}
/*
- * Default synchronous end-of-IO handler.. Just mark it up-to-date and
- * unlock the buffer. This is what ll_rw_block uses too.
+ * End-of-IO handler helper function which does not touch the bh after
+ * unlocking it.
+ * Note: unlock_buffer() sort-of does touch the bh after unlocking it, but
+ * a race there is benign: unlock_buffer() only use the bh's address for
+ * hashing after unlocking the buffer, so it doesn't actually touch the bh
+ * itself.
*/
-void end_buffer_read_sync(struct buffer_head *bh, int uptodate)
+static void __end_buffer_read_notouch(struct buffer_head *bh, int uptodate)
{
if (uptodate) {
set_buffer_uptodate(bh);
clear_buffer_uptodate(bh);
}
unlock_buffer(bh);
+}
+
+/*
+ * Default synchronous end-of-IO handler.. Just mark it up-to-date and
+ * unlock the buffer. This is what ll_rw_block uses too.
+ */
+void end_buffer_read_sync(struct buffer_head *bh, int uptodate)
+{
+ __end_buffer_read_notouch(bh, uptodate);
put_bh(bh);
}
goto done;
}
+/*
+ * If a page has any new buffers, zero them out here, and mark them uptodate
+ * and dirty so they'll be written out (in order to prevent uninitialised
+ * block data from leaking). And clear the new bit.
+ */
+void page_zero_new_buffers(struct page *page, unsigned from, unsigned to)
+{
+ unsigned int block_start, block_end;
+ struct buffer_head *head, *bh;
+
+ BUG_ON(!PageLocked(page));
+ if (!page_has_buffers(page))
+ return;
+
+ bh = head = page_buffers(page);
+ block_start = 0;
+ do {
+ block_end = block_start + bh->b_size;
+
+ if (buffer_new(bh)) {
+ if (block_end > from && block_start < to) {
+ if (!PageUptodate(page)) {
+ unsigned start, size;
+
+ start = max(from, block_start);
+ size = min(to, block_end) - start;
+
+ zero_user_page(page, start, size, KM_USER0);
+ set_buffer_uptodate(bh);
+ }
+
+ clear_buffer_new(bh);
+ mark_buffer_dirty(bh);
+ }
+ }
+
+ block_start = block_end;
+ bh = bh->b_this_page;
+ } while (bh != head);
+}
+EXPORT_SYMBOL(page_zero_new_buffers);
+
static int __block_prepare_write(struct inode *inode, struct page *page,
unsigned from, unsigned to, get_block_t *get_block)
{
unmap_underlying_metadata(bh->b_bdev,
bh->b_blocknr);
if (PageUptodate(page)) {
+ clear_buffer_new(bh);
set_buffer_uptodate(bh);
+ mark_buffer_dirty(bh);
continue;
}
if (block_end > to || block_start < from) {
if (!buffer_uptodate(*wait_bh))
err = -EIO;
}
- if (!err) {
- bh = head;
- do {
- if (buffer_new(bh))
- clear_buffer_new(bh);
- } while ((bh = bh->b_this_page) != head);
- return 0;
- }
- /* Error case: */
- /*
- * Zero out any newly allocated blocks to avoid exposing stale
- * data. If BH_New is set, we know that the block was newly
- * allocated in the above loop.
- */
- bh = head;
- block_start = 0;
- do {
- block_end = block_start+blocksize;
- if (block_end <= from)
- goto next_bh;
- if (block_start >= to)
- break;
- if (buffer_new(bh)) {
- clear_buffer_new(bh);
- zero_user_page(page, block_start, bh->b_size, KM_USER0);
- set_buffer_uptodate(bh);
- mark_buffer_dirty(bh);
- }
-next_bh:
- block_start = block_end;
- bh = bh->b_this_page;
- } while (bh != head);
+ if (unlikely(err))
+ page_zero_new_buffers(page, from, to);
return err;
}
set_buffer_uptodate(bh);
mark_buffer_dirty(bh);
}
+ clear_buffer_new(bh);
}
/*
return 0;
}
+/*
+ * block_write_begin takes care of the basic task of block allocation and
+ * bringing partial write blocks uptodate first.
+ *
+ * If *pagep is not NULL, then block_write_begin uses the locked page
+ * at *pagep rather than allocating its own. In this case, the page will
+ * not be unlocked or deallocated on failure.
+ */
+int block_write_begin(struct file *file, struct address_space *mapping,
+ loff_t pos, unsigned len, unsigned flags,
+ struct page **pagep, void **fsdata,
+ get_block_t *get_block)
+{
+ struct inode *inode = mapping->host;
+ int status = 0;
+ struct page *page;
+ pgoff_t index;
+ unsigned start, end;
+ int ownpage = 0;
+
+ index = pos >> PAGE_CACHE_SHIFT;
+ start = pos & (PAGE_CACHE_SIZE - 1);
+ end = start + len;
+
+ page = *pagep;
+ if (page == NULL) {
+ ownpage = 1;
+ page = __grab_cache_page(mapping, index);
+ if (!page) {
+ status = -ENOMEM;
+ goto out;
+ }
+ *pagep = page;
+ } else
+ BUG_ON(!PageLocked(page));
+
+ status = __block_prepare_write(inode, page, start, end, get_block);
+ if (unlikely(status)) {
+ ClearPageUptodate(page);
+
+ if (ownpage) {
+ unlock_page(page);
+ page_cache_release(page);
+ *pagep = NULL;
+
+ /*
+ * prepare_write() may have instantiated a few blocks
+ * outside i_size. Trim these off again. Don't need
+ * i_size_read because we hold i_mutex.
+ */
+ if (pos + len > inode->i_size)
+ vmtruncate(inode, inode->i_size);
+ }
+ goto out;
+ }
+
+out:
+ return status;
+}
+EXPORT_SYMBOL(block_write_begin);
+
+int block_write_end(struct file *file, struct address_space *mapping,
+ loff_t pos, unsigned len, unsigned copied,
+ struct page *page, void *fsdata)
+{
+ struct inode *inode = mapping->host;
+ unsigned start;
+
+ start = pos & (PAGE_CACHE_SIZE - 1);
+
+ if (unlikely(copied < len)) {
+ /*
+ * The buffers that were written will now be uptodate, so we
+ * don't have to worry about a readpage reading them and
+ * overwriting a partial write. However if we have encountered
+ * a short write and only partially written into a buffer, it
+ * will not be marked uptodate, so a readpage might come in and
+ * destroy our partial write.
+ *
+ * Do the simplest thing, and just treat any short write to a
+ * non uptodate page as a zero-length write, and force the
+ * caller to redo the whole thing.
+ */
+ if (!PageUptodate(page))
+ copied = 0;
+
+ page_zero_new_buffers(page, start+copied, start+len);
+ }
+ flush_dcache_page(page);
+
+ /* This could be a short (even 0-length) commit */
+ __block_commit_write(inode, page, start, start+copied);
+
+ return copied;
+}
+EXPORT_SYMBOL(block_write_end);
+
+int generic_write_end(struct file *file, struct address_space *mapping,
+ loff_t pos, unsigned len, unsigned copied,
+ struct page *page, void *fsdata)
+{
+ struct inode *inode = mapping->host;
+
+ copied = block_write_end(file, mapping, pos, len, copied, page, fsdata);
+
+ /*
+ * No need to use i_size_read() here, the i_size
+ * cannot change under us because we hold i_mutex.
+ *
+ * But it's important to update i_size while still holding page lock:
+ * page writeout could otherwise come in and zero beyond i_size.
+ */
+ if (pos+copied > inode->i_size) {
+ i_size_write(inode, pos+copied);
+ mark_inode_dirty(inode);
+ }
+
+ unlock_page(page);
+ page_cache_release(page);
+
+ return copied;
+}
+EXPORT_SYMBOL(generic_write_end);
+
/*
* Generic "read page" function for block devices that have the normal
* get_block functionality. This is most of the block device filesystems.
return 0;
}
+/*
+ * block_page_mkwrite() is not allowed to change the file size as it gets
+ * called from a page fault handler when a page is first dirtied. Hence we must
+ * be careful to check for EOF conditions here. We set the page up correctly
+ * for a written page which means we get ENOSPC checking when writing into
+ * holes and correct delalloc and unwritten extent mapping on filesystems that
+ * support these features.
+ *
+ * We are not allowed to take the i_mutex here so we have to play games to
+ * protect against truncate races as the page could now be beyond EOF. Because
+ * vmtruncate() writes the inode size before removing pages, once we have the
+ * page lock we can determine safely if the page is beyond EOF. If it is not
+ * beyond EOF, then the page is guaranteed safe against truncation until we
+ * unlock the page.
+ */
+int
+block_page_mkwrite(struct vm_area_struct *vma, struct page *page,
+ get_block_t get_block)
+{
+ struct inode *inode = vma->vm_file->f_path.dentry->d_inode;
+ unsigned long end;
+ loff_t size;
+ int ret = -EINVAL;
+
+ lock_page(page);
+ size = i_size_read(inode);
+ if ((page->mapping != inode->i_mapping) ||
+ (page_offset(page) > size)) {
+ /* page got truncated out from underneath us */
+ goto out_unlock;
+ }
+
+ /* page is wholly or partially inside EOF */
+ if (((page->index + 1) << PAGE_CACHE_SHIFT) > size)
+ end = size & ~PAGE_CACHE_MASK;
+ else
+ end = PAGE_CACHE_SIZE;
+
+ ret = block_prepare_write(page, 0, end, get_block);
+ if (!ret)
+ ret = block_commit_write(page, 0, end);
+
+out_unlock:
+ unlock_page(page);
+ return ret;
+}
/*
* nobh_prepare_write()'s prereads are special: the buffer_heads are freed
* immediately, while under the page lock. So it needs a special end_io
* handler which does not touch the bh after unlocking it.
- *
- * Note: unlock_buffer() sort-of does touch the bh after unlocking it, but
- * a race there is benign: unlock_buffer() only use the bh's address for
- * hashing after unlocking the buffer, so it doesn't actually touch the bh
- * itself.
*/
static void end_buffer_read_nobh(struct buffer_head *bh, int uptodate)
{
- if (uptodate) {
- set_buffer_uptodate(bh);
- } else {
- /* This happens, due to failed READA attempts. */
- clear_buffer_uptodate(bh);
- }
- unlock_buffer(bh);
+ __end_buffer_read_notouch(bh, uptodate);
}
/*
struct inode *inode = page->mapping->host;
const unsigned blkbits = inode->i_blkbits;
const unsigned blocksize = 1 << blkbits;
- struct buffer_head map_bh;
- struct buffer_head *read_bh[MAX_BUF_PER_PAGE];
+ struct buffer_head *head, *bh;
unsigned block_in_page;
- unsigned block_start;
+ unsigned block_start, block_end;
sector_t block_in_file;
char *kaddr;
int nr_reads = 0;
- int i;
int ret = 0;
int is_mapped_to_disk = 1;
+ if (page_has_buffers(page))
+ return block_prepare_write(page, from, to, get_block);
+
if (PageMappedToDisk(page))
return 0;
+ /*
+ * Allocate buffers so that we can keep track of state, and potentially
+ * attach them to the page if an error occurs. In the common case of
+ * no error, they will just be freed again without ever being attached
+ * to the page (which is all OK, because we're under the page lock).
+ *
+ * Be careful: the buffer linked list is a NULL terminated one, rather
+ * than the circular one we're used to.
+ */
+ head = alloc_page_buffers(page, blocksize, 0);
+ if (!head)
+ return -ENOMEM;
+
block_in_file = (sector_t)page->index << (PAGE_CACHE_SHIFT - blkbits);
- map_bh.b_page = page;
/*
* We loop across all blocks in the page, whether or not they are
* part of the affected region. This is so we can discover if the
* page is fully mapped-to-disk.
*/
- for (block_start = 0, block_in_page = 0;
+ for (block_start = 0, block_in_page = 0, bh = head;
block_start < PAGE_CACHE_SIZE;
- block_in_page++, block_start += blocksize) {
- unsigned block_end = block_start + blocksize;
+ block_in_page++, block_start += blocksize, bh = bh->b_this_page) {
int create;
- map_bh.b_state = 0;
+ block_end = block_start + blocksize;
+ bh->b_state = 0;
create = 1;
if (block_start >= to)
create = 0;
- map_bh.b_size = blocksize;
ret = get_block(inode, block_in_file + block_in_page,
- &map_bh, create);
+ bh, create);
if (ret)
goto failed;
- if (!buffer_mapped(&map_bh))
+ if (!buffer_mapped(bh))
is_mapped_to_disk = 0;
- if (buffer_new(&map_bh))
- unmap_underlying_metadata(map_bh.b_bdev,
- map_bh.b_blocknr);
- if (PageUptodate(page))
+ if (buffer_new(bh))
+ unmap_underlying_metadata(bh->b_bdev, bh->b_blocknr);
+ if (PageUptodate(page)) {
+ set_buffer_uptodate(bh);
continue;
- if (buffer_new(&map_bh) || !buffer_mapped(&map_bh)) {
+ }
+ if (buffer_new(bh) || !buffer_mapped(bh)) {
kaddr = kmap_atomic(page, KM_USER0);
if (block_start < from)
memset(kaddr+block_start, 0, from-block_start);
kunmap_atomic(kaddr, KM_USER0);
continue;
}
- if (buffer_uptodate(&map_bh))
+ if (buffer_uptodate(bh))
continue; /* reiserfs does this */
if (block_start < from || block_end > to) {
- struct buffer_head *bh = alloc_buffer_head(GFP_NOFS);
-
- if (!bh) {
- ret = -ENOMEM;
- goto failed;
- }
- bh->b_state = map_bh.b_state;
- atomic_set(&bh->b_count, 0);
- bh->b_this_page = NULL;
- bh->b_page = page;
- bh->b_blocknr = map_bh.b_blocknr;
- bh->b_size = blocksize;
- bh->b_data = (char *)(long)block_start;
- bh->b_bdev = map_bh.b_bdev;
- bh->b_private = NULL;
- read_bh[nr_reads++] = bh;
+ lock_buffer(bh);
+ bh->b_end_io = end_buffer_read_nobh;
+ submit_bh(READ, bh);
+ nr_reads++;
}
}
if (nr_reads) {
- struct buffer_head *bh;
-
/*
* The page is locked, so these buffers are protected from
* any VM or truncate activity. Hence we don't need to care
* for the buffer_head refcounts.
*/
- for (i = 0; i < nr_reads; i++) {
- bh = read_bh[i];
- lock_buffer(bh);
- bh->b_end_io = end_buffer_read_nobh;
- submit_bh(READ, bh);
- }
- for (i = 0; i < nr_reads; i++) {
- bh = read_bh[i];
+ for (bh = head; bh; bh = bh->b_this_page) {
wait_on_buffer(bh);
if (!buffer_uptodate(bh))
ret = -EIO;
- free_buffer_head(bh);
- read_bh[i] = NULL;
}
if (ret)
goto failed;
if (is_mapped_to_disk)
SetPageMappedToDisk(page);
+ do {
+ bh = head;
+ head = head->b_this_page;
+ free_buffer_head(bh);
+ } while (head);
+
return 0;
failed:
- for (i = 0; i < nr_reads; i++) {
- if (read_bh[i])
- free_buffer_head(read_bh[i]);
- }
-
/*
- * Error recovery is pretty slack. Clear the page and mark it dirty
- * so we'll later zero out any blocks which _were_ allocated.
+ * Error recovery is a bit difficult. We need to zero out blocks that
+ * were newly allocated, and dirty them to ensure they get written out.
+ * Buffers need to be attached to the page at this point, otherwise
+ * the handling of potential IO errors during writeout would be hard
+ * (could try doing synchronous writeout, but what if that fails too?)
*/
- zero_user_page(page, 0, PAGE_CACHE_SIZE, KM_USER0);
- SetPageUptodate(page);
- set_page_dirty(page);
+ spin_lock(&page->mapping->private_lock);
+ bh = head;
+ block_start = 0;
+ do {
+ if (PageUptodate(page))
+ set_buffer_uptodate(bh);
+ if (PageDirty(page))
+ set_buffer_dirty(bh);
+
+ block_end = block_start+blocksize;
+ if (block_end <= from)
+ goto next;
+ if (block_start >= to)
+ goto next;
+
+ if (buffer_new(bh)) {
+ clear_buffer_new(bh);
+ if (!buffer_uptodate(bh)) {
+ zero_user_page(page, block_start, bh->b_size, KM_USER0);
+ set_buffer_uptodate(bh);
+ }
+ mark_buffer_dirty(bh);
+ }
+next:
+ block_start = block_end;
+ if (!bh->b_this_page)
+ bh->b_this_page = head;
+ bh = bh->b_this_page;
+ } while (bh != head);
+ attach_page_buffers(page, head);
+ spin_unlock(&page->mapping->private_lock);
+
return ret;
}
EXPORT_SYMBOL(nobh_prepare_write);
struct inode *inode = page->mapping->host;
loff_t pos = ((loff_t)page->index << PAGE_CACHE_SHIFT) + to;
+ if (page_has_buffers(page))
+ return generic_commit_write(file, page, from, to);
+
SetPageUptodate(page);
set_page_dirty(page);
if (pos > inode->i_size) {
return tmp.b_blocknr;
}
-static int end_bio_bh_io_sync(struct bio *bio, unsigned int bytes_done, int err)
+static void end_bio_bh_io_sync(struct bio *bio, int err)
{
struct buffer_head *bh = bio->bi_private;
- if (bio->bi_size)
- return 1;
-
if (err == -EOPNOTSUPP) {
set_bit(BIO_EOPNOTSUPP, &bio->bi_flags);
set_bit(BH_Eopnotsupp, &bh->b_state);
bh->b_end_io(bh, test_bit(BIO_UPTODATE, &bio->bi_flags));
bio_put(bio);
- return 0;
}
int submit_bh(int rw, struct buffer_head * bh)
EXPORT_SYMBOL(__wait_on_buffer);
EXPORT_SYMBOL(block_commit_write);
EXPORT_SYMBOL(block_prepare_write);
+EXPORT_SYMBOL(block_page_mkwrite);
EXPORT_SYMBOL(block_read_full_page);
EXPORT_SYMBOL(block_sync_page);
EXPORT_SYMBOL(block_truncate_page);