressure, swap entries can be reinserted back
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2277) * into process space after the mmlist loop above passes over them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2278) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2279) * Limit the number of retries? No: when mmget_not_zero() above fails,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2280) * that mm is likely to be freeing swap from exit_mmap(), which proceeds
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2281) * at its own independent pace; and even shmem_writepage() could have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2282) * been preempted after get_swap_page(), temporarily hiding that swap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2283) * It's easy and robust (though cpu-intensive) just to keep retrying.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2284) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2285) if (READ_ONCE(si->inuse_pages)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2286) if (!signal_pending(current))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2287) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2288) retval = -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2289) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2290) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2291) return (retval == FRONTSWAP_PAGES_UNUSED) ? 0 : retval;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2292) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2294) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2295) * After a successful try_to_unuse, if no swap is now in use, we know
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2296) * we can empty the mmlist. swap_lock must be held on entry and exit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2297) * Note that mmlist_lock nests inside swap_lock, and an mm must be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2298) * added to the mmlist just after page_duplicate - before would be racy.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2299) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2300) static void drain_mmlist(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2301) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2302) struct list_head *p, *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2303) unsigned int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2304)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2305) for (type = 0; type < nr_swapfiles; type++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2306) if (swap_info[type]->inuse_pages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2307) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2308) spin_lock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2309) list_for_each_safe(p, next, &init_mm.mmlist)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2310) list_del_init(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2311) spin_unlock(&mmlist_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2312) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2313)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2314) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2315) * Use this swapdev's extent info to locate the (PAGE_SIZE) block which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2316) * corresponds to page offset for the specified swap entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2317) * Note that the type of this function is sector_t, but it returns page offset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2318) * into the bdev, not sector offset.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2319) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2320) static sector_t map_swap_entry(swp_entry_t entry, struct block_device **bdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2321) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2322) struct swap_info_struct *sis;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2323) struct swap_extent *se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2324) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2326) sis = swp_swap_info(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2327) *bdev = sis->bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2328)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2329) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2330) se = offset_to_swap_extent(sis, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2331) return se->start_block + (offset - se->start_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2334) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2335) * Returns the page offset into bdev for the specified page's swap entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2336) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2337) sector_t map_swap_page(struct page *page, struct block_device **bdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2338) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2339) swp_entry_t entry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2340) entry.val = page_private(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2341) return map_swap_entry(entry, bdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2342) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2344) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2345) * Free all of a swapdev's extent information
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2346) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2347) static void destroy_swap_extents(struct swap_info_struct *sis)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2348) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2349) while (!RB_EMPTY_ROOT(&sis->swap_extent_root)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2350) struct rb_node *rb = sis->swap_extent_root.rb_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2351) struct swap_extent *se = rb_entry(rb, struct swap_extent, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2352)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2353) rb_erase(rb, &sis->swap_extent_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2354) kfree(se);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2355) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2356)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2357) if (sis->flags & SWP_ACTIVATED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2358) struct file *swap_file = sis->swap_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2359) struct address_space *mapping = swap_file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2360)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2361) sis->flags &= ~SWP_ACTIVATED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2362) if (mapping->a_ops->swap_deactivate)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2363) mapping->a_ops->swap_deactivate(swap_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2364) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2365) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2367) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2368) * Add a block range (and the corresponding page range) into this swapdev's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2369) * extent tree.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2370) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2371) * This function rather assumes that it is called in ascending page order.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2372) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2373) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2374) add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2375) unsigned long nr_pages, sector_t start_block)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2376) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2377) struct rb_node **link = &sis->swap_extent_root.rb_node, *parent = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2378) struct swap_extent *se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2379) struct swap_extent *new_se;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2380)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2381) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2382) * place the new node at the right most since the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2383) * function is called in ascending page order.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2384) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2385) while (*link) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2386) parent = *link;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2387) link = &parent->rb_right;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2388) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2389)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2390) if (parent) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2391) se = rb_entry(parent, struct swap_extent, rb_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2392) BUG_ON(se->start_page + se->nr_pages != start_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2393) if (se->start_block + se->nr_pages == start_block) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2394) /* Merge it */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2395) se->nr_pages += nr_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2396) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2397) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2398) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2399)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2400) /* No merge, insert a new extent. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2401) new_se = kmalloc(sizeof(*se), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2402) if (new_se == NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2403) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2404) new_se->start_page = start_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2405) new_se->nr_pages = nr_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2406) new_se->start_block = start_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2408) rb_link_node(&new_se->rb_node, parent, link);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2409) rb_insert_color(&new_se->rb_node, &sis->swap_extent_root);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2410) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2411) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2412) EXPORT_SYMBOL_GPL(add_swap_extent);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2413)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2414) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2415) * A `swap extent' is a simple thing which maps a contiguous range of pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2416) * onto a contiguous range of disk blocks. An ordered list of swap extents
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2417) * is built at swapon time and is then used at swap_writepage/swap_readpage
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2418) * time for locating where on disk a page belongs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2419) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2420) * If the swapfile is an S_ISBLK block device, a single extent is installed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2421) * This is done so that the main operating code can treat S_ISBLK and S_ISREG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2422) * swap files identically.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2423) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2424) * Whether the swapdev is an S_ISREG file or an S_ISBLK blockdev, the swap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2425) * extent list operates in PAGE_SIZE disk blocks. Both S_ISREG and S_ISBLK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2426) * swapfiles are handled *identically* after swapon time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2427) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2428) * For S_ISREG swapfiles, setup_swap_extents() will walk all the file's blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2429) * and will parse them into an ordered extent list, in PAGE_SIZE chunks. If
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2430) * some stray blocks are found which do not fall within the PAGE_SIZE alignment
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2431) * requirements, they are simply tossed out - we will never use those blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2432) * for swapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2433) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2434) * For all swap devices we set S_SWAPFILE across the life of the swapon. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2435) * prevents users from writing to the swap device, which will corrupt memory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2436) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2437) * The amount of disk space which a single swap extent represents varies.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2438) * Typically it is in the 1-4 megabyte range. So we can have hundreds of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2439) * extents in the list. To avoid much list walking, we cache the previous
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2440) * search location in `curr_swap_extent', and start new searches from there.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2441) * This is extremely effective. The average number of iterations in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2442) * map_swap_page() has been measured at about 0.3 per page. - akpm.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2443) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2444) static int setup_swap_extents(struct swap_info_struct *sis, sector_t *span)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2445) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2446) struct file *swap_file = sis->swap_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2447) struct address_space *mapping = swap_file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2448) struct inode *inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2449) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2451) if (S_ISBLK(inode->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2452) ret = add_swap_extent(sis, 0, sis->max, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2453) *span = sis->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2454) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2455) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2456)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2457) if (mapping->a_ops->swap_activate) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2458) ret = mapping->a_ops->swap_activate(sis, swap_file, span);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2459) if (ret >= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2460) sis->flags |= SWP_ACTIVATED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2461) if (!ret) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2462) sis->flags |= SWP_FS_OPS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2463) ret = add_swap_extent(sis, 0, sis->max, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2464) *span = sis->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2465) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2466) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2467) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2468)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2469) return generic_swapfile_activate(sis, swap_file, span);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2470) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2471)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2472) static int swap_node(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2473) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2474) struct block_device *bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2475)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2476) if (p->bdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2477) bdev = p->bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2478) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2479) bdev = p->swap_file->f_inode->i_sb->s_bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2480)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2481) return bdev ? bdev->bd_disk->node_id : NUMA_NO_NODE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2482) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2483)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2484) static void setup_swap_info(struct swap_info_struct *p, int prio,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2485) unsigned char *swap_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2486) struct swap_cluster_info *cluster_info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2487) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2488) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2489)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2490) if (prio >= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2491) p->prio = prio;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2492) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2493) p->prio = --least_priority;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2494) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2495) * the plist prio is negated because plist ordering is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2496) * low-to-high, while swap ordering is high-to-low
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2497) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2498) p->list.prio = -p->prio;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2499) for_each_node(i) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2500) if (p->prio >= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2501) p->avail_lists[i].prio = -p->prio;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2502) else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2503) if (swap_node(p) == i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2504) p->avail_lists[i].prio = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2505) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2506) p->avail_lists[i].prio = -p->prio;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2507) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2508) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2509) p->swap_map = swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2510) p->cluster_info = cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2511) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2512)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2513) static void _enable_swap_info(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2514) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2515) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2516)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2517) p->flags |= SWP_WRITEOK | SWP_VALID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2518) trace_android_vh_account_swap_pages(p, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2519) if (!skip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2520) atomic_long_add(p->pages, &nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2521) total_swap_pages += p->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2522) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2523) assert_spin_locked(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2524) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2525) * both lists are plists, and thus priority ordered.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2526) * swap_active_head needs to be priority ordered for swapoff(),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2527) * which on removal of any swap_info_struct with an auto-assigned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2528) * (i.e. negative) priority increments the auto-assigned priority
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2529) * of any lower-priority swap_info_structs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2530) * swap_avail_head needs to be priority ordered for get_swap_page(),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2531) * which allocates swap pages from the highest available priority
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2532) * swap_info_struct.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2533) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2534) plist_add(&p->list, &swap_active_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2535) add_to_avail_list(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2536) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2537)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2538) static void enable_swap_info(struct swap_info_struct *p, int prio,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2539) unsigned char *swap_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2540) struct swap_cluster_info *cluster_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2541) unsigned long *frontswap_map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2542) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2543) frontswap_init(p->type, frontswap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2544) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2545) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2546) setup_swap_info(p, prio, swap_map, cluster_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2547) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2548) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2549) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2550) * Guarantee swap_map, cluster_info, etc. fields are valid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2551) * between get/put_swap_device() if SWP_VALID bit is set
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2552) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2553) synchronize_rcu();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2554) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2555) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2556) _enable_swap_info(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2557) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2558) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2559) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2560)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2561) static void reinsert_swap_info(struct swap_info_struct *p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2562) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2563) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2564) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2565) setup_swap_info(p, p->prio, p->swap_map, p->cluster_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2566) _enable_swap_info(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2567) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2568) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2569) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2570)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2571) bool has_usable_swap(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2572) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2573) bool ret = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2574)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2575) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2576) if (plist_head_empty(&swap_active_head))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2577) ret = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2578) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2579) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2580) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2581)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2582) SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2583) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2584) struct swap_info_struct *p = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2585) unsigned char *swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2586) struct swap_cluster_info *cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2587) unsigned long *frontswap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2588) struct file *swap_file, *victim;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2589) struct address_space *mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2590) struct inode *inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2591) struct filename *pathname;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2592) int err, found = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2593) unsigned int old_block_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2594) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2595)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2596) if (!capable(CAP_SYS_ADMIN))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2597) return -EPERM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2598)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2599) BUG_ON(!current->mm);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2600)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2601) pathname = getname(specialfile);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2602) if (IS_ERR(pathname))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2603) return PTR_ERR(pathname);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2605) victim = file_open_name(pathname, O_RDWR|O_LARGEFILE, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2606) err = PTR_ERR(victim);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2607) if (IS_ERR(victim))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2608) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2609)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2610) mapping = victim->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2611) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2612) plist_for_each_entry(p, &swap_active_head, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2613) if (p->flags & SWP_WRITEOK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2614) if (p->swap_file->f_mapping == mapping) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2615) found = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2616) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2617) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2618) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2619) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2620) if (!found) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2621) err = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2622) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2623) goto out_dput;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2624) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2625) if (!security_vm_enough_memory_mm(current->mm, p->pages))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2626) vm_unacct_memory(p->pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2627) else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2628) err = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2629) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2630) goto out_dput;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2631) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2632) del_from_avail_list(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2633) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2634) if (p->prio < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2635) struct swap_info_struct *si = p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2636) int nid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2637)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2638) plist_for_each_entry_continue(si, &swap_active_head, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2639) si->prio++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2640) si->list.prio--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2641) for_each_node(nid) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2642) if (si->avail_lists[nid].prio != 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2643) si->avail_lists[nid].prio--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2644) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2645) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2646) least_priority++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2647) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2648) plist_del(&p->list, &swap_active_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2649) trace_android_vh_account_swap_pages(p, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2650) if (!skip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2651) atomic_long_sub(p->pages, &nr_swap_pages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2652) total_swap_pages -= p->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2653) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2654) p->flags &= ~SWP_WRITEOK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2655) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2656) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2657)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2658) disable_swap_slots_cache_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2659)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2660) set_current_oom_origin();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2661) err = try_to_unuse(p->type, false, 0); /* force unuse all pages */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2662) clear_current_oom_origin();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2663)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2664) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2665) /* re-insert swap space back into swap_list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2666) reinsert_swap_info(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2667) reenable_swap_slots_cache_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2668) goto out_dput;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2669) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2670)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2671) reenable_swap_slots_cache_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2672)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2673) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2674) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2675) p->flags &= ~SWP_VALID; /* mark swap device as invalid */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2676) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2677) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2678) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2679) * wait for swap operations protected by get/put_swap_device()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2680) * to complete
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2681) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2682) synchronize_rcu();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2683)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2684) flush_work(&p->discard_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2685)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2686) destroy_swap_extents(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2687) if (p->flags & SWP_CONTINUED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2688) free_swap_count_continuations(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2689)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2690) if (!p->bdev || !blk_queue_nonrot(bdev_get_queue(p->bdev)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2691) atomic_dec(&nr_rotate_swap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2692)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2693) mutex_lock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2694) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2695) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2696) drain_mmlist();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2697)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2698) /* wait for anyone still in scan_swap_map */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2699) p->highest_bit = 0; /* cuts scans short */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2700) while (p->flags >= SWP_SCANNING) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2701) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2702) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2703) schedule_timeout_uninterruptible(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2704) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2705) spin_lock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2706) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2707)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2708) swap_file = p->swap_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2709) old_block_size = p->old_block_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2710) p->swap_file = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2711) p->max = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2712) swap_map = p->swap_map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2713) p->swap_map = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2714) cluster_info = p->cluster_info;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2715) p->cluster_info = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2716) frontswap_map = frontswap_map_get(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2717) spin_unlock(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2718) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2719) arch_swap_invalidate_area(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2720) frontswap_invalidate_area(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2721) frontswap_map_set(p, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2722) mutex_unlock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2723) free_percpu(p->percpu_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2724) p->percpu_cluster = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2725) free_percpu(p->cluster_next_cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2726) p->cluster_next_cpu = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2727) vfree(swap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2728) kvfree(cluster_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2729) kvfree(frontswap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2730) /* Destroy swap account information */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2731) swap_cgroup_swapoff(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2732) exit_swap_address_space(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2734) inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2735) if (S_ISBLK(inode->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2736) struct block_device *bdev = I_BDEV(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2737)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2738) set_blocksize(bdev, old_block_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2739) blkdev_put(bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2740) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2741)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2742) inode_lock(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2743) inode->i_flags &= ~S_SWAPFILE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2744) inode_unlock(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2745) filp_close(swap_file, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2746)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2747) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2748) * Clear the SWP_USED flag after all resources are freed so that swapon
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2749) * can reuse this swap_info in alloc_swap_info() safely. It is ok to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2750) * not hold p->lock after we cleared its SWP_WRITEOK.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2751) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2752) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2753) p->flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2754) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2755)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2756) err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2757) atomic_inc(&proc_poll_event);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2758) wake_up_interruptible(&proc_poll_wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2759)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2760) out_dput:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2761) filp_close(victim, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2762) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2763) putname(pathname);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2764) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2765) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2766)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2767) #ifdef CONFIG_PROC_FS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2768) static __poll_t swaps_poll(struct file *file, poll_table *wait)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2769) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2770) struct seq_file *seq = file->private_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2771)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2772) poll_wait(file, &proc_poll_wait, wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2773)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2774) if (seq->poll_event != atomic_read(&proc_poll_event)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2775) seq->poll_event = atomic_read(&proc_poll_event);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2776) return EPOLLIN | EPOLLRDNORM | EPOLLERR | EPOLLPRI;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2777) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2778)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2779) return EPOLLIN | EPOLLRDNORM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2780) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2781)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2782) /* iterator */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2783) static void *swap_start(struct seq_file *swap, loff_t *pos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2784) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2785) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2786) int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2787) loff_t l = *pos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2788)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2789) mutex_lock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2791) if (!l)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2792) return SEQ_START_TOKEN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2793)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2794) for (type = 0; (si = swap_type_to_swap_info(type)); type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2795) if (!(si->flags & SWP_USED) || !si->swap_map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2796) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2797) if (!--l)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2798) return si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2799) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2800)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2801) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2802) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2803)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2804) static void *swap_next(struct seq_file *swap, void *v, loff_t *pos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2805) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2806) struct swap_info_struct *si = v;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2807) int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2808)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2809) if (v == SEQ_START_TOKEN)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2810) type = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2811) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2812) type = si->type + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2813)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2814) ++(*pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2815) for (; (si = swap_type_to_swap_info(type)); type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2816) if (!(si->flags & SWP_USED) || !si->swap_map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2817) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2818) return si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2819) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2820)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2821) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2822) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2823)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2824) static void swap_stop(struct seq_file *swap, void *v)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2825) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2826) mutex_unlock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2827) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2828)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2829) static int swap_show(struct seq_file *swap, void *v)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2830) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2831) struct swap_info_struct *si = v;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2832) struct file *file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2833) int len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2834) unsigned int bytes, inuse;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2835)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2836) if (si == SEQ_START_TOKEN) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2837) seq_puts(swap,"Filename\t\t\t\tType\t\tSize\t\tUsed\t\tPriority\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2838) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2839) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2840)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2841) bytes = si->pages << (PAGE_SHIFT - 10);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2842) inuse = si->inuse_pages << (PAGE_SHIFT - 10);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2843)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2844) file = si->swap_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2845) len = seq_file_path(swap, file, " \t\n\\");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2846) seq_printf(swap, "%*s%s\t%u\t%s%u\t%s%d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2847) len < 40 ? 40 - len : 1, " ",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2848) S_ISBLK(file_inode(file)->i_mode) ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2849) "partition" : "file\t",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2850) bytes, bytes < 10000000 ? "\t" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2851) inuse, inuse < 10000000 ? "\t" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2852) si->prio);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2853) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2854) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2855)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2856) static const struct seq_operations swaps_op = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2857) .start = swap_start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2858) .next = swap_next,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2859) .stop = swap_stop,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2860) .show = swap_show
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2861) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2862)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2863) static int swaps_open(struct inode *inode, struct file *file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2864) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2865) struct seq_file *seq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2866) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2867)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2868) ret = seq_open(file, &swaps_op);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2869) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2870) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2871)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2872) seq = file->private_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2873) seq->poll_event = atomic_read(&proc_poll_event);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2874) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2875) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2876)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2877) static const struct proc_ops swaps_proc_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2878) .proc_flags = PROC_ENTRY_PERMANENT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2879) .proc_open = swaps_open,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2880) .proc_read = seq_read,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2881) .proc_lseek = seq_lseek,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2882) .proc_release = seq_release,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2883) .proc_poll = swaps_poll,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2884) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2885)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2886) static int __init procswaps_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2887) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2888) proc_create("swaps", 0, NULL, &swaps_proc_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2889) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2890) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2891) __initcall(procswaps_init);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2892) #endif /* CONFIG_PROC_FS */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2893)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2894) #ifdef MAX_SWAPFILES_CHECK
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2895) static int __init max_swapfiles_check(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2896) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2897) MAX_SWAPFILES_CHECK();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2898) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2899) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2900) late_initcall(max_swapfiles_check);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2901) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2902)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2903) static struct swap_info_struct *alloc_swap_info(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2904) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2905) struct swap_info_struct *p = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2906) struct swap_info_struct *defer = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2907) unsigned int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2908) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2909) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2910)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2911) trace_android_vh_alloc_si(&p, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2912) if (!skip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2913) p = kvzalloc(struct_size(p, avail_lists, nr_node_ids), GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2914) if (!p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2915) return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2916)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2917) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2918) for (type = 0; type < nr_swapfiles; type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2919) if (!(swap_info[type]->flags & SWP_USED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2920) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2921) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2922) if (type >= MAX_SWAPFILES) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2923) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2924) kvfree(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2925) return ERR_PTR(-EPERM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2926) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2927) if (type >= nr_swapfiles) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2928) p->type = type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2929) WRITE_ONCE(swap_info[type], p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2930) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2931) * Write swap_info[type] before nr_swapfiles, in case a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2932) * racing procfs swap_start() or swap_next() is reading them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2933) * (We never shrink nr_swapfiles, we never free this entry.)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2934) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2935) smp_wmb();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2936) WRITE_ONCE(nr_swapfiles, nr_swapfiles + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2937) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2938) defer = p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2939) p = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2940) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2941) * Do not memset this entry: a racing procfs swap_next()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2942) * would be relying on p->type to remain valid.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2943) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2944) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2945) p->swap_extent_root = RB_ROOT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2946) plist_node_init(&p->list, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2947) for_each_node(i)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2948) plist_node_init(&p->avail_lists[i], 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2949) p->flags = SWP_USED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2950) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2951) kvfree(defer);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2952) spin_lock_init(&p->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2953) spin_lock_init(&p->cont_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2954)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2955) return p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2956) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2957)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2958) static int claim_swapfile(struct swap_info_struct *p, struct inode *inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2959) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2960) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2961)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2962) if (S_ISBLK(inode->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2963) p->bdev = blkdev_get_by_dev(inode->i_rdev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2964) FMODE_READ | FMODE_WRITE | FMODE_EXCL, p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2965) if (IS_ERR(p->bdev)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2966) error = PTR_ERR(p->bdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2967) p->bdev = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2968) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2969) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2970) p->old_block_size = block_size(p->bdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2971) error = set_blocksize(p->bdev, PAGE_SIZE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2972) if (error < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2973) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2974) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2975) * Zoned block devices contain zones that have a sequential
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2976) * write only restriction. Hence zoned block devices are not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2977) * suitable for swapping. Disallow them here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2978) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2979) if (blk_queue_is_zoned(p->bdev->bd_disk->queue))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2980) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2981) p->flags |= SWP_BLKDEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2982) } else if (S_ISREG(inode->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2983) p->bdev = inode->i_sb->s_bdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2984) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2985)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2986) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2987) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2988)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2989)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2990) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2991) * Find out how many pages are allowed for a single swap device. There
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2992) * are two limiting factors:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2993) * 1) the number of bits for the swap offset in the swp_entry_t type, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2994) * 2) the number of bits in the swap pte, as defined by the different
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2995) * architectures.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2996) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2997) * In order to find the largest possible bit mask, a swap entry with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2998) * swap type 0 and swap offset ~0UL is created, encoded to a swap pte,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2999) * decoded to a swp_entry_t again, and finally the swap offset is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3000) * extracted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3001) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3002) * This will mask all the bits from the initial ~0UL mask that can't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3003) * be encoded in either the swp_entry_t or the architecture definition
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3004) * of a swap pte.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3005) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3006) unsigned long generic_max_swapfile_size(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3007) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3008) return swp_offset(pte_to_swp_entry(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3009) swp_entry_to_pte(swp_entry(0, ~0UL)))) + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3010) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3011)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3012) /* Can be overridden by an architecture for additional checks. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3013) __weak unsigned long max_swapfile_size(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3014) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3015) return generic_max_swapfile_size();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3016) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3017)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3018) static unsigned long read_swap_header(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3019) union swap_header *swap_header,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3020) struct inode *inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3021) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3022) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3023) unsigned long maxpages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3024) unsigned long swapfilepages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3025) unsigned long last_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3026)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3027) if (memcmp("SWAPSPACE2", swap_header->magic.magic, 10)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3028) pr_err("Unable to find swap-space signature\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3029) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3030) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3031)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3032) /* swap partition endianess hack... */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3033) if (swab32(swap_header->info.version) == 1) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3034) swab32s(&swap_header->info.version);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3035) swab32s(&swap_header->info.last_page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3036) swab32s(&swap_header->info.nr_badpages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3037) if (swap_header->info.nr_badpages > MAX_SWAP_BADPAGES)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3038) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3039) for (i = 0; i < swap_header->info.nr_badpages; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3040) swab32s(&swap_header->info.badpages[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3041) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3042) /* Check the swap header's sub-version */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3043) if (swap_header->info.version != 1) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3044) pr_warn("Unable to handle swap header version %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3045) swap_header->info.version);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3046) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3047) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3048)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3049) p->lowest_bit = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3050) p->cluster_next = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3051) p->cluster_nr = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3052)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3053) maxpages = max_swapfile_size();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3054) last_page = swap_header->info.last_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3055) if (!last_page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3056) pr_warn("Empty swap-file\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3057) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3058) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3059) if (last_page > maxpages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3060) pr_warn("Truncating oversized swap area, only using %luk out of %luk\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3061) maxpages << (PAGE_SHIFT - 10),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3062) last_page << (PAGE_SHIFT - 10));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3063) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3064) if (maxpages > last_page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3065) maxpages = last_page + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3066) /* p->max is an unsigned int: don't overflow it */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3067) if ((unsigned int)maxpages == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3068) maxpages = UINT_MAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3069) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3070) p->highest_bit = maxpages - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3071)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3072) if (!maxpages)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3073) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3074) swapfilepages = i_size_read(inode) >> PAGE_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3075) if (swapfilepages && maxpages > swapfilepages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3076) pr_warn("Swap area shorter than signature indicates\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3077) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3078) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3079) if (swap_header->info.nr_badpages && S_ISREG(inode->i_mode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3080) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3081) if (swap_header->info.nr_badpages > MAX_SWAP_BADPAGES)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3082) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3084) return maxpages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3085) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3086)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3087) #define SWAP_CLUSTER_INFO_COLS \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3088) DIV_ROUND_UP(L1_CACHE_BYTES, sizeof(struct swap_cluster_info))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3089) #define SWAP_CLUSTER_SPACE_COLS \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3090) DIV_ROUND_UP(SWAP_ADDRESS_SPACE_PAGES, SWAPFILE_CLUSTER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3091) #define SWAP_CLUSTER_COLS \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3092) max_t(unsigned int, SWAP_CLUSTER_INFO_COLS, SWAP_CLUSTER_SPACE_COLS)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3093)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3094) static int setup_swap_map_and_extents(struct swap_info_struct *p,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3095) union swap_header *swap_header,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3096) unsigned char *swap_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3097) struct swap_cluster_info *cluster_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3098) unsigned long maxpages,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3099) sector_t *span)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3100) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3101) unsigned int j, k;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3102) unsigned int nr_good_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3103) int nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3104) unsigned long nr_clusters = DIV_ROUND_UP(maxpages, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3105) unsigned long col = p->cluster_next / SWAPFILE_CLUSTER % SWAP_CLUSTER_COLS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3106) unsigned long i, idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3108) nr_good_pages = maxpages - 1; /* omit header page */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3109)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3110) cluster_list_init(&p->free_clusters);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3111) cluster_list_init(&p->discard_clusters);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3113) for (i = 0; i < swap_header->info.nr_badpages; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3114) unsigned int page_nr = swap_header->info.badpages[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3115) if (page_nr == 0 || page_nr > swap_header->info.last_page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3116) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3117) if (page_nr < maxpages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3118) swap_map[page_nr] = SWAP_MAP_BAD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3119) nr_good_pages--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3120) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3121) * Haven't marked the cluster free yet, no list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3122) * operation involved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3123) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3124) inc_cluster_info_page(p, cluster_info, page_nr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3125) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3126) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3127)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3128) /* Haven't marked the cluster free yet, no list operation involved */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3129) for (i = maxpages; i < round_up(maxpages, SWAPFILE_CLUSTER); i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3130) inc_cluster_info_page(p, cluster_info, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3132) if (nr_good_pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3133) swap_map[0] = SWAP_MAP_BAD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3134) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3135) * Not mark the cluster free yet, no list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3136) * operation involved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3137) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3138) inc_cluster_info_page(p, cluster_info, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3139) p->max = maxpages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3140) p->pages = nr_good_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3141) nr_extents = setup_swap_extents(p, span);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3142) if (nr_extents < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3143) return nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3144) nr_good_pages = p->pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3145) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3146) if (!nr_good_pages) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3147) pr_warn("Empty swap-file\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3148) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3149) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3150)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3151) if (!cluster_info)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3152) return nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3155) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3156) * Reduce false cache line sharing between cluster_info and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3157) * sharing same address space.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3158) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3159) for (k = 0; k < SWAP_CLUSTER_COLS; k++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3160) j = (k + col) % SWAP_CLUSTER_COLS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3161) for (i = 0; i < DIV_ROUND_UP(nr_clusters, SWAP_CLUSTER_COLS); i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3162) idx = i * SWAP_CLUSTER_COLS + j;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3163) if (idx >= nr_clusters)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3164) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3165) if (cluster_count(&cluster_info[idx]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3166) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3167) cluster_set_flag(&cluster_info[idx], CLUSTER_FLAG_FREE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3168) cluster_list_add_tail(&p->free_clusters, cluster_info,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3169) idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3170) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3171) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3172) return nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3173) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3175) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3176) * Helper to sys_swapon determining if a given swap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3177) * backing device queue supports DISCARD operations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3178) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3179) static bool swap_discardable(struct swap_info_struct *si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3180) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3181) struct request_queue *q = bdev_get_queue(si->bdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3182)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3183) if (!q || !blk_queue_discard(q))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3184) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3186) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3187) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3189) SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3190) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3191) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3192) struct filename *name;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3193) struct file *swap_file = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3194) struct address_space *mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3195) int prio;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3196) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3197) union swap_header *swap_header;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3198) int nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3199) sector_t span;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3200) unsigned long maxpages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3201) unsigned char *swap_map = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3202) struct swap_cluster_info *cluster_info = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3203) unsigned long *frontswap_map = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3204) struct page *page = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3205) struct inode *inode = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3206) bool inced_nr_rotate_swap = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3208) if (swap_flags & ~SWAP_FLAGS_VALID)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3209) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3211) if (!capable(CAP_SYS_ADMIN))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3212) return -EPERM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3214) if (!swap_avail_heads)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3215) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3216)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3217) p = alloc_swap_info();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3218) if (IS_ERR(p))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3219) return PTR_ERR(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3220)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3221) INIT_WORK(&p->discard_work, swap_discard_work);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3222)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3223) name = getname(specialfile);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3224) if (IS_ERR(name)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3225) error = PTR_ERR(name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3226) name = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3227) goto bad_swap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3228) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3229) swap_file = file_open_name(name, O_RDWR|O_LARGEFILE, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3230) if (IS_ERR(swap_file)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3231) error = PTR_ERR(swap_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3232) swap_file = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3233) goto bad_swap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3234) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3235)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3236) p->swap_file = swap_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3237) mapping = swap_file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3238) inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3240) error = claim_swapfile(p, inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3241) if (unlikely(error))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3242) goto bad_swap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3244) inode_lock(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3245) if (IS_SWAPFILE(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3246) error = -EBUSY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3247) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3248) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3250) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3251) * Read the swap header.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3252) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3253) if (!mapping->a_ops->readpage) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3254) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3255) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3256) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3257) page = read_mapping_page(mapping, 0, swap_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3258) if (IS_ERR(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3259) error = PTR_ERR(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3260) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3261) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3262) swap_header = kmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3264) maxpages = read_swap_header(p, swap_header, inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3265) if (unlikely(!maxpages)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3266) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3267) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3268) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3269)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3270) /* OK, set up the swap map and apply the bad block list */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3271) swap_map = vzalloc(maxpages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3272) if (!swap_map) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3273) error = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3274) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3275) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3276)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3277) if (p->bdev && blk_queue_stable_writes(p->bdev->bd_disk->queue))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3278) p->flags |= SWP_STABLE_WRITES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3279)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3280) if (p->bdev && p->bdev->bd_disk->fops->rw_page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3281) p->flags |= SWP_SYNCHRONOUS_IO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3282)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3283) if (p->bdev && blk_queue_nonrot(bdev_get_queue(p->bdev))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3284) int cpu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3285) unsigned long ci, nr_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3286)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3287) p->flags |= SWP_SOLIDSTATE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3288) p->cluster_next_cpu = alloc_percpu(unsigned int);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3289) if (!p->cluster_next_cpu) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3290) error = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3291) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3292) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3293) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3294) * select a random position to start with to help wear leveling
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3295) * SSD
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3296) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3297) for_each_possible_cpu(cpu) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3298) per_cpu(*p->cluster_next_cpu, cpu) =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3299) 1 + prandom_u32_max(p->highest_bit);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3300) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3301) nr_cluster = DIV_ROUND_UP(maxpages, SWAPFILE_CLUSTER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3302)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3303) cluster_info = kvcalloc(nr_cluster, sizeof(*cluster_info),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3304) GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3305) if (!cluster_info) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3306) error = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3307) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3308) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3309)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3310) for (ci = 0; ci < nr_cluster; ci++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3311) spin_lock_init(&((cluster_info + ci)->lock));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3312)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3313) p->percpu_cluster = alloc_percpu(struct percpu_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3314) if (!p->percpu_cluster) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3315) error = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3316) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3317) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3318) for_each_possible_cpu(cpu) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3319) struct percpu_cluster *cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3320) cluster = per_cpu_ptr(p->percpu_cluster, cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3321) cluster_set_null(&cluster->index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3322) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3323) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3324) atomic_inc(&nr_rotate_swap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3325) inced_nr_rotate_swap = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3326) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3327)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3328) error = swap_cgroup_swapon(p->type, maxpages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3329) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3330) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3331)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3332) nr_extents = setup_swap_map_and_extents(p, swap_header, swap_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3333) cluster_info, maxpages, &span);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3334) if (unlikely(nr_extents < 0)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3335) error = nr_extents;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3336) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3337) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3338) /* frontswap enabled? set up bit-per-page map for frontswap */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3339) if (IS_ENABLED(CONFIG_FRONTSWAP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3340) frontswap_map = kvcalloc(BITS_TO_LONGS(maxpages),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3341) sizeof(long),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3342) GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3344) if (p->bdev &&(swap_flags & SWAP_FLAG_DISCARD) && swap_discardable(p)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3345) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3346) * When discard is enabled for swap with no particular
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3347) * policy flagged, we set all swap discard flags here in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3348) * order to sustain backward compatibility with older
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3349) * swapon(8) releases.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3350) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3351) p->flags |= (SWP_DISCARDABLE | SWP_AREA_DISCARD |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3352) SWP_PAGE_DISCARD);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3354) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3355) * By flagging sys_swapon, a sysadmin can tell us to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3356) * either do single-time area discards only, or to just
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3357) * perform discards for released swap page-clusters.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3358) * Now it's time to adjust the p->flags accordingly.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3359) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3360) if (swap_flags & SWAP_FLAG_DISCARD_ONCE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3361) p->flags &= ~SWP_PAGE_DISCARD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3362) else if (swap_flags & SWAP_FLAG_DISCARD_PAGES)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3363) p->flags &= ~SWP_AREA_DISCARD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3364)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3365) /* issue a swapon-time discard if it's still required */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3366) if (p->flags & SWP_AREA_DISCARD) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3367) int err = discard_swap(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3368) if (unlikely(err))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3369) pr_err("swapon: discard_swap(%p): %d\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3370) p, err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3371) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3372) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3374) error = init_swap_address_space(p->type, maxpages);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3375) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3376) goto bad_swap_unlock_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3378) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3379) * Flush any pending IO and dirty mappings before we start using this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3380) * swap device.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3381) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3382) inode->i_flags |= S_SWAPFILE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3383) error = inode_drain_writes(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3384) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3385) inode->i_flags &= ~S_SWAPFILE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3386) goto free_swap_address_space;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3387) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3389) mutex_lock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3390) prio = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3391) if (swap_flags & SWAP_FLAG_PREFER)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3392) prio =
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3393) (swap_flags & SWAP_FLAG_PRIO_MASK) >> SWAP_FLAG_PRIO_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3394) enable_swap_info(p, prio, swap_map, cluster_info, frontswap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3395)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3396) trace_android_vh_init_swap_info_struct(p, swap_avail_heads);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3397) pr_info("Adding %uk swap on %s. Priority:%d extents:%d across:%lluk %s%s%s%s%s\n",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3398) p->pages<<(PAGE_SHIFT-10), name->name, p->prio,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3399) nr_extents, (unsigned long long)span<<(PAGE_SHIFT-10),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3400) (p->flags & SWP_SOLIDSTATE) ? "SS" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3401) (p->flags & SWP_DISCARDABLE) ? "D" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3402) (p->flags & SWP_AREA_DISCARD) ? "s" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3403) (p->flags & SWP_PAGE_DISCARD) ? "c" : "",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3404) (frontswap_map) ? "FS" : "");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3405)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3406) mutex_unlock(&swapon_mutex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3407) atomic_inc(&proc_poll_event);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3408) wake_up_interruptible(&proc_poll_wait);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3409)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3410) error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3411) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3412) free_swap_address_space:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3413) exit_swap_address_space(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3414) bad_swap_unlock_inode:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3415) inode_unlock(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3416) bad_swap:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3417) free_percpu(p->percpu_cluster);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3418) p->percpu_cluster = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3419) free_percpu(p->cluster_next_cpu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3420) p->cluster_next_cpu = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3421) if (inode && S_ISBLK(inode->i_mode) && p->bdev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3422) set_blocksize(p->bdev, p->old_block_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3423) blkdev_put(p->bdev, FMODE_READ | FMODE_WRITE | FMODE_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3424) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3425) inode = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3426) destroy_swap_extents(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3427) swap_cgroup_swapoff(p->type);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3428) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3429) p->swap_file = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3430) p->flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3431) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3432) vfree(swap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3433) kvfree(cluster_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3434) kvfree(frontswap_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3435) if (inced_nr_rotate_swap)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3436) atomic_dec(&nr_rotate_swap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3437) if (swap_file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3438) filp_close(swap_file, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3439) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3440) if (page && !IS_ERR(page)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3441) kunmap(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3442) put_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3443) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3444) if (name)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3445) putname(name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3446) if (inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3447) inode_unlock(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3448) if (!error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3449) enable_swap_slots_cache();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3450) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3451) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3453) void si_swapinfo(struct sysinfo *val)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3454) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3455) unsigned int type;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3456) unsigned long nr_to_be_unused = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3457)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3458) spin_lock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3459) for (type = 0; type < nr_swapfiles; type++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3460) struct swap_info_struct *si = swap_info[type];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3461) bool skip = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3462)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3463) trace_android_vh_si_swapinfo(si, &skip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3464) if (!skip && (si->flags & SWP_USED) && !(si->flags & SWP_WRITEOK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3465) nr_to_be_unused += si->inuse_pages;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3466) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3467) val->freeswap = atomic_long_read(&nr_swap_pages) + nr_to_be_unused;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3468) val->totalswap = total_swap_pages + nr_to_be_unused;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3469) spin_unlock(&swap_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3470) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3471) EXPORT_SYMBOL_GPL(si_swapinfo);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3473) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3474) * Verify that a swap entry is valid and increment its swap map count.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3475) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3476) * Returns error code in following case.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3477) * - success -> 0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3478) * - swp_entry is invalid -> EINVAL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3479) * - swp_entry is migration entry -> EINVAL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3480) * - swap-cache reference is requested but there is already one. -> EEXIST
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3481) * - swap-cache reference is requested but the entry is not used. -> ENOENT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3482) * - swap-mapped reference requested but needs continued swap count. -> ENOMEM
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3483) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3484) static int __swap_duplicate(swp_entry_t entry, unsigned char usage)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3485) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3486) struct swap_info_struct *p;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3487) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3488) unsigned long offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3489) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3490) unsigned char has_cache;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3491) int err = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3492)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3493) p = get_swap_device(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3494) if (!p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3495) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3496)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3497) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3498) ci = lock_cluster_or_swap_info(p, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3499)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3500) count = p->swap_map[offset];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3501)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3502) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3503) * swapin_readahead() doesn't check if a swap entry is valid, so the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3504) * swap entry could be SWAP_MAP_BAD. Check here with lock held.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3505) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3506) if (unlikely(swap_count(count) == SWAP_MAP_BAD)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3507) err = -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3508) goto unlock_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3509) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3510)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3511) has_cache = count & SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3512) count &= ~SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3513) err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3514)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3515) if (usage == SWAP_HAS_CACHE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3516)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3517) /* set SWAP_HAS_CACHE if there is no cache and entry is used */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3518) if (!has_cache && count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3519) has_cache = SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3520) else if (has_cache) /* someone else added cache */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3521) err = -EEXIST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3522) else /* no users remaining */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3523) err = -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3525) } else if (count || has_cache) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3526)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3527) if ((count & ~COUNT_CONTINUED) < SWAP_MAP_MAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3528) count += usage;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3529) else if ((count & ~COUNT_CONTINUED) > SWAP_MAP_MAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3530) err = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3531) else if (swap_count_continued(p, offset, count))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3532) count = COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3533) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3534) err = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3535) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3536) err = -ENOENT; /* unused swap entry */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3537)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3538) WRITE_ONCE(p->swap_map[offset], count | has_cache);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3539)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3540) unlock_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3541) unlock_cluster_or_swap_info(p, ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3542) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3543) if (p)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3544) put_swap_device(p);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3545) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3546) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3547)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3548) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3549) * Help swapoff by noting that swap entry belongs to shmem/tmpfs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3550) * (in which case its reference count is never incremented).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3551) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3552) void swap_shmem_alloc(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3553) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3554) __swap_duplicate(entry, SWAP_MAP_SHMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3555) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3556)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3557) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3558) * Increase reference count of swap entry by 1.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3559) * Returns 0 for success, or -ENOMEM if a swap_count_continuation is required
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3560) * but could not be atomically allocated. Returns 0, just as if it succeeded,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3561) * if __swap_duplicate() fails for another reason (-EINVAL or -ENOENT), which
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3562) * might occur if a page table entry has got corrupted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3563) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3564) int swap_duplicate(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3565) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3566) int err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3567)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3568) while (!err && __swap_duplicate(entry, 1) == -ENOMEM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3569) err = add_swap_count_continuation(entry, GFP_ATOMIC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3570) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3571) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3573) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3574) * @entry: swap entry for which we allocate swap cache.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3575) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3576) * Called when allocating swap cache for existing swap entry,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3577) * This can return error codes. Returns 0 at success.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3578) * -EEXIST means there is a swap cache.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3579) * Note: return code is different from swap_duplicate().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3580) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3581) int swapcache_prepare(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3582) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3583) return __swap_duplicate(entry, SWAP_HAS_CACHE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3584) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3586) struct swap_info_struct *swp_swap_info(swp_entry_t entry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3587) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3588) return swap_type_to_swap_info(swp_type(entry));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3589) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3590) EXPORT_SYMBOL_GPL(swp_swap_info);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3591)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3592) struct swap_info_struct *page_swap_info(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3593) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3594) swp_entry_t entry = { .val = page_private(page) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3595) return swp_swap_info(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3596) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3597)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3598) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3599) * out-of-line __page_file_ methods to avoid include hell.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3600) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3601) struct address_space *__page_file_mapping(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3602) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3603) return page_swap_info(page)->swap_file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3604) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3605) EXPORT_SYMBOL_GPL(__page_file_mapping);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3607) pgoff_t __page_file_index(struct page *page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3608) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3609) swp_entry_t swap = { .val = page_private(page) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3610) return swp_offset(swap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3611) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3612) EXPORT_SYMBOL_GPL(__page_file_index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3613)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3614) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3615) * add_swap_count_continuation - called when a swap count is duplicated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3616) * beyond SWAP_MAP_MAX, it allocates a new page and links that to the entry's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3617) * page of the original vmalloc'ed swap_map, to hold the continuation count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3618) * (for that entry and for its neighbouring PAGE_SIZE swap entries). Called
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3619) * again when count is duplicated beyond SWAP_MAP_MAX * SWAP_CONT_MAX, etc.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3620) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3621) * These continuation pages are seldom referenced: the common paths all work
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3622) * on the original swap_map, only referring to a continuation page when the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3623) * low "digit" of a count is incremented or decremented through SWAP_MAP_MAX.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3624) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3625) * add_swap_count_continuation(, GFP_ATOMIC) can be called while holding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3626) * page table locks; if it fails, add_swap_count_continuation(, GFP_KERNEL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3627) * can be called after dropping locks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3628) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3629) int add_swap_count_continuation(swp_entry_t entry, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3630) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3631) struct swap_info_struct *si;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3632) struct swap_cluster_info *ci;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3633) struct page *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3634) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3635) struct page *list_page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3636) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3637) unsigned char count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3638) int ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3639)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3640) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3641) * When debugging, it's easier to use __GFP_ZERO here; but it's better
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3642) * for latency not to zero a page while GFP_ATOMIC and holding locks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3643) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3644) page = alloc_page(gfp_mask | __GFP_HIGHMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3645)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3646) si = get_swap_device(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3647) if (!si) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3648) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3649) * An acceptable race has occurred since the failing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3650) * __swap_duplicate(): the swap device may be swapoff
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3651) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3652) goto outer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3653) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3654) spin_lock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3655)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3656) offset = swp_offset(entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3657)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3658) ci = lock_cluster(si, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3659)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3660) count = si->swap_map[offset] & ~SWAP_HAS_CACHE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3661)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3662) if ((count & ~COUNT_CONTINUED) != SWAP_MAP_MAX) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3663) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3664) * The higher the swap count, the more likely it is that tasks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3665) * will race to add swap count continuation: we need to avoid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3666) * over-provisioning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3667) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3668) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3669) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3670)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3671) if (!page) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3672) ret = -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3673) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3674) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3675)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3676) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3677) * We are fortunate that although vmalloc_to_page uses pte_offset_map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3678) * no architecture is using highmem pages for kernel page tables: so it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3679) * will not corrupt the GFP_ATOMIC caller's atomic page table kmaps.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3680) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3681) head = vmalloc_to_page(si->swap_map + offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3682) offset &= ~PAGE_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3683)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3684) spin_lock(&si->cont_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3685) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3686) * Page allocation does not initialize the page's lru field,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3687) * but it does always reset its private field.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3688) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3689) if (!page_private(head)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3690) BUG_ON(count & COUNT_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3691) INIT_LIST_HEAD(&head->lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3692) set_page_private(head, SWP_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3693) si->flags |= SWP_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3694) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3695)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3696) list_for_each_entry(list_page, &head->lru, lru) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3697) unsigned char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3698)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3699) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3700) * If the previous map said no continuation, but we've found
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3701) * a continuation page, free our allocation and use this one.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3702) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3703) if (!(count & COUNT_CONTINUED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3704) goto out_unlock_cont;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3705)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3706) map = kmap_atomic(list_page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3707) count = *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3708) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3709)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3710) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3711) * If this continuation count now has some space in it,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3712) * free our allocation and use this one.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3713) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3714) if ((count & ~COUNT_CONTINUED) != SWAP_CONT_MAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3715) goto out_unlock_cont;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3716) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3717)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3718) list_add_tail(&page->lru, &head->lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3719) page = NULL; /* now it's attached, don't free it */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3720) out_unlock_cont:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3721) spin_unlock(&si->cont_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3722) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3723) unlock_cluster(ci);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3724) spin_unlock(&si->lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3725) put_swap_device(si);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3726) outer:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3727) if (page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3728) __free_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3729) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3730) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3731)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3732) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3733) * swap_count_continued - when the original swap_map count is incremented
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3734) * from SWAP_MAP_MAX, check if there is already a continuation page to carry
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3735) * into, carry if so, or else fail until a new continuation page is allocated;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3736) * when the original swap_map count is decremented from 0 with continuation,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3737) * borrow from the continuation and report whether it still holds more.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3738) * Called while __swap_duplicate() or swap_entry_free() holds swap or cluster
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3739) * lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3740) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3741) static bool swap_count_continued(struct swap_info_struct *si,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3742) pgoff_t offset, unsigned char count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3743) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3744) struct page *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3745) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3746) unsigned char *map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3747) bool ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3748)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3749) head = vmalloc_to_page(si->swap_map + offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3750) if (page_private(head) != SWP_CONTINUED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3751) BUG_ON(count & COUNT_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3752) return false; /* need to add count continuation */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3753) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3754)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3755) spin_lock(&si->cont_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3756) offset &= ~PAGE_MASK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3757) page = list_next_entry(head, lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3758) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3759)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3760) if (count == SWAP_MAP_MAX) /* initial increment from swap_map */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3761) goto init_map; /* jump over SWAP_CONT_MAX checks */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3762)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3763) if (count == (SWAP_MAP_MAX | COUNT_CONTINUED)) { /* incrementing */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3764) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3765) * Think of how you add 1 to 999
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3766) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3767) while (*map == (SWAP_CONT_MAX | COUNT_CONTINUED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3768) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3769) page = list_next_entry(page, lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3770) BUG_ON(page == head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3771) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3772) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3773) if (*map == SWAP_CONT_MAX) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3774) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3775) page = list_next_entry(page, lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3776) if (page == head) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3777) ret = false; /* add count continuation */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3778) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3779) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3780) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3781) init_map: *map = 0; /* we didn't zero the page */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3782) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3783) *map += 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3784) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3785) while ((page = list_prev_entry(page, lru)) != head) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3786) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3787) *map = COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3788) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3789) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3790) ret = true; /* incremented */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3791)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3792) } else { /* decrementing */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3793) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3794) * Think of how you subtract 1 from 1000
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3795) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3796) BUG_ON(count != COUNT_CONTINUED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3797) while (*map == COUNT_CONTINUED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3798) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3799) page = list_next_entry(page, lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3800) BUG_ON(page == head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3801) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3802) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3803) BUG_ON(*map == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3804) *map -= 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3805) if (*map == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3806) count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3807) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3808) while ((page = list_prev_entry(page, lru)) != head) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3809) map = kmap_atomic(page) + offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3810) *map = SWAP_CONT_MAX | count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3811) count = COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3812) kunmap_atomic(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3813) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3814) ret = count == COUNT_CONTINUED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3815) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3816) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3817) spin_unlock(&si->cont_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3818) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3819) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3820)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3821) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3822) * free_swap_count_continuations - swapoff free all the continuation pages
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3823) * appended to the swap_map, after swap_map is quiesced, before vfree'ing it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3824) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3825) static void free_swap_count_continuations(struct swap_info_struct *si)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3826) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3827) pgoff_t offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3828)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3829) for (offset = 0; offset < si->max; offset += PAGE_SIZE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3830) struct page *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3831) head = vmalloc_to_page(si->swap_map + offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3832) if (page_private(head)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3833) struct page *page, *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3834)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3835) list_for_each_entry_safe(page, next, &head->lru, lru) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3836) list_del(&page->lru);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3837) __free_page(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3838) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3839) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3840) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3841) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3842)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3843) #if defined(CONFIG_MEMCG) && defined(CONFIG_BLK_CGROUP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3844) void __cgroup_throttle_swaprate(struct page *page, gfp_t gfp_mask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3845) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3846) struct swap_info_struct *si, *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3847) int nid = page_to_nid(page);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3848)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3849) if (!(gfp_mask & __GFP_IO))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3850) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3851)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3852) if (!blk_cgroup_congested())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3853) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3854)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3855) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3856) * We've already scheduled a throttle, avoid taking the global swap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3857) * lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3858) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3859) if (current->throttle_queue)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3860) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3861)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3862) spin_lock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3863) plist_for_each_entry_safe(si, next, &swap_avail_heads[nid],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3864) avail_lists[nid]) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3865) if (si->bdev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3866) blkcg_schedule_throttle(bdev_get_queue(si->bdev), true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3867) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3868) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3869) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3870) spin_unlock(&swap_avail_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3871) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3872) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3873)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3874) static int __init swapfile_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3875) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3876) int nid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3877)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3878) swap_avail_heads = kmalloc_array(nr_node_ids, sizeof(struct plist_head),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3879) GFP_KERNEL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3880) if (!swap_avail_heads) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3881) pr_emerg("Not enough memory for swap heads, swap is disabled\n");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3882) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3883) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3884)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3885) for_each_node(nid)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3886) plist_head_init(&swap_avail_heads[nid]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3887)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3888) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3889) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3890) subsys_initcall(swapfile_init);
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags