#include <linux/bio.h>
#include <linux/fiemap.h>
#include <linux/namei.h>
+#include <trace/events/ext3.h>
#include "xattr.h"
#include "acl.h"
static int ext3_writepage_trans_blocks(struct inode *inode);
+static int ext3_block_truncate_page(struct inode *inode, loff_t from);
/*
* Test whether an inode is a fast symlink.
might_sleep();
+ trace_ext3_forget(inode, is_metadata, blocknr);
BUFFER_TRACE(bh, "enter");
jbd_debug(4, "forgetting bh %p: is_metadata = %d, mode %o, "
*/
void ext3_evict_inode (struct inode *inode)
{
+ struct ext3_inode_info *ei = EXT3_I(inode);
struct ext3_block_alloc_info *rsv;
handle_t *handle;
int want_delete = 0;
+ trace_ext3_evict_inode(inode);
if (!inode->i_nlink && !is_bad_inode(inode)) {
dquot_initialize(inode);
want_delete = 1;
}
+ /*
+ * When journalling data dirty buffers are tracked only in the journal.
+ * So although mm thinks everything is clean and ready for reaping the
+ * inode might still have some pages to write in the running
+ * transaction or waiting to be checkpointed. Thus calling
+ * journal_invalidatepage() (via truncate_inode_pages()) to discard
+ * these buffers can cause data loss. Also even if we did not discard
+ * these buffers, we would have no way to find them after the inode
+ * is reaped and thus user could see stale data if he tries to read
+ * them before the transaction is checkpointed. So be careful and
+ * force everything to disk here... We use ei->i_datasync_tid to
+ * store the newest transaction containing inode's data.
+ *
+ * Note that directories do not have this problem because they don't
+ * use page cache.
+ */
+ if (inode->i_nlink && ext3_should_journal_data(inode) &&
+ (S_ISLNK(inode->i_mode) || S_ISREG(inode->i_mode))) {
+ tid_t commit_tid = atomic_read(&ei->i_datasync_tid);
+ journal_t *journal = EXT3_SB(inode->i_sb)->s_journal;
+
+ log_start_commit(journal, commit_tid);
+ log_wait_commit(journal, commit_tid);
+ filemap_write_and_wait(&inode->i_data);
+ }
truncate_inode_pages(&inode->i_data, 0);
ext3_discard_reservation(inode);
- rsv = EXT3_I(inode)->i_block_alloc_info;
- EXT3_I(inode)->i_block_alloc_info = NULL;
+ rsv = ei->i_block_alloc_info;
+ ei->i_block_alloc_info = NULL;
if (unlikely(rsv))
kfree(rsv);
if (inode->i_blocks)
ext3_truncate(inode);
/*
- * Kill off the orphan record which ext3_truncate created.
- * AKPM: I think this can be inside the above `if'.
- * Note that ext3_orphan_del() has to be able to cope with the
- * deletion of a non-existent orphan - this is because we don't
- * know if ext3_truncate() actually created an orphan record.
- * (Well, we could do this if we need to, but heck - it works)
+ * Kill off the orphan record created when the inode lost the last
+ * link. Note that ext3_orphan_del() has to be able to cope with the
+ * deletion of a non-existent orphan - ext3_truncate() could
+ * have removed the record.
*/
ext3_orphan_del(handle, inode);
- EXT3_I(inode)->i_dtime = get_seconds();
+ ei->i_dtime = get_seconds();
/*
* One subtle ordering requirement: if anything has gone wrong
ext3_fsblk_t first_block = 0;
+ trace_ext3_get_blocks_enter(inode, iblock, maxblocks, create);
J_ASSERT(handle != NULL || create == 0);
depth = ext3_block_to_path(inode,iblock,offsets,&blocks_to_boundary);
if (!create || err == -EIO)
goto cleanup;
+ /*
+ * Block out ext3_truncate while we alter the tree
+ */
mutex_lock(&ei->truncate_mutex);
/*
*/
count = ext3_blks_to_allocate(partial, indirect_blks,
maxblocks, blocks_to_boundary);
- /*
- * Block out ext3_truncate while we alter the tree
- */
err = ext3_alloc_branch(handle, inode, indirect_blks, &count, goal,
offsets + (partial - chain), partial);
}
BUFFER_TRACE(bh_result, "returned");
out:
+ trace_ext3_get_blocks_exit(inode, iblock,
+ depth ? le32_to_cpu(chain[depth-1].key) : 0,
+ count, err);
return err;
}
ext3_truncate(inode);
}
+/*
+ * Truncate blocks that were not used by direct IO write. We have to zero out
+ * the last file block as well because direct IO might have written to it.
+ */
+static void ext3_truncate_failed_direct_write(struct inode *inode)
+{
+ ext3_block_truncate_page(inode, inode->i_size);
+ ext3_truncate(inode);
+}
+
static int ext3_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
* we allocate blocks but write fails for some reason */
int needed_blocks = ext3_writepage_trans_blocks(inode) + 1;
+ trace_ext3_write_begin(inode, pos, len, flags);
+
index = pos >> PAGE_CACHE_SHIFT;
from = pos & (PAGE_CACHE_SIZE - 1);
to = from + len;
unsigned from, to;
int ret = 0, ret2;
+ trace_ext3_ordered_write_end(inode, pos, len, copied);
copied = block_write_end(file, mapping, pos, len, copied, page, fsdata);
from = pos & (PAGE_CACHE_SIZE - 1);
struct inode *inode = file->f_mapping->host;
int ret;
+ trace_ext3_writeback_write_end(inode, pos, len, copied);
copied = block_write_end(file, mapping, pos, len, copied, page, fsdata);
update_file_sizes(inode, pos, copied);
/*
{
handle_t *handle = ext3_journal_current_handle();
struct inode *inode = mapping->host;
+ struct ext3_inode_info *ei = EXT3_I(inode);
int ret = 0, ret2;
int partial = 0;
unsigned from, to;
+ trace_ext3_journalled_write_end(inode, pos, len, copied);
from = pos & (PAGE_CACHE_SIZE - 1);
to = from + len;
if (pos + len > inode->i_size && ext3_can_truncate(inode))
ext3_orphan_add(handle, inode);
ext3_set_inode_state(inode, EXT3_STATE_JDATA);
- if (inode->i_size > EXT3_I(inode)->i_disksize) {
- EXT3_I(inode)->i_disksize = inode->i_size;
+ atomic_set(&ei->i_datasync_tid, handle->h_transaction->t_tid);
+ if (inode->i_size > ei->i_disksize) {
+ ei->i_disksize = inode->i_size;
ret2 = ext3_mark_inode_dirty(handle, inode);
if (!ret)
ret = ret2;
if (ext3_journal_current_handle())
goto out_fail;
+ trace_ext3_ordered_writepage(page);
if (!page_has_buffers(page)) {
create_empty_buffers(page, inode->i_sb->s_blocksize,
(1 << BH_Dirty)|(1 << BH_Uptodate));
if (ext3_journal_current_handle())
goto out_fail;
+ trace_ext3_writeback_writepage(page);
if (page_has_buffers(page)) {
if (!walk_page_buffers(NULL, page_buffers(page), 0,
PAGE_CACHE_SIZE, NULL, buffer_unmapped)) {
if (ext3_journal_current_handle())
goto no_write;
+ trace_ext3_journalled_writepage(page);
handle = ext3_journal_start(inode, ext3_writepage_trans_blocks(inode));
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
if (ret == 0)
ret = err;
ext3_set_inode_state(inode, EXT3_STATE_JDATA);
+ atomic_set(&EXT3_I(inode)->i_datasync_tid,
+ handle->h_transaction->t_tid);
unlock_page(page);
} else {
/*
static int ext3_readpage(struct file *file, struct page *page)
{
+ trace_ext3_readpage(page);
return mpage_readpage(page, ext3_get_block);
}
{
journal_t *journal = EXT3_JOURNAL(page->mapping->host);
+ trace_ext3_invalidatepage(page, offset);
+
/*
* If it's a full truncate we just forget about the pending dirtying
*/
{
journal_t *journal = EXT3_JOURNAL(page->mapping->host);
+ trace_ext3_releasepage(page);
WARN_ON(PageChecked(page));
if (!page_has_buffers(page))
return 0;
size_t count = iov_length(iov, nr_segs);
int retries = 0;
+ trace_ext3_direct_IO_enter(inode, offset, iov_length(iov, nr_segs), rw);
+
if (rw == WRITE) {
loff_t final_size = offset + count;
loff_t end = offset + iov_length(iov, nr_segs);
if (end > isize)
- vmtruncate(inode, isize);
+ ext3_truncate_failed_direct_write(inode);
}
if (ret == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries))
goto retry;
/* This is really bad luck. We've written the data
* but cannot extend i_size. Truncate allocated blocks
* and pretend the write failed... */
- ext3_truncate(inode);
+ ext3_truncate_failed_direct_write(inode);
ret = PTR_ERR(handle);
goto out;
}
ret = err;
}
out:
+ trace_ext3_direct_IO_exit(inode, offset,
+ iov_length(iov, nr_segs), rw, ret);
return ret;
}
* This required during truncate. We need to physically zero the tail end
* of that block so it doesn't yield old data if the file is later grown.
*/
-static int ext3_block_truncate_page(handle_t *handle, struct page *page,
- struct address_space *mapping, loff_t from)
+static int ext3_block_truncate_page(struct inode *inode, loff_t from)
{
ext3_fsblk_t index = from >> PAGE_CACHE_SHIFT;
- unsigned offset = from & (PAGE_CACHE_SIZE-1);
+ unsigned offset = from & (PAGE_CACHE_SIZE - 1);
unsigned blocksize, iblock, length, pos;
- struct inode *inode = mapping->host;
+ struct page *page;
+ handle_t *handle = NULL;
struct buffer_head *bh;
int err = 0;
+ /* Truncated on block boundary - nothing to do */
blocksize = inode->i_sb->s_blocksize;
+ if ((from & (blocksize - 1)) == 0)
+ return 0;
+
+ page = grab_cache_page(inode->i_mapping, index);
+ if (!page)
+ return -ENOMEM;
length = blocksize - (offset & (blocksize - 1));
iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
goto unlock;
}
+ /* data=writeback mode doesn't need transaction to zero-out data */
+ if (!ext3_should_writeback_data(inode)) {
+ /* We journal at most one block */
+ handle = ext3_journal_start(inode, 1);
+ if (IS_ERR(handle)) {
+ clear_highpage(page);
+ flush_dcache_page(page);
+ err = PTR_ERR(handle);
+ goto unlock;
+ }
+ }
+
if (ext3_should_journal_data(inode)) {
BUFFER_TRACE(bh, "get write access");
err = ext3_journal_get_write_access(handle, bh);
if (err)
- goto unlock;
+ goto stop;
}
zero_user(page, offset, length);
err = ext3_journal_dirty_data(handle, bh);
mark_buffer_dirty(bh);
}
+stop:
+ if (handle)
+ ext3_journal_stop(handle);
unlock:
unlock_page(page);
int ext3_can_truncate(struct inode *inode)
{
- if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
- return 0;
if (S_ISREG(inode->i_mode))
return 1;
if (S_ISDIR(inode->i_mode))
struct ext3_inode_info *ei = EXT3_I(inode);
__le32 *i_data = ei->i_data;
int addr_per_block = EXT3_ADDR_PER_BLOCK(inode->i_sb);
- struct address_space *mapping = inode->i_mapping;
int offsets[4];
Indirect chain[4];
Indirect *partial;
int n;
long last_block;
unsigned blocksize = inode->i_sb->s_blocksize;
- struct page *page;
+
+ trace_ext3_truncate_enter(inode);
if (!ext3_can_truncate(inode))
goto out_notrans;
if (inode->i_size == 0 && ext3_should_writeback_data(inode))
ext3_set_inode_state(inode, EXT3_STATE_FLUSH_ON_CLOSE);
- /*
- * We have to lock the EOF page here, because lock_page() nests
- * outside journal_start().
- */
- if ((inode->i_size & (blocksize - 1)) == 0) {
- /* Block boundary? Nothing to do */
- page = NULL;
- } else {
- page = grab_cache_page(mapping,
- inode->i_size >> PAGE_CACHE_SHIFT);
- if (!page)
- goto out_notrans;
- }
-
handle = start_transaction(inode);
- if (IS_ERR(handle)) {
- if (page) {
- clear_highpage(page);
- flush_dcache_page(page);
- unlock_page(page);
- page_cache_release(page);
- }
+ if (IS_ERR(handle))
goto out_notrans;
- }
last_block = (inode->i_size + blocksize-1)
>> EXT3_BLOCK_SIZE_BITS(inode->i_sb);
-
- if (page)
- ext3_block_truncate_page(handle, page, mapping, inode->i_size);
-
n = ext3_block_to_path(inode, last_block, offsets, NULL);
if (n == 0)
goto out_stop; /* error */
ext3_orphan_del(handle, inode);
ext3_journal_stop(handle);
+ trace_ext3_truncate_exit(inode);
return;
out_notrans:
/*
*/
if (inode->i_nlink)
ext3_orphan_del(NULL, inode);
+ trace_ext3_truncate_exit(inode);
}
static ext3_fsblk_t ext3_get_inode_block(struct super_block *sb,
* has in-inode xattrs, or we don't have this inode in memory.
* Read the block from disk.
*/
+ trace_ext3_load_inode(inode);
get_bh(bh);
bh->b_end_io = end_buffer_read_sync;
submit_bh(READ_META, bh);
}
error = ext3_orphan_add(handle, inode);
+ if (error) {
+ ext3_journal_stop(handle);
+ goto err_out;
+ }
EXT3_I(inode)->i_disksize = attr->ia_size;
- rc = ext3_mark_inode_dirty(handle, inode);
- if (!error)
- error = rc;
+ error = ext3_mark_inode_dirty(handle, inode);
ext3_journal_stop(handle);
+ if (error) {
+ /* Some hard fs error must have happened. Bail out. */
+ ext3_orphan_del(NULL, inode);
+ goto err_out;
+ }
+ rc = ext3_block_truncate_page(inode, attr->ia_size);
+ if (rc) {
+ /* Cleanup orphan list and exit */
+ handle = ext3_journal_start(inode, 3);
+ if (IS_ERR(handle)) {
+ ext3_orphan_del(NULL, inode);
+ goto err_out;
+ }
+ ext3_orphan_del(handle, inode);
+ ext3_journal_stop(handle);
+ goto err_out;
+ }
}
if ((attr->ia_valid & ATTR_SIZE) &&
attr->ia_size != i_size_read(inode)) {
- rc = vmtruncate(inode, attr->ia_size);
- if (rc)
- goto err_out;
+ truncate_setsize(inode, attr->ia_size);
+ ext3_truncate(inode);
}
setattr_copy(inode, attr);
int err;
might_sleep();
+ trace_ext3_mark_inode_dirty(inode, _RET_IP_);
err = ext3_reserve_inode_write(handle, inode, &iloc);
if (!err)
err = ext3_mark_iloc_dirty(handle, inode, &iloc);
git.openpandora.org / pandora-kernel.git / blobdiff