提交 d2997b10 编写于 作者: K KAMEZAWA Hiroyuki 提交者: Linus Torvalds

hibernation: freeze swap at hibernation

When taking a memory snapshot in hibernate_snapshot(), all (directly
called) memory allocations use GFP_ATOMIC.  Hence swap misusage during
hibernation never occurs.

But from a pessimistic point of view, there is no guarantee that no page
allcation has __GFP_WAIT.  It is better to have a global indication "we
enter hibernation, don't use swap!".

This patch tries to freeze new-swap-allocation during hibernation.  (All
user processes are frozenm so swapin is not a concern).

This way, no updates will happen to swap_map[] between
hibernate_snapshot() and save_image().  Swap is thawed when swsusp_free()
is called.  We can be assured that swap corruption will not occur.
Signed-off-by: NKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Hugh Dickins <hughd@google.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Ondrej Zary <linux@rainbow-software.org>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
上级 966cca02
...@@ -316,7 +316,6 @@ extern long nr_swap_pages; ...@@ -316,7 +316,6 @@ extern long nr_swap_pages;
extern long total_swap_pages; extern long total_swap_pages;
extern void si_swapinfo(struct sysinfo *); extern void si_swapinfo(struct sysinfo *);
extern swp_entry_t get_swap_page(void); extern swp_entry_t get_swap_page(void);
extern swp_entry_t get_swap_page_of_type(int);
extern int valid_swaphandles(swp_entry_t, unsigned long *); extern int valid_swaphandles(swp_entry_t, unsigned long *);
extern int add_swap_count_continuation(swp_entry_t, gfp_t); extern int add_swap_count_continuation(swp_entry_t, gfp_t);
extern void swap_shmem_alloc(swp_entry_t); extern void swap_shmem_alloc(swp_entry_t);
...@@ -333,6 +332,13 @@ extern int reuse_swap_page(struct page *); ...@@ -333,6 +332,13 @@ extern int reuse_swap_page(struct page *);
extern int try_to_free_swap(struct page *); extern int try_to_free_swap(struct page *);
struct backing_dev_info; struct backing_dev_info;
#ifdef CONFIG_HIBERNATION
void hibernation_freeze_swap(void);
void hibernation_thaw_swap(void);
swp_entry_t get_swap_for_hibernation(int type);
void swap_free_for_hibernation(swp_entry_t val);
#endif
/* linux/mm/thrash.c */ /* linux/mm/thrash.c */
extern struct mm_struct *swap_token_mm; extern struct mm_struct *swap_token_mm;
extern void grab_swap_token(struct mm_struct *); extern void grab_swap_token(struct mm_struct *);
......
...@@ -338,6 +338,7 @@ int hibernation_snapshot(int platform_mode) ...@@ -338,6 +338,7 @@ int hibernation_snapshot(int platform_mode)
goto Close; goto Close;
suspend_console(); suspend_console();
hibernation_freeze_swap();
saved_mask = clear_gfp_allowed_mask(GFP_IOFS); saved_mask = clear_gfp_allowed_mask(GFP_IOFS);
error = dpm_suspend_start(PMSG_FREEZE); error = dpm_suspend_start(PMSG_FREEZE);
if (error) if (error)
......
...@@ -1086,6 +1086,7 @@ void swsusp_free(void) ...@@ -1086,6 +1086,7 @@ void swsusp_free(void)
buffer = NULL; buffer = NULL;
alloc_normal = 0; alloc_normal = 0;
alloc_highmem = 0; alloc_highmem = 0;
hibernation_thaw_swap();
} }
/* Helper functions used for the shrinking of memory. */ /* Helper functions used for the shrinking of memory. */
......
...@@ -136,10 +136,10 @@ sector_t alloc_swapdev_block(int swap) ...@@ -136,10 +136,10 @@ sector_t alloc_swapdev_block(int swap)
{ {
unsigned long offset; unsigned long offset;
offset = swp_offset(get_swap_page_of_type(swap)); offset = swp_offset(get_swap_for_hibernation(swap));
if (offset) { if (offset) {
if (swsusp_extents_insert(offset)) if (swsusp_extents_insert(offset))
swap_free(swp_entry(swap, offset)); swap_free_for_hibernation(swp_entry(swap, offset));
else else
return swapdev_block(swap, offset); return swapdev_block(swap, offset);
} }
...@@ -163,7 +163,7 @@ void free_all_swap_pages(int swap) ...@@ -163,7 +163,7 @@ void free_all_swap_pages(int swap)
ext = container_of(node, struct swsusp_extent, node); ext = container_of(node, struct swsusp_extent, node);
rb_erase(node, &swsusp_extents); rb_erase(node, &swsusp_extents);
for (offset = ext->start; offset <= ext->end; offset++) for (offset = ext->start; offset <= ext->end; offset++)
swap_free(swp_entry(swap, offset)); swap_free_for_hibernation(swp_entry(swap, offset));
kfree(ext); kfree(ext);
} }
......
...@@ -47,6 +47,8 @@ long nr_swap_pages; ...@@ -47,6 +47,8 @@ long nr_swap_pages;
long total_swap_pages; long total_swap_pages;
static int least_priority; static int least_priority;
static bool swap_for_hibernation;
static const char Bad_file[] = "Bad swap file entry "; static const char Bad_file[] = "Bad swap file entry ";
static const char Unused_file[] = "Unused swap file entry "; static const char Unused_file[] = "Unused swap file entry ";
static const char Bad_offset[] = "Bad swap offset entry "; static const char Bad_offset[] = "Bad swap offset entry ";
...@@ -451,6 +453,8 @@ swp_entry_t get_swap_page(void) ...@@ -451,6 +453,8 @@ swp_entry_t get_swap_page(void)
spin_lock(&swap_lock); spin_lock(&swap_lock);
if (nr_swap_pages <= 0) if (nr_swap_pages <= 0)
goto noswap; goto noswap;
if (swap_for_hibernation)
goto noswap;
nr_swap_pages--; nr_swap_pages--;
for (type = swap_list.next; type >= 0 && wrapped < 2; type = next) { for (type = swap_list.next; type >= 0 && wrapped < 2; type = next) {
...@@ -483,28 +487,6 @@ swp_entry_t get_swap_page(void) ...@@ -483,28 +487,6 @@ swp_entry_t get_swap_page(void)
return (swp_entry_t) {0}; return (swp_entry_t) {0};
} }
/* The only caller of this function is now susupend routine */
swp_entry_t get_swap_page_of_type(int type)
{
struct swap_info_struct *si;
pgoff_t offset;
spin_lock(&swap_lock);
si = swap_info[type];
if (si && (si->flags & SWP_WRITEOK)) {
nr_swap_pages--;
/* This is called for allocating swap entry, not cache */
offset = scan_swap_map(si, 1);
if (offset) {
spin_unlock(&swap_lock);
return swp_entry(type, offset);
}
nr_swap_pages++;
}
spin_unlock(&swap_lock);
return (swp_entry_t) {0};
}
static struct swap_info_struct *swap_info_get(swp_entry_t entry) static struct swap_info_struct *swap_info_get(swp_entry_t entry)
{ {
struct swap_info_struct *p; struct swap_info_struct *p;
...@@ -764,6 +746,74 @@ int mem_cgroup_count_swap_user(swp_entry_t ent, struct page **pagep) ...@@ -764,6 +746,74 @@ int mem_cgroup_count_swap_user(swp_entry_t ent, struct page **pagep)
#endif #endif
#ifdef CONFIG_HIBERNATION #ifdef CONFIG_HIBERNATION
static pgoff_t hibernation_offset[MAX_SWAPFILES];
/*
* Once hibernation starts to use swap, we freeze swap_map[]. Otherwise,
* saved swap_map[] image to the disk will be an incomplete because it's
* changing without synchronization with hibernation snap shot.
* At resume, we just make swap_for_hibernation=false. We can forget
* used maps easily.
*/
void hibernation_freeze_swap(void)
{
int i;
spin_lock(&swap_lock);
printk(KERN_INFO "PM: Freeze Swap\n");
swap_for_hibernation = true;
for (i = 0; i < MAX_SWAPFILES; i++)
hibernation_offset[i] = 1;
spin_unlock(&swap_lock);
}
void hibernation_thaw_swap(void)
{
spin_lock(&swap_lock);
if (swap_for_hibernation) {
printk(KERN_INFO "PM: Thaw Swap\n");
swap_for_hibernation = false;
}
spin_unlock(&swap_lock);
}
/*
* Because updateing swap_map[] can make not-saved-status-change,
* we use our own easy allocator.
* Please see kernel/power/swap.c, Used swaps are recorded into
* RB-tree.
*/
swp_entry_t get_swap_for_hibernation(int type)
{
pgoff_t off;
swp_entry_t val = {0};
struct swap_info_struct *si;
spin_lock(&swap_lock);
si = swap_info[type];
if (!si || !(si->flags & SWP_WRITEOK))
goto done;
for (off = hibernation_offset[type]; off < si->max; ++off) {
if (!si->swap_map[off])
break;
}
if (off < si->max) {
val = swp_entry(type, off);
hibernation_offset[type] = off + 1;
}
done:
spin_unlock(&swap_lock);
return val;
}
void swap_free_for_hibernation(swp_entry_t ent)
{
/* Nothing to do */
}
/* /*
* Find the swap type that corresponds to given device (if any). * Find the swap type that corresponds to given device (if any).
* *
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册