VServer 1.9.2 (patch-2.6.8.1-vs1.9.2.diff)

[linux-2.6.git] / mm / swapfile.c

diff --git a/mm/swapfile.c b/mm/swapfile.c
index fce4cd4..89bc19e 100644 (file)
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -30,10 +30,11 @@
 #include <asm/tlbflush.h>
 #include <linux/swapops.h>
 #include <linux/vs_base.h>
+#include <linux/vs_memory.h>
 
 spinlock_t swaplock = SPIN_LOCK_UNLOCKED;
 unsigned int nr_swapfiles;
-int total_swap_pages;
+long total_swap_pages;
 static int swap_overflow;
 
 EXPORT_SYMBOL(total_swap_pages);
@@ -111,7 +112,7 @@ static inline int scan_swap_map(struct swap_info_struct *si)
  check_next_cluster:
        if (offset+SWAPFILE_CLUSTER-1 <= si->highest_bit)
        {
-               int nr;
+               unsigned long nr;
                for (nr = offset; nr < offset+SWAPFILE_CLUSTER; nr++)
                        if (si->swap_map[nr])
                        {
@@ -469,6 +470,13 @@ static unsigned long unuse_pmd(struct vm_area_struct * vma, pmd_t *dir,
                if (unlikely(pte_same(*pte, swp_pte))) {
                        unuse_pte(vma, offset + address, pte, entry, page);
                        pte_unmap(pte);
+
+                       /*
+                        * Move the page to the active list so it is not
+                        * immediately swapped out again after swapon.
+                        */
+                       activate_page(page);
+
                        /* add 1 since address may be 0 */
                        return 1 + offset + address;
                }
@@ -543,7 +551,15 @@ static int unuse_process(struct mm_struct * mm,
        /*
         * Go through process' page directory.
         */
-       down_read(&mm->mmap_sem);
+       if (!down_read_trylock(&mm->mmap_sem)) {
+               /*
+                * Our reference to the page stops try_to_unmap_one from
+                * unmapping its ptes, so swapoff can make progress.
+                */
+               unlock_page(page);
+               down_read(&mm->mmap_sem);
+               lock_page(page);
+       }
        spin_lock(&mm->page_table_lock);
        for (vma = mm->mmap; vma; vma = vma->vm_next) {
                if (!is_vm_hugetlb_page(vma)) {
@@ -647,7 +663,7 @@ static int try_to_unuse(unsigned int type)
         * open() method called) - so swap entries may be invisible
         * to swapoff for a while, then reappear - but that is rare.
         */
-       while ((i = find_next_to_unuse(si, i))) {
+       while ((i = find_next_to_unuse(si, i)) != 0) {
                if (signal_pending(current)) {
                        retval = -EINTR;
                        break;
@@ -1067,6 +1083,7 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
        unsigned short *swap_map;
        struct file *swap_file, *victim;
        struct address_space *mapping;
+       struct inode *inode;
        char * pathname;
        int i, type, prev;
        int err;
@@ -1079,7 +1096,7 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
        if (IS_ERR(pathname))
                goto out;
 
-       victim = filp_open(pathname, O_RDWR, 0);
+       victim = filp_open(pathname, O_RDWR|O_LARGEFILE, 0);
        putname(pathname);
        err = PTR_ERR(victim);
        if (IS_ERR(victim))
@@ -1160,12 +1177,15 @@ asmlinkage long sys_swapoff(const char __user * specialfile)
        swap_list_unlock();
        up(&swapon_sem);
        vfree(swap_map);
-       if (S_ISBLK(mapping->host->i_mode)) {
-               struct block_device *bdev = I_BDEV(mapping->host);
+       inode = mapping->host;
+       if (S_ISBLK(inode->i_mode)) {
+               struct block_device *bdev = I_BDEV(inode);
                set_blocksize(bdev, p->old_block_size);
                bd_release(bdev);
        } else {
-               up(&mapping->host->i_sem);
+               down(&inode->i_sem);
+               inode->i_flags &= ~S_SWAPFILE;
+               up(&inode->i_sem);
        }
        filp_close(swap_file, NULL);
        err = 0;
@@ -1284,7 +1304,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
        int i, prev;
        int error;
        static int least_priority;
-       union swap_header *swap_header = 0;
+       union swap_header *swap_header = NULL;
        int swap_header_version;
        int nr_good_pages = 0;
        unsigned long maxpages = 1;
@@ -1345,7 +1365,7 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
                name = NULL;
                goto bad_swap_2;
        }
-       swap_file = filp_open(name, O_RDWR, 0);
+       swap_file = filp_open(name, O_RDWR|O_LARGEFILE, 0);
        error = PTR_ERR(swap_file);
        if (IS_ERR(swap_file)) {
                swap_file = NULL;
@@ -1383,6 +1403,10 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags)
                p->bdev = inode->i_sb->s_bdev;
                down(&inode->i_sem);
                did_down = 1;
+               if (IS_SWAPFILE(inode)) {
+                       error = -EBUSY;
+                       goto bad_swap;
+               }
        } else {
                goto bad_swap;
        }
@@ -1555,8 +1579,11 @@ out:
        }
        if (name)
                putname(name);
-       if (error && did_down)
+       if (did_down) {
+               if (!error)
+                       inode->i_flags |= S_SWAPFILE;
                up(&inode->i_sem);
+       }
        return error;
 }