summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
-rw-r--r--Documentation/vm/locking15
-rw-r--r--include/linux/swap.h11
-rw-r--r--mm/filemap.c7
-rw-r--r--mm/rmap.c3
-rw-r--r--mm/swapfile.c125
5 files changed, 66 insertions, 95 deletions
diff --git a/Documentation/vm/locking b/Documentation/vm/locking
index c3ef09ae3bb1..f366fa956179 100644
--- a/Documentation/vm/locking
+++ b/Documentation/vm/locking
@@ -83,19 +83,18 @@ single address space optimization, so that the zap_page_range (from
vmtruncate) does not lose sending ipi's to cloned threads that might
be spawned underneath it and go to user mode to drag in pte's into tlbs.
-swap_list_lock/swap_device_lock
--------------------------------
+swap_lock
+--------------
The swap devices are chained in priority order from the "swap_list" header.
The "swap_list" is used for the round-robin swaphandle allocation strategy.
The #free swaphandles is maintained in "nr_swap_pages". These two together
-are protected by the swap_list_lock.
+are protected by the swap_lock.
-The swap_device_lock, which is per swap device, protects the reference
-counts on the corresponding swaphandles, maintained in the "swap_map"
-array, and the "highest_bit" and "lowest_bit" fields.
+The swap_lock also protects all the device reference counts on the
+corresponding swaphandles, maintained in the "swap_map" array, and the
+"highest_bit" and "lowest_bit" fields.
-Both of these are spinlocks, and are never acquired from intr level. The
-locking hierarchy is swap_list_lock -> swap_device_lock.
+The swap_lock is a spinlock, and is never acquired from intr level.
To prevent races between swap space deletion or async readahead swapins
deciding whether a swap handle is being used, ie worthy of being read in
diff --git a/include/linux/swap.h b/include/linux/swap.h
index db3b5de7c92f..3c9ff0048153 100644
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -121,7 +121,7 @@ enum {
*/
struct swap_info_struct {
unsigned int flags;
- spinlock_t sdev_lock;
+ int prio; /* swap priority */
struct file *swap_file;
struct block_device *bdev;
struct list_head extent_list;
@@ -135,7 +135,6 @@ struct swap_info_struct {
unsigned int pages;
unsigned int max;
unsigned int inuse_pages;
- int prio; /* swap priority */
int next; /* next entry on swap list */
};
@@ -221,13 +220,7 @@ extern int can_share_swap_page(struct page *);
extern int remove_exclusive_swap_page(struct page *);
struct backing_dev_info;
-extern struct swap_list_t swap_list;
-extern spinlock_t swaplock;
-
-#define swap_list_lock() spin_lock(&swaplock)
-#define swap_list_unlock() spin_unlock(&swaplock)
-#define swap_device_lock(p) spin_lock(&p->sdev_lock)
-#define swap_device_unlock(p) spin_unlock(&p->sdev_lock)
+extern spinlock_t swap_lock;
/* linux/mm/thrash.c */
extern struct mm_struct * swap_token_mm;
diff --git a/mm/filemap.c b/mm/filemap.c
index c11418dd94e8..edc54436fa94 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -54,9 +54,8 @@
*
* ->i_mmap_lock (vmtruncate)
* ->private_lock (__free_pte->__set_page_dirty_buffers)
- * ->swap_list_lock
- * ->swap_device_lock (exclusive_swap_page, others)
- * ->mapping->tree_lock
+ * ->swap_lock (exclusive_swap_page, others)
+ * ->mapping->tree_lock
*
* ->i_sem
* ->i_mmap_lock (truncate->unmap_mapping_range)
@@ -86,7 +85,7 @@
* ->page_table_lock (anon_vma_prepare and various)
*
* ->page_table_lock
- * ->swap_device_lock (try_to_unmap_one)
+ * ->swap_lock (try_to_unmap_one)
* ->private_lock (try_to_unmap_one)
* ->tree_lock (try_to_unmap_one)
* ->zone.lru_lock (follow_page->mark_page_accessed)
diff --git a/mm/rmap.c b/mm/rmap.c
index 08ac5c7fa91f..facb8cdca665 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -34,9 +34,8 @@
* anon_vma->lock
* mm->page_table_lock
* zone->lru_lock (in mark_page_accessed)
- * swap_list_lock (in swap_free etc's swap_info_get)
+ * swap_lock (in swap_duplicate, swap_info_get)
* mmlist_lock (in mmput, drain_mmlist and others)
- * swap_device_lock (in swap_duplicate, swap_info_get)
* mapping->private_lock (in __set_page_dirty_buffers)
* inode_lock (in set_page_dirty's __mark_inode_dirty)
* sb_lock (within inode_lock in fs/fs-writeback.c)
diff --git a/mm/swapfile.c b/mm/swapfile.c
index e675ae55f87d..4b6e8bf986bc 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -31,7 +31,7 @@
#include <asm/tlbflush.h>
#include <linux/swapops.h>
-DEFINE_SPINLOCK(swaplock);
+DEFINE_SPINLOCK(swap_lock);
unsigned int nr_swapfiles;
long total_swap_pages;
static int swap_overflow;
@@ -51,7 +51,7 @@ static DECLARE_MUTEX(swapon_sem);
/*
* We need this because the bdev->unplug_fn can sleep and we cannot
- * hold swap_list_lock while calling the unplug_fn. And swap_list_lock
+ * hold swap_lock while calling the unplug_fn. And swap_lock
* cannot be turned into a semaphore.
*/
static DECLARE_RWSEM(swap_unplug_sem);
@@ -105,7 +105,7 @@ static inline unsigned long scan_swap_map(struct swap_info_struct *si)
si->cluster_nr = SWAPFILE_CLUSTER - 1;
if (si->pages - si->inuse_pages < SWAPFILE_CLUSTER)
goto lowest;
- swap_device_unlock(si);
+ spin_unlock(&swap_lock);
offset = si->lowest_bit;
last_in_cluster = offset + SWAPFILE_CLUSTER - 1;
@@ -115,7 +115,7 @@ static inline unsigned long scan_swap_map(struct swap_info_struct *si)
if (si->swap_map[offset])
last_in_cluster = offset + SWAPFILE_CLUSTER;
else if (offset == last_in_cluster) {
- swap_device_lock(si);
+ spin_lock(&swap_lock);
si->cluster_next = offset-SWAPFILE_CLUSTER-1;
goto cluster;
}
@@ -124,7 +124,7 @@ static inline unsigned long scan_swap_map(struct swap_info_struct *si)
latency_ration = LATENCY_LIMIT;
}
}
- swap_device_lock(si);
+ spin_lock(&swap_lock);
goto lowest;
}
@@ -153,10 +153,10 @@ checks: if (!(si->flags & SWP_WRITEOK))
return offset;
}
- swap_device_unlock(si);
+ spin_unlock(&swap_lock);
while (++offset <= si->highest_bit) {
if (!si->swap_map[offset]) {
- swap_device_lock(si);
+ spin_lock(&swap_lock);
goto checks;
}
if (unlikely(--latency_ration < 0)) {
@@ -164,7 +164,7 @@ checks: if (!(si->flags & SWP_WRITEOK))
latency_ration = LATENCY_LIMIT;
}
}
- swap_device_lock(si);
+ spin_lock(&swap_lock);
goto lowest;
no_page:
@@ -179,7 +179,7 @@ swp_entry_t get_swap_page(void)
int type, next;
int wrapped = 0;
- swap_list_lock();
+ spin_lock(&swap_lock);
if (nr_swap_pages <= 0)
goto noswap;
nr_swap_pages--;
@@ -199,19 +199,17 @@ swp_entry_t get_swap_page(void)
continue;
swap_list.next = next;
- swap_device_lock(si);
- swap_list_unlock();
offset = scan_swap_map(si);
- swap_device_unlock(si);
- if (offset)
+ if (offset) {
+ spin_unlock(&swap_lock);
return swp_entry(type, offset);
- swap_list_lock();
+ }
next = swap_list.next;
}
nr_swap_pages++;
noswap:
- swap_list_unlock();
+ spin_unlock(&swap_lock);
return (swp_entry_t) {0};
}
@@ -233,8 +231,7 @@ static struct swap_info_struct * swap_info_get(swp_entry_t entry)
goto bad_offset;
if (!p->swap_map[offset])
goto bad_free;
- swap_list_lock();
- swap_device_lock(p);
+ spin_lock(&swap_lock);
return p;
bad_free:
@@ -252,12 +249,6 @@ out:
return NULL;
}
-static void swap_info_put(struct swap_info_struct * p)
-{
- swap_device_unlock(p);
- swap_list_unlock();
-}
-
static int swap_entry_free(struct swap_info_struct *p, unsigned long offset)
{
int count = p->swap_map[offset];
@@ -290,7 +281,7 @@ void swap_free(swp_entry_t entry)
p = swap_info_get(entry);
if (p) {
swap_entry_free(p, swp_offset(entry));
- swap_info_put(p);
+ spin_unlock(&swap_lock);
}
}
@@ -308,7 +299,7 @@ static inline int page_swapcount(struct page *page)
if (p) {
/* Subtract the 1 for the swap cache itself */
count = p->swap_map[swp_offset(entry)] - 1;
- swap_info_put(p);
+ spin_unlock(&swap_lock);
}
return count;
}
@@ -365,7 +356,7 @@ int remove_exclusive_swap_page(struct page *page)
}
write_unlock_irq(&swapper_space.tree_lock);
}
- swap_info_put(p);
+ spin_unlock(&swap_lock);
if (retval) {
swap_free(entry);
@@ -388,7 +379,7 @@ void free_swap_and_cache(swp_entry_t entry)
if (p) {
if (swap_entry_free(p, swp_offset(entry)) == 1)
page = find_trylock_page(&swapper_space, entry.val);
- swap_info_put(p);
+ spin_unlock(&swap_lock);
}
if (page) {
int one_user;
@@ -558,10 +549,10 @@ static unsigned int find_next_to_unuse(struct swap_info_struct *si,
int count;
/*
- * No need for swap_device_lock(si) here: we're just looking
+ * No need for swap_lock here: we're just looking
* for whether an entry is in use, not modifying it; false
* hits are okay, and sys_swapoff() has already prevented new
- * allocations from this area (while holding swap_list_lock()).
+ * allocations from this area (while holding swap_lock).
*/
for (;;) {
if (++i >= max) {
@@ -751,9 +742,9 @@ static int try_to_unuse(unsigned int type)
* report them; but do report if we reset SWAP_MAP_MAX.
*/
if (*swap_map == SWAP_MAP_MAX) {
- swap_device_lock(si);
+ spin_lock(&swap_lock);
*swap_map = 1;
- swap_device_unlock(si);
+ spin_unlock(&swap_lock);
reset_overflow = 1;
}
@@ -817,9 +808,9 @@ static int try_to_unuse(unsigned int type)
}
/*
- * After a successful try_to_unuse, if no swap is now in use, we know we
- * can empty the mmlist. swap_list_lock must be held on entry and exit.
- * Note that mmlist_lock nests inside swap_list_lock, and an mm must be
+ * After a successful try_to_unuse, if no swap is now in use, we know
+ * we can empty the mmlist. swap_lock must be held on entry and exit.
+ * Note that mmlist_lock nests inside swap_lock, and an mm must be
* added to the mmlist just after page_duplicate - before would be racy.
*/
static void drain_mmlist(void)
@@ -1092,7 +1083,7 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
mapping = victim->f_mapping;
prev = -1;
- swap_list_lock();
+ spin_lock(&swap_lock);
for (type = swap_list.head; type >= 0; type = swap_info[type].next) {
p = swap_info + type;
if ((p->flags & SWP_ACTIVE) == SWP_ACTIVE) {
@@ -1103,14 +1094,14 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
}
if (type < 0) {
err = -EINVAL;
- swap_list_unlock();
+ spin_unlock(&swap_lock);
goto out_dput;
}
if (!security_vm_enough_memory(p->pages))
vm_unacct_memory(p->pages);
else {
err = -ENOMEM;
- swap_list_unlock();
+ spin_unlock(&swap_lock);
goto out_dput;
}
if (prev < 0) {
@@ -1124,10 +1115,8 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
}
nr_swap_pages -= p->pages;
total_swap_pages -= p->pages;
- swap_device_lock(p);
p->flags &= ~SWP_WRITEOK;
- swap_device_unlock(p);
- swap_list_unlock();
+ spin_unlock(&swap_lock);
current->flags |= PF_SWAPOFF;
err = try_to_unuse(type);
@@ -1135,7 +1124,7 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
if (err) {
/* re-insert swap space back into swap_list */
- swap_list_lock();
+ spin_lock(&swap_lock);
for (prev = -1, i = swap_list.head; i >= 0; prev = i, i = swap_info[i].next)
if (p->prio >= swap_info[i].prio)
break;
@@ -1146,10 +1135,8 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
swap_info[prev].next = p - swap_info;
nr_swap_pages += p->pages;
total_swap_pages += p->pages;
- swap_device_lock(p);
p->flags |= SWP_WRITEOK;
- swap_device_unlock(p);
- swap_list_unlock();
+ spin_unlock(&swap_lock);
goto out_dput;
}
@@ -1157,30 +1144,27 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
down_write(&swap_unplug_sem);
up_write(&swap_unplug_sem);
+ destroy_swap_extents(p);
+ down(&swapon_sem);
+ spin_lock(&swap_lock);
+ drain_mmlist();
+
/* wait for anyone still in scan_swap_map */
- swap_device_lock(p);
p->highest_bit = 0; /* cuts scans short */
while (p->flags >= SWP_SCANNING) {
- swap_device_unlock(p);
+ spin_unlock(&swap_lock);
set_current_state(TASK_UNINTERRUPTIBLE);
schedule_timeout(1);
- swap_device_lock(p);
+ spin_lock(&swap_lock);
}
- swap_device_unlock(p);
- destroy_swap_extents(p);
- down(&swapon_sem);
- swap_list_lock();
- drain_mmlist();
- swap_device_lock(p);
swap_file = p->swap_file;
p->swap_file = NULL;
p->max = 0;
swap_map = p->swap_map;
p->swap_map = NULL;
p->flags = 0;
- swap_device_unlock(p);
- swap_list_unlock();
+ spin_unlock(&swap_lock);
up(&swapon_sem);
vfree(swap_map);
inode = mapping->host;
@@ -1324,7 +1308,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
if (!capable(CAP_SYS_ADMIN))
return -EPERM;
- swap_list_lock();
+ spin_lock(&swap_lock);
p = swap_info;
for (type = 0 ; type < nr_swapfiles ; type++,p++)
if (!(p->flags & SWP_USED))
@@ -1343,7 +1327,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
* swp_entry_t or the architecture definition of a swap pte.
*/
if (type > swp_type(pte_to_swp_entry(swp_entry_to_pte(swp_entry(~0UL,0))))) {
- swap_list_unlock();
+ spin_unlock(&swap_lock);
goto out;
}
if (type >= nr_swapfiles)
@@ -1357,7 +1341,6 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
p->highest_bit = 0;
p->cluster_nr = 0;
p->inuse_pages = 0;
- spin_lock_init(&p->sdev_lock);
p->next = -1;
if (swap_flags & SWAP_FLAG_PREFER) {
p->prio =
@@ -1365,7 +1348,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
} else {
p->prio = --least_priority;
}
- swap_list_unlock();
+ spin_unlock(&swap_lock);
name = getname(specialfile);
error = PTR_ERR(name);
if (IS_ERR(name)) {
@@ -1542,8 +1525,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
}
down(&swapon_sem);
- swap_list_lock();
- swap_device_lock(p);
+ spin_lock(&swap_lock);
p->flags = SWP_ACTIVE;
nr_swap_pages += nr_good_pages;
total_swap_pages += nr_good_pages;
@@ -1567,8 +1549,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
} else {
swap_info[prev].next = p - swap_info;
}
- swap_device_unlock(p);
- swap_list_unlock();
+ spin_unlock(&swap_lock);
up(&swapon_sem);
error = 0;
goto out;
@@ -1579,14 +1560,14 @@ bad_swap:
}
destroy_swap_extents(p);
bad_swap_2:
- swap_list_lock();
+ spin_lock(&swap_lock);
swap_map = p->swap_map;
p->swap_file = NULL;
p->swap_map = NULL;
p->flags = 0;
if (!(swap_flags & SWAP_FLAG_PREFER))
++least_priority;
- swap_list_unlock();
+ spin_unlock(&swap_lock);
vfree(swap_map);
if (swap_file)
filp_close(swap_file, NULL);
@@ -1610,7 +1591,7 @@ void si_swapinfo(struct sysinfo *val)
unsigned int i;
unsigned long nr_to_be_unused = 0;
- swap_list_lock();
+ spin_lock(&swap_lock);
for (i = 0; i < nr_swapfiles; i++) {
if (!(swap_info[i].flags & SWP_USED) ||
(swap_info[i].flags & SWP_WRITEOK))
@@ -1619,7 +1600,7 @@ void si_swapinfo(struct sysinfo *val)
}
val->freeswap = nr_swap_pages + nr_to_be_unused;
val->totalswap = total_swap_pages + nr_to_be_unused;
- swap_list_unlock();
+ spin_unlock(&swap_lock);
}
/*
@@ -1640,7 +1621,7 @@ int swap_duplicate(swp_entry_t entry)
p = type + swap_info;
offset = swp_offset(entry);
- swap_device_lock(p);
+ spin_lock(&swap_lock);
if (offset < p->max && p->swap_map[offset]) {
if (p->swap_map[offset] < SWAP_MAP_MAX - 1) {
p->swap_map[offset]++;
@@ -1652,7 +1633,7 @@ int swap_duplicate(swp_entry_t entry)
result = 1;
}
}
- swap_device_unlock(p);
+ spin_unlock(&swap_lock);
out:
return result;
@@ -1668,7 +1649,7 @@ get_swap_info_struct(unsigned type)
}
/*
- * swap_device_lock prevents swap_map being freed. Don't grab an extra
+ * swap_lock prevents swap_map being freed. Don't grab an extra
* reference on the swaphandle, it doesn't matter if it becomes unused.
*/
int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
@@ -1684,7 +1665,7 @@ int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
toff++, i--;
*offset = toff;
- swap_device_lock(swapdev);
+ spin_lock(&swap_lock);
do {
/* Don't read-ahead past the end of the swap area */
if (toff >= swapdev->max)
@@ -1697,6 +1678,6 @@ int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
toff++;
ret++;
} while (--i);
- swap_device_unlock(swapdev);
+ spin_unlock(&swap_lock);
return ret;
}