/**
* vxfs_get_page - read a page into memory.
* @ip: inode to read from
* @n: page number
*
* Description:
* vxfs_get_page reads the @n th page of @ip into the pagecache.
*
* Returns:
* The wanted page on success, else a NULL pointer.
*/
struct page *
vxfs_get_page(struct address_space *mapping, u_long n)
{
struct page * pp;
pp = read_cache_page(mapping, n,
(filler_t*)mapping->a_ops->readpage, NULL);
if (!IS_ERR(pp)) {
wait_on_page(pp);
kmap(pp);
if (!Page_Uptodate(pp))
goto fail;
/** if (!PageChecked(pp)) **/
/** vxfs_check_page(pp); **/
if (PageError(pp))
goto fail;
}
return (pp);
fail:
vxfs_put_page(pp);
return ERR_PTR(-EIO);
}
/* read a range of the data via the page cache */
static int blkmtd_read(struct mtd_info *mtd, loff_t from, size_t len,
size_t *retlen, u_char *buf)
{
mtd_raw_dev_data_t *rawdevice = mtd->priv;
int err = 0;
int offset;
int pagenr, pages;
*retlen = 0;
DEBUG(2, "blkmtd: read: dev = `%s' from = %ld len = %d buf = %p\n",
bdevname(rawdevice->binding->bd_dev), (long int)from, len, buf);
pagenr = from >> PAGE_SHIFT;
offset = from - (pagenr << PAGE_SHIFT);
pages = (offset+len+PAGE_SIZE-1) >> PAGE_SHIFT;
DEBUG(3, "blkmtd: read: pagenr = %d offset = %d, pages = %d\n", pagenr, offset, pages);
/* just loop through each page, getting it via readpage() - slow but easy */
while(pages) {
struct page *page;
int cpylen;
DEBUG(3, "blkmtd: read: looking for page: %d\n", pagenr);
page = read_cache_page(&rawdevice->as, pagenr, (filler_t *)blkmtd_readpage, rawdevice->file);
if(IS_ERR(page)) {
return PTR_ERR(page);
}
wait_on_page(page);
if(!Page_Uptodate(page)) {
/* error reading page */
printk("blkmtd: read: page not uptodate\n");
page_cache_release(page);
return -EIO;
}
cpylen = (PAGE_SIZE > len) ? len : PAGE_SIZE;
if(offset+cpylen > PAGE_SIZE)
cpylen = PAGE_SIZE-offset;
memcpy(buf + *retlen, page_address(page) + offset, cpylen);
offset = 0;
len -= cpylen;
*retlen += cpylen;
pagenr++;
pages--;
page_cache_release(page);
}
DEBUG(2, "blkmtd: end read: retlen = %d, err = %d\n", *retlen, err);
return err;
}
static struct page * dir_get_page(struct inode *dir, unsigned long n)
{
struct address_space *mapping = dir->i_mapping;
struct page *page = read_cache_page(mapping, n,
(filler_t*)mapping->a_ops->readpage, NULL);
if (!IS_ERR(page)) {
wait_on_page(page);
kmap(page);
if (!Page_Uptodate(page))
goto fail;
}
return page;
fail:
dir_put_page(page);
return ERR_PTR(-EIO);
}
/*
* The swap lock map insists that pages be in the page cache!
* Therefore we can't use it. Later when we can remove the need for the
* lock map and we can reduce the number of functions exported.
*/
void rw_swap_page_nolock(int rw, swp_entry_t entry, char *buf)
{
struct page *page = virt_to_page(buf);
if (!PageLocked(page))
PAGE_BUG(page);
if (PageSwapCache(page))
PAGE_BUG(page);
if (page->mapping)
PAGE_BUG(page);
/* needs sync_page to wait I/O completation */
page->mapping = &swapper_space;
if (!rw_swap_page_base(rw, entry, page))
UnlockPage(page);
wait_on_page(page);
page->mapping = NULL;
}
/*
* Setting up a new swap file needs a simple wrapper just to read the
* swap signature. SysV shared memory also needs a simple wrapper.
*/
void rw_swap_page_nocache(int rw, unsigned long entry, char *buffer)
{
struct page *page;
page = mem_map + MAP_NR((unsigned long) buffer);
wait_on_page(page);
set_bit(PG_locked, &page->flags);
if (test_and_set_bit(PG_swap_cache, &page->flags)) {
printk ("VM: read_swap_page: page already in swap cache!\n");
return;
}
if (page->inode) {
printk ("VM: read_swap_page: page already in page cache!\n");
return;
}
page->inode = &swapper_inode;
page->offset = entry;
atomic_inc(&page->count); /* Protect from shrink_mmap() */
rw_swap_page(rw, entry, buffer, 1);
atomic_dec(&page->count);
page->inode = 0;
clear_bit(PG_swap_cache, &page->flags);
}
static int shrink_cache(int nr_pages, zone_t * classzone, unsigned int gfp_mask, int priority)
{
struct list_head * entry;
int max_scan = nr_inactive_pages / priority;
int max_mapped = min((nr_pages << (10 - priority)), max_scan / 10);
spin_lock(&pagemap_lru_lock);
while (--max_scan >= 0 && (entry = inactive_list.prev) != &inactive_list) {
struct page * page;
/* lock depth is 1 or 2 */
if (unlikely(current->need_resched)) {
spin_unlock(&pagemap_lru_lock);
__set_current_state(TASK_RUNNING);
schedule();
spin_lock(&pagemap_lru_lock);
continue;
}
page = list_entry(entry, struct page, lru);
if (unlikely(!PageLRU(page)))
BUG();
if (unlikely(PageActive(page)))
BUG();
list_del(entry);
list_add(entry, &inactive_list);
/*
* Zero page counts can happen because we unlink the pages
* _after_ decrementing the usage count..
*/
if (unlikely(!page_count(page)))
continue;
if (!memclass(page->zone, classzone))
continue;
/* Racy check to avoid trylocking when not worthwhile */
if (!page->buffers && (page_count(page) != 1 || !page->mapping))
goto page_mapped;
/*
* The page is locked. IO in progress?
* Move it to the back of the list.
*/
if (unlikely(TryLockPage(page))) {
if (PageLaunder(page) && (gfp_mask & __GFP_FS)) {
page_cache_get(page);
spin_unlock(&pagemap_lru_lock);
wait_on_page(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
}
continue;
}
if ((PageDirty(page) || DelallocPage(page)) && is_page_cache_freeable(page) && page->mapping) {
/*
* It is not critical here to write it only if
* the page is unmapped beause any direct writer
* like O_DIRECT would set the PG_dirty bitflag
* on the phisical page after having successfully
* pinned it and after the I/O to the page is finished,
* so the direct writes to the page cannot get lost.
*/
int (*writepage)(struct page *);
writepage = page->mapping->a_ops->writepage;
if ((gfp_mask & __GFP_FS) && writepage) {
ClearPageDirty(page);
SetPageLaunder(page);
page_cache_get(page);
spin_unlock(&pagemap_lru_lock);
writepage(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
continue;
}
}
/*
* If the page has buffers, try to free the buffer mappings
* associated with this page. If we succeed we try to free
* the page as well.
*/
if (page->buffers) {
spin_unlock(&pagemap_lru_lock);
/* avoid to free a locked page */
page_cache_get(page);
if (try_to_release_page(page, gfp_mask)) {
if (!page->mapping) {
/*
* We must not allow an anon page
* with no buffers to be visible on
* the LRU, so we unlock the page after
* taking the lru lock
*/
spin_lock(&pagemap_lru_lock);
UnlockPage(page);
__lru_cache_del(page);
/* effectively free the page here */
page_cache_release(page);
if (--nr_pages)
continue;
break;
} else {
/*
* The page is still in pagecache so undo the stuff
* before the try_to_release_page since we've not
* finished and we can now try the next step.
*/
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
}
} else {
/* failed to drop the buffers so stop here */
UnlockPage(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
continue;
}
}
spin_lock(&pagecache_lock);
/*
* this is the non-racy check for busy page.
*/
if (!page->mapping || !is_page_cache_freeable(page)) {
spin_unlock(&pagecache_lock);
UnlockPage(page);
page_mapped:
if (--max_mapped >= 0)
continue;
/*
* Alert! We've found too many mapped pages on the
* inactive list, so we start swapping out now!
*/
spin_unlock(&pagemap_lru_lock);
swap_out(priority, gfp_mask, classzone);
return nr_pages;
}
/*
* It is critical to check PageDirty _after_ we made sure
* the page is freeable* so not in use by anybody.
*/
if (PageDirty(page)) {
spin_unlock(&pagecache_lock);
UnlockPage(page);
continue;
}
/* point of no return */
if (likely(!PageSwapCache(page))) {
__remove_inode_page(page);
spin_unlock(&pagecache_lock);
} else {
swp_entry_t swap;
swap.val = page->index;
__delete_from_swap_cache(page);
spin_unlock(&pagecache_lock);
swap_free(swap);
}
__lru_cache_del(page);
UnlockPage(page);
/* effectively free the page here */
page_cache_release(page);
if (--nr_pages)
continue;
break;
}
spin_unlock(&pagemap_lru_lock);
return nr_pages;
}
int lock_kiovec(int nr, struct kiobuf *iovec[], int wait)
{
struct kiobuf *iobuf;
int i, j;
struct page *page, **ppage;
int doublepage = 0;
int repeat = 0;
repeat:
for (i = 0; i < nr; i++) {
iobuf = iovec[i];
if (iobuf->locked)
continue;
iobuf->locked = 1;
ppage = iobuf->maplist;
for (j = 0; j < iobuf->nr_pages; ppage++, j++) {
page = *ppage;
if (!page)
continue;
if (TryLockPage(page))
goto retry;
}
}
return 0;
retry:
/*
* We couldn't lock one of the pages. Undo the locking so far,
* wait on the page we got to, and try again.
*/
unlock_kiovec(nr, iovec);
if (!wait)
return -EAGAIN;
/*
* Did the release also unlock the page we got stuck on?
*/
if (!PageLocked(page)) {
/*
* If so, we may well have the page mapped twice
* in the IO address range. Bad news. Of
* course, it _might_ just be a coincidence,
* but if it happens more than once, chances
* are we have a double-mapped page.
*/
if (++doublepage >= 3)
return -EINVAL;
/* Try again... */
wait_on_page(page);
}
if (++repeat < 16)
goto repeat;
return -EAGAIN;
}
/*
* We completely avoid races by reading each swap page in advance,
* and then search for the process using it. All the necessary
* page table adjustments can then be made atomically.
*/
static int try_to_unuse(unsigned int type)
{
struct swap_info_struct * si = &swap_info[type];
struct mm_struct *start_mm;
unsigned short *swap_map;
unsigned short swcount;
struct page *page;
swp_entry_t entry;
int i = 0;
int retval = 0;
int reset_overflow = 0;
/*
* When searching mms for an entry, a good strategy is to
* start at the first mm we freed the previous entry from
* (though actually we don't notice whether we or coincidence
* freed the entry). Initialize this start_mm with a hold.
*
* A simpler strategy would be to start at the last mm we
* freed the previous entry from; but that would take less
* advantage of mmlist ordering (now preserved by swap_out()),
* which clusters forked address spaces together, most recent
* child immediately after parent. If we race with dup_mmap(),
* we very much want to resolve parent before child, otherwise
* we may miss some entries: using last mm would invert that.
*/
start_mm = &init_mm;
atomic_inc(&init_mm.mm_users);
/*
* Keep on scanning until all entries have gone. Usually,
* one pass through swap_map is enough, but not necessarily:
* mmput() removes mm from mmlist before exit_mmap() and its
* zap_page_range(). That's not too bad, those entries are
* on their way out, and handled faster there than here.
* do_munmap() behaves similarly, taking the range out of mm's
* vma list before zap_page_range(). But unfortunately, when
* unmapping a part of a vma, it takes the whole out first,
* then reinserts what's left after (might even reschedule if
* 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))) {
/*
* Get a page for the entry, using the existing swap
* cache page if there is one. Otherwise, get a clean
* page and read the swap into it.
*/
swap_map = &si->swap_map[i];
entry = SWP_ENTRY(type, i);
page = read_swap_cache_async(entry);
if (!page) {
/*
* Either swap_duplicate() failed because entry
* has been freed independently, and will not be
* reused since sys_swapoff() already disabled
* allocation from here, or alloc_page() failed.
*/
if (!*swap_map)
continue;
retval = -ENOMEM;
break;
}
/*
* Don't hold on to start_mm if it looks like exiting.
*/
if (atomic_read(&start_mm->mm_users) == 1) {
mmput(start_mm);
start_mm = &init_mm;
atomic_inc(&init_mm.mm_users);
}
/*
* Wait for and lock page. When do_swap_page races with
* try_to_unuse, do_swap_page can handle the fault much
* faster than try_to_unuse can locate the entry. This
* apparently redundant "wait_on_page" lets try_to_unuse
* defer to do_swap_page in such a case - in some tests,
* do_swap_page and try_to_unuse repeatedly compete.
*/
wait_on_page(page);
lock_page(page);
/*
* Remove all references to entry, without blocking.
* Whenever we reach init_mm, there's no address space
* to search, but use it as a reminder to search shmem.
*/
swcount = *swap_map;
if (swcount > 1) {
flush_page_to_ram(page);
if (start_mm == &init_mm)
shmem_unuse(entry, page);
else
unuse_process(start_mm, entry, page);
}
if (*swap_map > 1) {
int set_start_mm = (*swap_map >= swcount);
struct list_head *p = &start_mm->mmlist;
struct mm_struct *new_start_mm = start_mm;
struct mm_struct *mm;
spin_lock(&mmlist_lock);
while (*swap_map > 1 &&
(p = p->next) != &start_mm->mmlist) {
mm = list_entry(p, struct mm_struct, mmlist);
swcount = *swap_map;
if (mm == &init_mm) {
set_start_mm = 1;
shmem_unuse(entry, page);
} else
unuse_process(mm, entry, page);
if (set_start_mm && *swap_map < swcount) {
new_start_mm = mm;
set_start_mm = 0;
}
}
atomic_inc(&new_start_mm->mm_users);
spin_unlock(&mmlist_lock);
mmput(start_mm);
start_mm = new_start_mm;
}
/*
* How could swap count reach 0x7fff when the maximum
* pid is 0x7fff, and there's no way to repeat a swap
* page within an mm (except in shmem, where it's the
* shared object which takes the reference count)?
* We believe SWAP_MAP_MAX cannot occur in Linux 2.4.
*
* If that's wrong, then we should worry more about
* exit_mmap() and do_munmap() cases described above:
* we might be resetting SWAP_MAP_MAX too early here.
* We know "Undead"s can happen, they're okay, so don't
* report them; but do report if we reset SWAP_MAP_MAX.
*/
if (*swap_map == SWAP_MAP_MAX) {
swap_list_lock();
swap_device_lock(si);
nr_swap_pages++;
*swap_map = 1;
swap_device_unlock(si);
swap_list_unlock();
reset_overflow = 1;
}
/*
* If a reference remains (rare), we would like to leave
* the page in the swap cache; but try_to_swap_out could
* then re-duplicate the entry once we drop page lock,
* so we might loop indefinitely; also, that page could
* not be swapped out to other storage meanwhile. So:
* delete from cache even if there's another reference,
* after ensuring that the data has been saved to disk -
* since if the reference remains (rarer), it will be
* read from disk into another page. Splitting into two
* pages would be incorrect if swap supported "shared
* private" pages, but they are handled by tmpfs files.
* Note shmem_unuse already deleted its from swap cache.
*/
if ((*swap_map > 1) && PageDirty(page) && PageSwapCache(page)) {
rw_swap_page(WRITE, page);
lock_page(page);
}
if (PageSwapCache(page))
delete_from_swap_cache(page);
/*
* So we could skip searching mms once swap count went
* to 1, we did not mark any present ptes as dirty: must
* mark page dirty so try_to_swap_out will preserve it.
*/
SetPageDirty(page);
UnlockPage(page);
page_cache_release(page);
/*
* Make sure that we aren't completely killing
* interactive performance. Interruptible check on
* signal_pending() would be nice, but changes the spec?
*/
if (current->need_resched)
schedule();
}
/*
* Reads or writes a swap page.
* wait=1: start I/O and wait for completion. wait=0: start asynchronous I/O.
*
* Important prevention of race condition: The first thing we do is set a lock
* on this swap page, which lasts until I/O completes. This way a
* write_swap_page(entry) immediately followed by a read_swap_page(entry)
* on the same entry will first complete the write_swap_page(). Fortunately,
* not more than one write_swap_page() request can be pending per entry. So
* all races the caller must catch are: multiple read_swap_page() requests
* on the same entry.
*/
void rw_swap_page(int rw, unsigned long entry, char * buf, int wait)
{
unsigned long type, offset;
struct swap_info_struct * p;
struct page *page;
type = SWP_TYPE(entry);
if (type >= nr_swapfiles) {
printk("Internal error: bad swap-device\n");
return;
}
p = &swap_info[type];
offset = SWP_OFFSET(entry);
if (offset >= p->max) {
printk("rw_swap_page: weirdness\n");
return;
}
if (p->swap_map && !p->swap_map[offset]) {
printk("Hmm.. Trying to use unallocated swap (%08lx)\n", entry);
return;
}
if (!(p->flags & SWP_USED)) {
printk("Trying to swap to unused swap-device\n");
return;
}
/* Make sure we are the only process doing I/O with this swap page. */
while (set_bit(offset,p->swap_lockmap)) {
run_task_queue(&tq_disk);
sleep_on(&lock_queue);
}
if (rw == READ)
kstat.pswpin++;
else
kstat.pswpout++;
page = mem_map + MAP_NR(buf);
atomic_inc(&page->count);
wait_on_page(page);
if (p->swap_device) {
if (!wait) {
set_bit(PG_free_after, &page->flags);
set_bit(PG_decr_after, &page->flags);
set_bit(PG_swap_unlock_after, &page->flags);
page->swap_unlock_entry = entry;
atomic_inc(&nr_async_pages);
}
ll_rw_page(rw,p->swap_device,offset,buf);
/*
* NOTE! We don't decrement the page count if we
* don't wait - that will happen asynchronously
* when the IO completes.
*/
if (!wait)
return;
wait_on_page(page);
} else if (p->swap_file) {
struct inode *swapf = p->swap_file;
unsigned int zones[PAGE_SIZE/512];
int i;
if (swapf->i_op->bmap == NULL
&& swapf->i_op->smap != NULL){
/*
With MsDOS, we use msdos_smap which return
a sector number (not a cluster or block number).
It is a patch to enable the UMSDOS project.
Other people are working on better solution.
It sounds like ll_rw_swap_file defined
it operation size (sector size) based on
PAGE_SIZE and the number of block to read.
So using bmap or smap should work even if
smap will require more blocks.
*/
int j;
unsigned int block = offset << 3;
for (i=0, j=0; j< PAGE_SIZE ; i++, j += 512){
if (!(zones[i] = swapf->i_op->smap(swapf,block++))) {
printk("rw_swap_page: bad swap file\n");
return;
}
}
}else{
int j;
unsigned int block = offset
<< (PAGE_SHIFT - swapf->i_sb->s_blocksize_bits);
for (i=0, j=0; j< PAGE_SIZE ; i++, j +=swapf->i_sb->s_blocksize)
if (!(zones[i] = bmap(swapf,block++))) {
printk("rw_swap_page: bad swap file\n");
}
}
ll_rw_swap_file(rw,swapf->i_dev, zones, i,buf);
} else
printk("rw_swap_page: no swap file or device\n");
atomic_dec(&page->count);
if (offset && !clear_bit(offset,p->swap_lockmap))
printk("rw_swap_page: lock already cleared\n");
wake_up(&lock_queue);
}
/*
* Returns a pointer to a buffer containing at least LEN bytes of
* filesystem starting at byte offset OFFSET into the filesystem.
*/
static void *cramfs_read(struct super_block *sb, unsigned int offset, unsigned int len)
{
struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
struct page *pages[BLKS_PER_BUF];
unsigned i, blocknr, buffer, unread;
unsigned long devsize;
int major, minor;
char *data;
if (!len)
return NULL;
blocknr = offset >> PAGE_CACHE_SHIFT;
offset &= PAGE_CACHE_SIZE - 1;
/* Check if an existing buffer already has the data.. */
for (i = 0; i < READ_BUFFERS; i++) {
unsigned int blk_offset;
if (buffer_dev[i] != sb)
continue;
if (blocknr < buffer_blocknr[i])
continue;
blk_offset = (blocknr - buffer_blocknr[i]) << PAGE_CACHE_SHIFT;
blk_offset += offset;
if (blk_offset + len > BUFFER_SIZE)
continue;
return read_buffers[i] + blk_offset;
}
devsize = mapping->host->i_size >> PAGE_CACHE_SHIFT;
major = MAJOR(sb->s_dev);
minor = MINOR(sb->s_dev);
if (blk_size[major])
devsize = blk_size[major][minor] >> 2;
/* Ok, read in BLKS_PER_BUF pages completely first. */
unread = 0;
for (i = 0; i < BLKS_PER_BUF; i++) {
struct page *page = NULL;
if (blocknr + i < devsize) {
page = read_cache_page(mapping, blocknr + i,
(filler_t *)mapping->a_ops->readpage,
NULL);
/* synchronous error? */
if (IS_ERR(page))
page = NULL;
}
pages[i] = page;
}
for (i = 0; i < BLKS_PER_BUF; i++) {
struct page *page = pages[i];
if (page) {
wait_on_page(page);
if (!Page_Uptodate(page)) {
/* asynchronous error */
page_cache_release(page);
pages[i] = NULL;
}
}
}
buffer = next_buffer;
next_buffer = NEXT_BUFFER(buffer);
buffer_blocknr[buffer] = blocknr;
buffer_dev[buffer] = sb;
data = read_buffers[buffer];
for (i = 0; i < BLKS_PER_BUF; i++) {
struct page *page = pages[i];
if (page) {
memcpy(data, kmap(page), PAGE_CACHE_SIZE);
kunmap(page);
page_cache_release(page);
} else
memset(data, 0, PAGE_CACHE_SIZE);
data += PAGE_CACHE_SIZE;
}
return read_buffers[buffer] + offset;
}
static int shrink_cache(int nr_pages, zone_t * classzone, unsigned int gfp_mask, int * failed_swapout)
{
struct list_head * entry;
int max_scan = (classzone->nr_inactive_pages + classzone->nr_active_pages) / vm_cache_scan_ratio;
int max_mapped = vm_mapped_ratio * nr_pages;
while (max_scan && classzone->nr_inactive_pages && (entry = inactive_list.prev) != &inactive_list) {
struct page * page;
if (unlikely(current->need_resched)) {
spin_unlock(&pagemap_lru_lock);
__set_current_state(TASK_RUNNING);
schedule();
spin_lock(&pagemap_lru_lock);
continue;
}
page = list_entry(entry, struct page, lru);
BUG_ON(!PageLRU(page));
BUG_ON(PageActive(page));
list_del(entry);
list_add(entry, &inactive_list);
/*
* Zero page counts can happen because we unlink the pages
* _after_ decrementing the usage count..
*/
if (unlikely(!page_count(page)))
continue;
if (!memclass(page_zone(page), classzone))
continue;
max_scan--;
/* Racy check to avoid trylocking when not worthwhile */
if (!page->buffers && (page_count(page) != 1 || !page->mapping))
goto page_mapped;
/*
* The page is locked. IO in progress?
* Move it to the back of the list.
*/
if (unlikely(TryLockPage(page))) {
if (PageLaunder(page) && (gfp_mask & __GFP_FS)) {
page_cache_get(page);
spin_unlock(&pagemap_lru_lock);
wait_on_page(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
}
continue;
}
if (PageDirty(page) && is_page_cache_freeable(page) && page->mapping) {
/*
* It is not critical here to write it only if
* the page is unmapped beause any direct writer
* like O_DIRECT would set the PG_dirty bitflag
* on the phisical page after having successfully
* pinned it and after the I/O to the page is finished,
* so the direct writes to the page cannot get lost.
*/
int (*writepage)(struct page *);
writepage = page->mapping->a_ops->writepage;
if ((gfp_mask & __GFP_FS) && writepage) {
ClearPageDirty(page);
SetPageLaunder(page);
page_cache_get(page);
spin_unlock(&pagemap_lru_lock);
writepage(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
continue;
}
}
/*
* If the page has buffers, try to free the buffer mappings
* associated with this page. If we succeed we try to free
* the page as well.
*/
if (page->buffers) {
spin_unlock(&pagemap_lru_lock);
/* avoid to free a locked page */
page_cache_get(page);
if (try_to_release_page(page, gfp_mask)) {
if (!page->mapping) {
/*
* We must not allow an anon page
* with no buffers to be visible on
* the LRU, so we unlock the page after
* taking the lru lock
*/
spin_lock(&pagemap_lru_lock);
UnlockPage(page);
__lru_cache_del(page);
/* effectively free the page here */
page_cache_release(page);
if (--nr_pages)
continue;
break;
} else {
/*
* The page is still in pagecache so undo the stuff
* before the try_to_release_page since we've not
* finished and we can now try the next step.
*/
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
}
} else {
/* failed to drop the buffers so stop here */
UnlockPage(page);
page_cache_release(page);
spin_lock(&pagemap_lru_lock);
continue;
}
}
spin_lock(&pagecache_lock);
/*
* This is the non-racy check for busy page.
* It is critical to check PageDirty _after_ we made sure
* the page is freeable so not in use by anybody.
* At this point we're guaranteed that page->buffers is NULL,
* nobody can refill page->buffers under us because we still
* hold the page lock.
*/
if (!page->mapping || page_count(page) > 1) {
spin_unlock(&pagecache_lock);
UnlockPage(page);
page_mapped:
if (--max_mapped < 0) {
spin_unlock(&pagemap_lru_lock);
nr_pages -= kmem_cache_reap(gfp_mask);
if (nr_pages <= 0)
goto out;
shrink_dcache_memory(vm_vfs_scan_ratio, gfp_mask);
shrink_icache_memory(vm_vfs_scan_ratio, gfp_mask);
#ifdef CONFIG_QUOTA
shrink_dqcache_memory(vm_vfs_scan_ratio, gfp_mask);
#endif
if (!*failed_swapout)
*failed_swapout = !swap_out(classzone);
max_mapped = nr_pages * vm_mapped_ratio;
spin_lock(&pagemap_lru_lock);
refill_inactive(nr_pages, classzone);
}
continue;
}
if (PageDirty(page)) {
spin_unlock(&pagecache_lock);
UnlockPage(page);
continue;
}
__lru_cache_del(page);
/* point of no return */
if (likely(!PageSwapCache(page))) {
__remove_inode_page(page);
spin_unlock(&pagecache_lock);
} else {
swp_entry_t swap;
swap.val = page->index;
__delete_from_swap_cache(page);
spin_unlock(&pagecache_lock);
swap_free(swap);
}
UnlockPage(page);
/* effectively free the page here */
page_cache_release(page);
if (--nr_pages)
continue;
break;
}
spin_unlock(&pagemap_lru_lock);
out:
return nr_pages;
}
static void rw_swap_page_base(int rw, unsigned long entry, struct page *page, int wait)
{
unsigned long type, offset;
struct swap_info_struct * p;
int zones[PAGE_SIZE/512];
int zones_used;
kdev_t dev = 0;
int block_size;
#ifdef DEBUG_SWAP
printk ("DebugVM: %s_swap_page entry %08lx, page %p (count %d), %s\n",
(rw == READ) ? "read" : "write",
entry, (char *) page_address(page), atomic_read(&page->count),
wait ? "wait" : "nowait");
#endif
type = SWP_TYPE(entry);
if (type >= nr_swapfiles) {
printk("Internal error: bad swap-device\n");
return;
}
/* Don't allow too many pending pages in flight.. */
if (atomic_read(&nr_async_pages) > pager_daemon.swap_cluster)
wait = 1;
p = &swap_info[type];
offset = SWP_OFFSET(entry);
if (offset >= p->max) {
printk("rw_swap_page: weirdness\n");
return;
}
if (p->swap_map && !p->swap_map[offset]) {
printk(KERN_ERR "rw_swap_page: "
"Trying to %s unallocated swap (%08lx)\n",
(rw == READ) ? "read" : "write", entry);
return;
}
if (!(p->flags & SWP_USED)) {
printk(KERN_ERR "rw_swap_page: "
"Trying to swap to unused swap-device\n");
return;
}
if (!PageLocked(page)) {
printk(KERN_ERR "VM: swap page is unlocked\n");
return;
}
if (PageSwapCache(page)) {
/* Make sure we are the only process doing I/O with this swap page. */
if (test_and_set_bit(offset, p->swap_lockmap))
{
struct wait_queue __wait;
__wait.task = current;
add_wait_queue(&lock_queue, &__wait);
for (;;) {
current->state = TASK_UNINTERRUPTIBLE;
mb();
if (!test_and_set_bit(offset, p->swap_lockmap))
break;
run_task_queue(&tq_disk);
schedule();
}
current->state = TASK_RUNNING;
remove_wait_queue(&lock_queue, &__wait);
}
/*
* Make sure that we have a swap cache association for this
* page. We need this to find which swap page to unlock once
* the swap IO has completed to the physical page. If the page
* is not already in the cache, just overload the offset entry
* as if it were: we are not allowed to manipulate the inode
* hashing for locked pages.
*/
if (page->offset != entry) {
printk ("swap entry mismatch");
return;
}
}
if (rw == READ) {
clear_bit(PG_uptodate, &page->flags);
kstat.pswpin++;
} else
kstat.pswpout++;
atomic_inc(&page->count);
if (p->swap_device) {
zones[0] = offset;
zones_used = 1;
dev = p->swap_device;
block_size = PAGE_SIZE;
} else if (p->swap_file) {
struct inode *swapf = p->swap_file->d_inode;
int i;
if (swapf->i_op->bmap == NULL
&& swapf->i_op->smap != NULL){
/*
With MS-DOS, we use msdos_smap which returns
a sector number (not a cluster or block number).
It is a patch to enable the UMSDOS project.
Other people are working on better solution.
It sounds like ll_rw_swap_file defined
its operation size (sector size) based on
PAGE_SIZE and the number of blocks to read.
So using bmap or smap should work even if
smap will require more blocks.
*/
int j;
unsigned int block = offset << 3;
for (i=0, j=0; j< PAGE_SIZE ; i++, j += 512){
if (!(zones[i] = swapf->i_op->smap(swapf,block++))) {
printk("rw_swap_page: bad swap file\n");
return;
}
}
block_size = 512;
}else{
int j;
unsigned int block = offset
<< (PAGE_SHIFT - swapf->i_sb->s_blocksize_bits);
block_size = swapf->i_sb->s_blocksize;
for (i=0, j=0; j< PAGE_SIZE ; i++, j += block_size)
if (!(zones[i] = bmap(swapf,block++))) {
printk("rw_swap_page: bad swap file\n");
return;
}
zones_used = i;
dev = swapf->i_dev;
}
} else {
printk(KERN_ERR "rw_swap_page: no swap file or device\n");
/* Do some cleaning up so if this ever happens we can hopefully
* trigger controlled shutdown.
*/
if (PageSwapCache(page)) {
if (!test_and_clear_bit(offset,p->swap_lockmap))
printk("swap_after_unlock_page: lock already cleared\n");
wake_up(&lock_queue);
}
atomic_dec(&page->count);
return;
}
if (!wait) {
set_bit(PG_decr_after, &page->flags);
atomic_inc(&nr_async_pages);
}
if (PageSwapCache(page)) {
/* only lock/unlock swap cache pages! */
set_bit(PG_swap_unlock_after, &page->flags);
}
set_bit(PG_free_after, &page->flags);
/* block_size == PAGE_SIZE/zones_used */
brw_page(rw, page, dev, zones, block_size, 0);
/* Note! For consistency we do all of the logic,
* decrementing the page count, and unlocking the page in the
* swap lock map - in the IO completion handler.
*/
if (!wait)
return;
wait_on_page(page);
/* This shouldn't happen, but check to be sure. */
if (atomic_read(&page->count) == 0)
printk(KERN_ERR "rw_swap_page: page unused while waiting!\n");
#ifdef DEBUG_SWAP
printk ("DebugVM: %s_swap_page finished on page %p (count %d)\n",
(rw == READ) ? "read" : "write",
(char *) page_adddress(page),
atomic_read(&page->count));
#endif
}
struct dentry *umsdos_solve_hlink (struct dentry *hlink)
{
/* root is our root for resolving pseudo-hardlink */
struct dentry *base = hlink->d_sb->s_root;
struct dentry *dentry_dst;
char *path, *pt;
int len;
struct address_space *mapping = hlink->d_inode->i_mapping;
struct page *page;
page=read_cache_page(mapping,0,(filler_t *)mapping->a_ops->readpage,NULL);
dentry_dst=(struct dentry *)page;
if (IS_ERR(page))
goto out;
wait_on_page(page);
if (!Page_Uptodate(page))
goto async_fail;
dentry_dst = ERR_PTR(-ENOMEM);
path = (char *) kmalloc (PATH_MAX, GFP_KERNEL);
if (path == NULL)
goto out_release;
memcpy(path, kmap(page), hlink->d_inode->i_size);
kunmap(page);
page_cache_release(page);
len = hlink->d_inode->i_size;
/* start at root dentry */
dentry_dst = dget(base);
path[len] = '\0';
pt = path;
if (*path == '/')
pt++; /* skip leading '/' */
if (base->d_inode == pseudo_root)
pt += (UMSDOS_PSDROOT_LEN + 1);
while (1) {
struct dentry *dir = dentry_dst, *demd;
char *start = pt;
int real;
while (*pt != '\0' && *pt != '/') pt++;
len = (int) (pt - start);
if (*pt == '/') *pt++ = '\0';
real = 1;
demd = umsdos_get_emd_dentry(dir);
if (!IS_ERR(demd)) {
if (demd->d_inode)
real = 0;
dput(demd);
}
#ifdef UMSDOS_DEBUG_VERBOSE
printk ("umsdos_solve_hlink: dir %s/%s, name=%s, real=%d\n",
dir->d_parent->d_name.name, dir->d_name.name, start, real);
#endif
dentry_dst = umsdos_lookup_dentry(dir, start, len, real);
/* XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX */
if (real)
d_drop(dir);
dput (dir);
if (IS_ERR(dentry_dst))
break;
/* not found? stop search ... */
if (!dentry_dst->d_inode) {
break;
}
if (*pt == '\0') /* we're finished! */
break;
} /* end while */
if (!IS_ERR(dentry_dst)) {
struct inode *inode = dentry_dst->d_inode;
if (inode) {
inode->u.umsdos_i.i_is_hlink = 1;
#ifdef UMSDOS_DEBUG_VERBOSE
printk ("umsdos_solve_hlink: resolved link %s/%s, ino=%ld\n",
dentry_dst->d_parent->d_name.name, dentry_dst->d_name.name, inode->i_ino);
#endif
} else {
#ifdef UMSDOS_DEBUG_VERBOSE
printk ("umsdos_solve_hlink: resolved link %s/%s negative!\n",
dentry_dst->d_parent->d_name.name, dentry_dst->d_name.name);
#endif
}
} else
printk(KERN_WARNING
"umsdos_solve_hlink: err=%ld\n", PTR_ERR(dentry_dst));
kfree (path);
out:
dput(hlink); /* original hlink no longer needed */
return dentry_dst;
async_fail:
dentry_dst = ERR_PTR(-EIO);
out_release:
page_cache_release(page);
goto out;
}
