linux/fs/ext4/inode.c

4857 lines
140 KiB
C
Raw Normal View History

/*
* linux/fs/ext4/inode.c
*
* Copyright (C) 1992, 1993, 1994, 1995
* Remy Card (card@masi.ibp.fr)
* Laboratoire MASI - Institut Blaise Pascal
* Universite Pierre et Marie Curie (Paris VI)
*
* from
*
* linux/fs/minix/inode.c
*
* Copyright (C) 1991, 1992 Linus Torvalds
*
* 64-bit file support on 64-bit platforms by Jakub Jelinek
* (jj@sunsite.ms.mff.cuni.cz)
*
* Assorted race fixes, rewrite of ext4_get_block() by Al Viro, 2000
*/
#include <linux/fs.h>
#include <linux/time.h>
#include <linux/jbd2.h>
#include <linux/highuid.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
#include <linux/string.h>
#include <linux/buffer_head.h>
#include <linux/writeback.h>
#include <linux/pagevec.h>
#include <linux/mpage.h>
#include <linux/namei.h>
#include <linux/uio.h>
#include <linux/bio.h>
#include <linux/workqueue.h>
#include <linux/kernel.h>
#include <linux/printk.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/slab.h>
#include <linux/ratelimit.h>
#include "ext4_jbd2.h"
#include "xattr.h"
#include "acl.h"
#include "truncate.h"
#include <trace/events/ext4.h>
#define MPAGE_DA_EXTENT_TAIL 0x01
static __u32 ext4_inode_csum(struct inode *inode, struct ext4_inode *raw,
struct ext4_inode_info *ei)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
__u16 csum_lo;
__u16 csum_hi = 0;
__u32 csum;
csum_lo = raw->i_checksum_lo;
raw->i_checksum_lo = 0;
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE &&
EXT4_FITS_IN_INODE(raw, ei, i_checksum_hi)) {
csum_hi = raw->i_checksum_hi;
raw->i_checksum_hi = 0;
}
csum = ext4_chksum(sbi, ei->i_csum_seed, (__u8 *)raw,
EXT4_INODE_SIZE(inode->i_sb));
raw->i_checksum_lo = csum_lo;
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE &&
EXT4_FITS_IN_INODE(raw, ei, i_checksum_hi))
raw->i_checksum_hi = csum_hi;
return csum;
}
static int ext4_inode_csum_verify(struct inode *inode, struct ext4_inode *raw,
struct ext4_inode_info *ei)
{
__u32 provided, calculated;
if (EXT4_SB(inode->i_sb)->s_es->s_creator_os !=
cpu_to_le32(EXT4_OS_LINUX) ||
!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
return 1;
provided = le16_to_cpu(raw->i_checksum_lo);
calculated = ext4_inode_csum(inode, raw, ei);
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE &&
EXT4_FITS_IN_INODE(raw, ei, i_checksum_hi))
provided |= ((__u32)le16_to_cpu(raw->i_checksum_hi)) << 16;
else
calculated &= 0xFFFF;
return provided == calculated;
}
static void ext4_inode_csum_set(struct inode *inode, struct ext4_inode *raw,
struct ext4_inode_info *ei)
{
__u32 csum;
if (EXT4_SB(inode->i_sb)->s_es->s_creator_os !=
cpu_to_le32(EXT4_OS_LINUX) ||
!EXT4_HAS_RO_COMPAT_FEATURE(inode->i_sb,
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
return;
csum = ext4_inode_csum(inode, raw, ei);
raw->i_checksum_lo = cpu_to_le16(csum & 0xFFFF);
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE &&
EXT4_FITS_IN_INODE(raw, ei, i_checksum_hi))
raw->i_checksum_hi = cpu_to_le16(csum >> 16);
}
static inline int ext4_begin_ordered_truncate(struct inode *inode,
loff_t new_size)
{
trace_ext4_begin_ordered_truncate(inode, new_size);
/*
* If jinode is zero, then we never opened the file for
* writing, so there's no need to call
* jbd2_journal_begin_ordered_truncate() since there's no
* outstanding writes we need to flush.
*/
if (!EXT4_I(inode)->jinode)
return 0;
return jbd2_journal_begin_ordered_truncate(EXT4_JOURNAL(inode),
EXT4_I(inode)->jinode,
new_size);
}
static void ext4_invalidatepage(struct page *page, unsigned long offset);
static int noalloc_get_block_write(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create);
static int ext4_set_bh_endio(struct buffer_head *bh, struct inode *inode);
static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate);
static int __ext4_journalled_writepage(struct page *page, unsigned int len);
static int ext4_bh_delay_or_unwritten(handle_t *handle, struct buffer_head *bh);
static int ext4_discard_partial_page_buffers_no_lock(handle_t *handle,
struct inode *inode, struct page *page, loff_t from,
loff_t length, int flags);
/*
* Test whether an inode is a fast symlink.
*/
static int ext4_inode_is_fast_symlink(struct inode *inode)
{
int ea_blocks = EXT4_I(inode)->i_file_acl ?
(inode->i_sb->s_blocksize >> 9) : 0;
return (S_ISLNK(inode->i_mode) && inode->i_blocks - ea_blocks == 0);
}
/*
* Restart the transaction associated with *handle. This does a commit,
* so before we call here everything must be consistently dirtied against
* this transaction.
*/
int ext4_truncate_restart_trans(handle_t *handle, struct inode *inode,
int nblocks)
{
int ret;
/*
* Drop i_data_sem to avoid deadlock with ext4_map_blocks. At this
* moment, get_block can be called only for blocks inside i_size since
* page cache has been already dropped and writes are blocked by
* i_mutex. So we can safely drop the i_data_sem here.
*/
BUG_ON(EXT4_JOURNAL(inode) == NULL);
jbd_debug(2, "restarting handle %p\n", handle);
up_write(&EXT4_I(inode)->i_data_sem);
ret = ext4_journal_restart(handle, nblocks);
down_write(&EXT4_I(inode)->i_data_sem);
ext4_discard_preallocations(inode);
return ret;
}
/*
* Called at the last iput() if i_nlink is zero.
*/
void ext4_evict_inode(struct inode *inode)
{
handle_t *handle;
int err;
trace_ext4_evict_inode(inode);
ext4: call ext4_ioend_wait and ext4_flush_completed_IO in ext4_evict_inode Flush inode's i_completed_io_list before calling ext4_io_wait to prevent the following deadlock scenario: A page fault happens while some process is writing inode A. During page fault, shrink_icache_memory is called that in turn evicts another inode B. Inode B has some pending io_end work so it calls ext4_ioend_wait() that waits for inode B's i_ioend_count to become zero. However, inode B's ioend work was queued behind some of inode A's ioend work on the same cpu's ext4-dio-unwritten workqueue. As the ext4-dio-unwritten thread on that cpu is processing inode A's ioend work, it tries to grab inode A's i_mutex lock. Since the i_mutex lock of inode A is still hold before the page fault happened, we enter a deadlock. Also moves ext4_flush_completed_IO and ext4_ioend_wait from ext4_destroy_inode() to ext4_evict_inode(). During inode deleteion, ext4_evict_inode() is called before ext4_destroy_inode() and in ext4_evict_inode(), we may call ext4_truncate() without holding i_mutex lock. As a result, there is a race between flush_completed_IO that is called from ext4_ext_truncate() and ext4_end_io_work, which may cause corruption on an io_end structure. This change moves ext4_flush_completed_IO and ext4_ioend_wait from ext4_destroy_inode() to ext4_evict_inode() to resolve the race between ext4_truncate() and ext4_end_io_work during inode deletion. Signed-off-by: Jiaying Zhang <jiayingz@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@kernel.org
2011-08-13 16:17:13 +00:00
ext4_ioend_wait(inode);
if (inode->i_nlink) {
/*
* 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 jbd2_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 (ext4_should_journal_data(inode) &&
(S_ISLNK(inode->i_mode) || S_ISREG(inode->i_mode))) {
journal_t *journal = EXT4_SB(inode->i_sb)->s_journal;
tid_t commit_tid = EXT4_I(inode)->i_datasync_tid;
jbd2_log_start_commit(journal, commit_tid);
jbd2_log_wait_commit(journal, commit_tid);
filemap_write_and_wait(&inode->i_data);
}
truncate_inode_pages(&inode->i_data, 0);
goto no_delete;
}
if (!is_bad_inode(inode))
dquot_initialize(inode);
if (ext4_should_order_data(inode))
ext4_begin_ordered_truncate(inode, 0);
truncate_inode_pages(&inode->i_data, 0);
if (is_bad_inode(inode))
goto no_delete;
/*
* Protect us against freezing - iput() caller didn't have to have any
* protection against it
*/
sb_start_intwrite(inode->i_sb);
handle = ext4_journal_start(inode, ext4_blocks_for_truncate(inode)+3);
if (IS_ERR(handle)) {
ext4_std_error(inode->i_sb, PTR_ERR(handle));
/*
* If we're going to skip the normal cleanup, we still need to
* make sure that the in-core orphan linked list is properly
* cleaned up.
*/
ext4_orphan_del(NULL, inode);
sb_end_intwrite(inode->i_sb);
goto no_delete;
}
if (IS_SYNC(inode))
ext4_handle_sync(handle);
inode->i_size = 0;
err = ext4_mark_inode_dirty(handle, inode);
if (err) {
ext4_warning(inode->i_sb,
"couldn't mark inode dirty (err %d)", err);
goto stop_handle;
}
if (inode->i_blocks)
ext4_truncate(inode);
/*
* ext4_ext_truncate() doesn't reserve any slop when it
* restarts journal transactions; therefore there may not be
* enough credits left in the handle to remove the inode from
* the orphan list and set the dtime field.
*/
if (!ext4_handle_has_enough_credits(handle, 3)) {
err = ext4_journal_extend(handle, 3);
if (err > 0)
err = ext4_journal_restart(handle, 3);
if (err != 0) {
ext4_warning(inode->i_sb,
"couldn't extend journal (err %d)", err);
stop_handle:
ext4_journal_stop(handle);
ext4_orphan_del(NULL, inode);
sb_end_intwrite(inode->i_sb);
goto no_delete;
}
}
/*
* Kill off the orphan record which ext4_truncate created.
* AKPM: I think this can be inside the above `if'.
* Note that ext4_orphan_del() has to be able to cope with the
* deletion of a non-existent orphan - this is because we don't
* know if ext4_truncate() actually created an orphan record.
* (Well, we could do this if we need to, but heck - it works)
*/
ext4_orphan_del(handle, inode);
EXT4_I(inode)->i_dtime = get_seconds();
/*
* One subtle ordering requirement: if anything has gone wrong
* (transaction abort, IO errors, whatever), then we can still
* do these next steps (the fs will already have been marked as
* having errors), but we can't free the inode if the mark_dirty
* fails.
*/
if (ext4_mark_inode_dirty(handle, inode))
/* If that failed, just do the required in-core inode clear. */
ext4_clear_inode(inode);
else
ext4_free_inode(handle, inode);
ext4_journal_stop(handle);
sb_end_intwrite(inode->i_sb);
return;
no_delete:
ext4_clear_inode(inode); /* We must guarantee clearing of inode... */
}
#ifdef CONFIG_QUOTA
qsize_t *ext4_get_reserved_space(struct inode *inode)
{
return &EXT4_I(inode)->i_reserved_quota;
}
#endif
/*
* Calculate the number of metadata blocks need to reserve
* to allocate a block located at @lblock
*/
static int ext4_calc_metadata_amount(struct inode *inode, ext4_lblk_t lblock)
{
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
return ext4_ext_calc_metadata_amount(inode, lblock);
return ext4_ind_calc_metadata_amount(inode, lblock);
}
/*
* Called with i_data_sem down, which is important since we can call
* ext4_discard_preallocations() from here.
*/
void ext4_da_update_reserve_space(struct inode *inode,
int used, int quota_claim)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
struct ext4_inode_info *ei = EXT4_I(inode);
spin_lock(&ei->i_block_reservation_lock);
trace_ext4_da_update_reserve_space(inode, used, quota_claim);
if (unlikely(used > ei->i_reserved_data_blocks)) {
ext4_msg(inode->i_sb, KERN_NOTICE, "%s: ino %lu, used %d "
"with only %d reserved data blocks",
__func__, inode->i_ino, used,
ei->i_reserved_data_blocks);
WARN_ON(1);
used = ei->i_reserved_data_blocks;
}
if (unlikely(ei->i_allocated_meta_blocks > ei->i_reserved_meta_blocks)) {
ext4_msg(inode->i_sb, KERN_NOTICE, "%s: ino %lu, allocated %d "
"with only %d reserved metadata blocks\n", __func__,
inode->i_ino, ei->i_allocated_meta_blocks,
ei->i_reserved_meta_blocks);
WARN_ON(1);
ei->i_allocated_meta_blocks = ei->i_reserved_meta_blocks;
}
/* Update per-inode reservations */
ei->i_reserved_data_blocks -= used;
ei->i_reserved_meta_blocks -= ei->i_allocated_meta_blocks;
percpu_counter_sub(&sbi->s_dirtyclusters_counter,
used + ei->i_allocated_meta_blocks);
ei->i_allocated_meta_blocks = 0;
if (ei->i_reserved_data_blocks == 0) {
/*
* We can release all of the reserved metadata blocks
* only when we have written all of the delayed
* allocation blocks.
*/
percpu_counter_sub(&sbi->s_dirtyclusters_counter,
ei->i_reserved_meta_blocks);
ei->i_reserved_meta_blocks = 0;
ei->i_da_metadata_calc_len = 0;
}
spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
/* Update quota subsystem for data blocks */
if (quota_claim)
dquot_claim_block(inode, EXT4_C2B(sbi, used));
else {
/*
* We did fallocate with an offset that is already delayed
* allocated. So on delayed allocated writeback we should
* not re-claim the quota for fallocated blocks.
*/
dquot_release_reservation_block(inode, EXT4_C2B(sbi, used));
}
/*
* If we have done all the pending block allocations and if
* there aren't any writers on the inode, we can discard the
* inode's preallocations.
*/
if ((ei->i_reserved_data_blocks == 0) &&
(atomic_read(&inode->i_writecount) == 0))
ext4_discard_preallocations(inode);
}
static int __check_block_validity(struct inode *inode, const char *func,
unsigned int line,
struct ext4_map_blocks *map)
{
if (!ext4_data_block_valid(EXT4_SB(inode->i_sb), map->m_pblk,
map->m_len)) {
ext4_error_inode(inode, func, line, map->m_pblk,
"lblock %lu mapped to illegal pblock "
"(length %d)", (unsigned long) map->m_lblk,
map->m_len);
return -EIO;
}
return 0;
}
#define check_block_validity(inode, map) \
__check_block_validity((inode), __func__, __LINE__, (map))
/*
* Return the number of contiguous dirty pages in a given inode
* starting at page frame idx.
*/
static pgoff_t ext4_num_dirty_pages(struct inode *inode, pgoff_t idx,
unsigned int max_pages)
{
struct address_space *mapping = inode->i_mapping;
pgoff_t index;
struct pagevec pvec;
pgoff_t num = 0;
int i, nr_pages, done = 0;
if (max_pages == 0)
return 0;
pagevec_init(&pvec, 0);
while (!done) {
index = idx;
nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
PAGECACHE_TAG_DIRTY,
(pgoff_t)PAGEVEC_SIZE);
if (nr_pages == 0)
break;
for (i = 0; i < nr_pages; i++) {
struct page *page = pvec.pages[i];
struct buffer_head *bh, *head;
lock_page(page);
if (unlikely(page->mapping != mapping) ||
!PageDirty(page) ||
PageWriteback(page) ||
page->index != idx) {
done = 1;
unlock_page(page);
break;
}
if (page_has_buffers(page)) {
bh = head = page_buffers(page);
do {
if (!buffer_delay(bh) &&
!buffer_unwritten(bh))
done = 1;
bh = bh->b_this_page;
} while (!done && (bh != head));
}
unlock_page(page);
if (done)
break;
idx++;
num++;
if (num >= max_pages) {
done = 1;
break;
}
}
pagevec_release(&pvec);
}
return num;
}
/*
* The ext4_map_blocks() function tries to look up the requested blocks,
* and returns if the blocks are already mapped.
*
* Otherwise it takes the write lock of the i_data_sem and allocate blocks
* and store the allocated blocks in the result buffer head and mark it
* mapped.
*
* If file type is extents based, it will call ext4_ext_map_blocks(),
* Otherwise, call with ext4_ind_map_blocks() to handle indirect mapping
* based files
*
* On success, it returns the number of blocks being mapped or allocate.
* if create==0 and the blocks are pre-allocated and uninitialized block,
* the result buffer head is unmapped. If the create ==1, it will make sure
* the buffer head is mapped.
*
* It returns 0 if plain look up failed (blocks have not been allocated), in
* that case, buffer head is unmapped
*
* It returns the error in case of allocation failure.
*/
int ext4_map_blocks(handle_t *handle, struct inode *inode,
struct ext4_map_blocks *map, int flags)
{
int retval;
map->m_flags = 0;
ext_debug("ext4_map_blocks(): inode %lu, flag %d, max_blocks %u,"
"logical block %lu\n", inode->i_ino, flags, map->m_len,
(unsigned long) map->m_lblk);
/*
* Try to see if we can get the block without requesting a new
* file system block.
*/
if (!(flags & EXT4_GET_BLOCKS_NO_LOCK))
down_read((&EXT4_I(inode)->i_data_sem));
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
retval = ext4_ext_map_blocks(handle, inode, map, flags &
EXT4_GET_BLOCKS_KEEP_SIZE);
} else {
retval = ext4_ind_map_blocks(handle, inode, map, flags &
EXT4_GET_BLOCKS_KEEP_SIZE);
}
if (!(flags & EXT4_GET_BLOCKS_NO_LOCK))
up_read((&EXT4_I(inode)->i_data_sem));
if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
int ret;
if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
/* delayed alloc may be allocated by fallocate and
* coverted to initialized by directIO.
* we need to handle delayed extent here.
*/
down_write((&EXT4_I(inode)->i_data_sem));
goto delayed_mapped;
}
ret = check_block_validity(inode, map);
if (ret != 0)
return ret;
}
/* If it is only a block(s) look up */
if ((flags & EXT4_GET_BLOCKS_CREATE) == 0)
return retval;
/*
* Returns if the blocks have already allocated
*
* Note that if blocks have been preallocated
* ext4_ext_get_block() returns the create = 0
* with buffer head unmapped.
*/
if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED)
return retval;
/*
* When we call get_blocks without the create flag, the
* BH_Unwritten flag could have gotten set if the blocks
* requested were part of a uninitialized extent. We need to
* clear this flag now that we are committed to convert all or
* part of the uninitialized extent to be an initialized
* extent. This is because we need to avoid the combination
* of BH_Unwritten and BH_Mapped flags being simultaneously
* set on the buffer_head.
*/
map->m_flags &= ~EXT4_MAP_UNWRITTEN;
/*
* New blocks allocate and/or writing to uninitialized extent
* will possibly result in updating i_data, so we take
* the write lock of i_data_sem, and call get_blocks()
* with create == 1 flag.
*/
down_write((&EXT4_I(inode)->i_data_sem));
/*
* if the caller is from delayed allocation writeout path
* we have already reserved fs blocks for allocation
* let the underlying get_block() function know to
* avoid double accounting
*/
if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE)
ext4_set_inode_state(inode, EXT4_STATE_DELALLOC_RESERVED);
/*
* We need to check for EXT4 here because migrate
* could have changed the inode type in between
*/
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
retval = ext4_ext_map_blocks(handle, inode, map, flags);
} else {
retval = ext4_ind_map_blocks(handle, inode, map, flags);
if (retval > 0 && map->m_flags & EXT4_MAP_NEW) {
/*
* We allocated new blocks which will result in
* i_data's format changing. Force the migrate
* to fail by clearing migrate flags
*/
ext4_clear_inode_state(inode, EXT4_STATE_EXT_MIGRATE);
}
/*
* Update reserved blocks/metadata blocks after successful
* block allocation which had been deferred till now. We don't
* support fallocate for non extent files. So we can update
* reserve space here.
*/
if ((retval > 0) &&
(flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE))
ext4_da_update_reserve_space(inode, retval, 1);
}
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-09 23:20:51 +00:00
if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE) {
ext4_clear_inode_state(inode, EXT4_STATE_DELALLOC_RESERVED);
if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
int ret;
delayed_mapped:
/* delayed allocation blocks has been allocated */
ret = ext4_es_remove_extent(inode, map->m_lblk,
map->m_len);
if (ret < 0)
retval = ret;
}
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-09 23:20:51 +00:00
}
up_write((&EXT4_I(inode)->i_data_sem));
if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) {
int ret = check_block_validity(inode, map);
if (ret != 0)
return ret;
}
return retval;
}
/* Maximum number of blocks we map for direct IO at once. */
#define DIO_MAX_BLOCKS 4096
static int _ext4_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh, int flags)
{
handle_t *handle = ext4_journal_current_handle();
struct ext4_map_blocks map;
int ret = 0, started = 0;
int dio_credits;
if (ext4_has_inline_data(inode))
return -ERANGE;
map.m_lblk = iblock;
map.m_len = bh->b_size >> inode->i_blkbits;
if (flags && !(flags & EXT4_GET_BLOCKS_NO_LOCK) && !handle) {
/* Direct IO write... */
if (map.m_len > DIO_MAX_BLOCKS)
map.m_len = DIO_MAX_BLOCKS;
dio_credits = ext4_chunk_trans_blocks(inode, map.m_len);
handle = ext4_journal_start(inode, dio_credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
return ret;
}
started = 1;
}
ret = ext4_map_blocks(handle, inode, &map, flags);
if (ret > 0) {
map_bh(bh, inode->i_sb, map.m_pblk);
bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags;
bh->b_size = inode->i_sb->s_blocksize * map.m_len;
ret = 0;
}
if (started)
ext4_journal_stop(handle);
return ret;
}
int ext4_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh, int create)
{
return _ext4_get_block(inode, iblock, bh,
create ? EXT4_GET_BLOCKS_CREATE : 0);
}
/*
* `handle' can be NULL if create is zero
*/
struct buffer_head *ext4_getblk(handle_t *handle, struct inode *inode,
ext4_lblk_t block, int create, int *errp)
{
struct ext4_map_blocks map;
struct buffer_head *bh;
int fatal = 0, err;
J_ASSERT(handle != NULL || create == 0);
map.m_lblk = block;
map.m_len = 1;
err = ext4_map_blocks(handle, inode, &map,
create ? EXT4_GET_BLOCKS_CREATE : 0);
ext4: fix possible non-initialized variable in htree_dirblock_to_tree() htree_dirblock_to_tree() declares a non-initialized 'err' variable, which is passed as a reference to another functions expecting them to set this variable with their error codes. It's passed to ext4_bread(), which then passes it to ext4_getblk(). If ext4_map_blocks() returns 0 due to a lookup failure, leaving the ext4_getblk() buffer_head uninitialized, it will make ext4_getblk() return to ext4_bread() without initialize the 'err' variable, and ext4_bread() will return to htree_dirblock_to_tree() with this variable still uninitialized. htree_dirblock_to_tree() will pass this variable with garbage back to ext4_htree_fill_tree(), which expects a number of directory entries added to the rb-tree. which, in case, might return a fake non-zero value due the garbage left in the 'err' variable, leading the kernel to an Oops in ext4_dx_readdir(), once this is expecting a filled rb-tree node, when in turn it will have a NULL-ed one, causing an invalid page request when trying to get a fname struct from this NULL-ed rb-tree node in this line: fname = rb_entry(info->curr_node, struct fname, rb_hash); The patch itself initializes the err variable in htree_dirblock_to_tree() to avoid usage mistakes by the called functions, and also fix ext4_getblk() to return a initialized 'err' variable when ext4_map_blocks() fails a lookup. Signed-off-by: Carlos Maiolino <cmaiolino@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-09-18 03:39:12 +00:00
/* ensure we send some value back into *errp */
*errp = 0;
if (err < 0)
*errp = err;
if (err <= 0)
return NULL;
bh = sb_getblk(inode->i_sb, map.m_pblk);
if (!bh) {
*errp = -EIO;
return NULL;
}
if (map.m_flags & EXT4_MAP_NEW) {
J_ASSERT(create != 0);
J_ASSERT(handle != NULL);
/*
* Now that we do not always journal data, we should
* keep in mind whether this should always journal the
* new buffer as metadata. For now, regular file
* writes use ext4_get_block instead, so it's not a
* problem.
*/
lock_buffer(bh);
BUFFER_TRACE(bh, "call get_create_access");
fatal = ext4_journal_get_create_access(handle, bh);
if (!fatal && !buffer_uptodate(bh)) {
memset(bh->b_data, 0, inode->i_sb->s_blocksize);
set_buffer_uptodate(bh);
}
unlock_buffer(bh);
BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
err = ext4_handle_dirty_metadata(handle, inode, bh);
if (!fatal)
fatal = err;
} else {
BUFFER_TRACE(bh, "not a new buffer");
}
if (fatal) {
*errp = fatal;
brelse(bh);
bh = NULL;
}
return bh;
}
struct buffer_head *ext4_bread(handle_t *handle, struct inode *inode,
ext4_lblk_t block, int create, int *err)
{
struct buffer_head *bh;
bh = ext4_getblk(handle, inode, block, create, err);
if (!bh)
return bh;
if (buffer_uptodate(bh))
return bh;
ll_rw_block(READ | REQ_META | REQ_PRIO, 1, &bh);
wait_on_buffer(bh);
if (buffer_uptodate(bh))
return bh;
put_bh(bh);
*err = -EIO;
return NULL;
}
int ext4_walk_page_buffers(handle_t *handle,
struct buffer_head *head,
unsigned from,
unsigned to,
int *partial,
int (*fn)(handle_t *handle,
struct buffer_head *bh))
{
struct buffer_head *bh;
unsigned block_start, block_end;
unsigned blocksize = head->b_size;
int err, ret = 0;
struct buffer_head *next;
for (bh = head, block_start = 0;
ret == 0 && (bh != head || !block_start);
block_start = block_end, bh = next) {
next = bh->b_this_page;
block_end = block_start + blocksize;
if (block_end <= from || block_start >= to) {
if (partial && !buffer_uptodate(bh))
*partial = 1;
continue;
}
err = (*fn)(handle, bh);
if (!ret)
ret = err;
}
return ret;
}
/*
* To preserve ordering, it is essential that the hole instantiation and
* the data write be encapsulated in a single transaction. We cannot
* close off a transaction and start a new one between the ext4_get_block()
* and the commit_write(). So doing the jbd2_journal_start at the start of
* prepare_write() is the right place.
*
* Also, this function can nest inside ext4_writepage() ->
* block_write_full_page(). In that case, we *know* that ext4_writepage()
* has generated enough buffer credits to do the whole page. So we won't
* block on the journal in that case, which is good, because the caller may
* be PF_MEMALLOC.
*
* By accident, ext4 can be reentered when a transaction is open via
* quota file writes. If we were to commit the transaction while thus
* reentered, there can be a deadlock - we would be holding a quota
* lock, and the commit would never complete if another thread had a
* transaction open and was blocking on the quota lock - a ranking
* violation.
*
* So what we do is to rely on the fact that jbd2_journal_stop/journal_start
* will _not_ run commit under these circumstances because handle->h_ref
* is elevated. We'll still have enough credits for the tiny quotafile
* write.
*/
int do_journal_get_write_access(handle_t *handle,
struct buffer_head *bh)
{
int dirty = buffer_dirty(bh);
int ret;
if (!buffer_mapped(bh) || buffer_freed(bh))
return 0;
/*
* __block_write_begin() could have dirtied some buffers. Clean
* the dirty bit as jbd2_journal_get_write_access() could complain
* otherwise about fs integrity issues. Setting of the dirty bit
* by __block_write_begin() isn't a real problem here as we clear
* the bit before releasing a page lock and thus writeback cannot
* ever write the buffer.
*/
if (dirty)
clear_buffer_dirty(bh);
ret = ext4_journal_get_write_access(handle, bh);
if (!ret && dirty)
ret = ext4_handle_dirty_metadata(handle, NULL, bh);
return ret;
}
static int ext4_get_block_write_nolock(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create);
static int ext4_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
{
struct inode *inode = mapping->host;
int ret, needed_blocks;
handle_t *handle;
int retries = 0;
struct page *page;
pgoff_t index;
unsigned from, to;
trace_ext4_write_begin(inode, pos, len, flags);
/*
* Reserve one block more for addition to orphan list in case
* we allocate blocks but write fails for some reason
*/
needed_blocks = ext4_writepage_trans_blocks(inode) + 1;
index = pos >> PAGE_CACHE_SHIFT;
from = pos & (PAGE_CACHE_SIZE - 1);
to = from + len;
if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) {
ret = ext4_try_to_write_inline_data(mapping, inode, pos, len,
flags, pagep);
if (ret < 0)
goto out;
if (ret == 1) {
ret = 0;
goto out;
}
}
retry:
handle = ext4_journal_start(inode, needed_blocks);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
goto out;
}
/* We cannot recurse into the filesystem as the transaction is already
* started */
flags |= AOP_FLAG_NOFS;
fs: symlink write_begin allocation context fix With the write_begin/write_end aops, page_symlink was broken because it could no longer pass a GFP_NOFS type mask into the point where the allocations happened. They are done in write_begin, which would always assume that the filesystem can be entered from reclaim. This bug could cause filesystem deadlocks. The funny thing with having a gfp_t mask there is that it doesn't really allow the caller to arbitrarily tinker with the context in which it can be called. It couldn't ever be GFP_ATOMIC, for example, because it needs to take the page lock. The only thing any callers care about is __GFP_FS anyway, so turn that into a single flag. Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on this flag in their write_begin function. Change __grab_cache_page to accept a nofs argument as well, to honour that flag (while we're there, change the name to grab_cache_page_write_begin which is more instructive and does away with random leading underscores). This is really a more flexible way to go in the end anyway -- if a filesystem happens to want any extra allocations aside from the pagecache ones in ints write_begin function, it may now use GFP_KERNEL (rather than GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a random example). [kosaki.motohiro@jp.fujitsu.com: fix ubifs] [kosaki.motohiro@jp.fujitsu.com: fix fuse] Signed-off-by: Nick Piggin <npiggin@suse.de> Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: <stable@kernel.org> [2.6.28.x] Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Cleaned up the calling convention: just pass in the AOP flags untouched to the grab_cache_page_write_begin() function. That just simplifies everybody, and may even allow future expansion of the logic. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-04 20:00:53 +00:00
page = grab_cache_page_write_begin(mapping, index, flags);
if (!page) {
ext4_journal_stop(handle);
ret = -ENOMEM;
goto out;
}
*pagep = page;
if (ext4_should_dioread_nolock(inode))
ret = __block_write_begin(page, pos, len, ext4_get_block_write);
else
ret = __block_write_begin(page, pos, len, ext4_get_block);
if (!ret && ext4_should_journal_data(inode)) {
ret = ext4_walk_page_buffers(handle, page_buffers(page),
from, to, NULL,
do_journal_get_write_access);
}
if (ret) {
unlock_page(page);
page_cache_release(page);
/*
* __block_write_begin may have instantiated a few blocks
* outside i_size. Trim these off again. Don't need
* i_size_read because we hold i_mutex.
*
* Add inode to orphan list in case we crash before
* truncate finishes
*/
if (pos + len > inode->i_size && ext4_can_truncate(inode))
ext4_orphan_add(handle, inode);
ext4_journal_stop(handle);
if (pos + len > inode->i_size) {
ext4_truncate_failed_write(inode);
/*
* If truncate failed early the inode might
* still be on the orphan list; we need to
* make sure the inode is removed from the
* orphan list in that case.
*/
if (inode->i_nlink)
ext4_orphan_del(NULL, inode);
}
}
if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
goto retry;
out:
return ret;
}
/* For write_end() in data=journal mode */
static int write_end_fn(handle_t *handle, struct buffer_head *bh)
{
if (!buffer_mapped(bh) || buffer_freed(bh))
return 0;
set_buffer_uptodate(bh);
return ext4_handle_dirty_metadata(handle, NULL, bh);
}
static int ext4_generic_write_end(struct file *file,
struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
int i_size_changed = 0;
struct inode *inode = mapping->host;
handle_t *handle = ext4_journal_current_handle();
if (ext4_has_inline_data(inode))
copied = ext4_write_inline_data_end(inode, pos, len,
copied, page);
else
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);
i_size_changed = 1;
}
if (pos + copied > EXT4_I(inode)->i_disksize) {
/* We need to mark inode dirty even if
* new_i_size is less that inode->i_size
* bu greater than i_disksize.(hint delalloc)
*/
ext4_update_i_disksize(inode, (pos + copied));
i_size_changed = 1;
}
unlock_page(page);
page_cache_release(page);
/*
* Don't mark the inode dirty under page lock. First, it unnecessarily
* makes the holding time of page lock longer. Second, it forces lock
* ordering of page lock and transaction start for journaling
* filesystems.
*/
if (i_size_changed)
ext4_mark_inode_dirty(handle, inode);
return copied;
}
/*
* We need to pick up the new inode size which generic_commit_write gave us
* `file' can be NULL - eg, when called from page_symlink().
*
* ext4 never places buffers on inode->i_mapping->private_list. metadata
* buffers are managed internally.
*/
static int ext4_ordered_write_end(struct file *file,
struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
handle_t *handle = ext4_journal_current_handle();
struct inode *inode = mapping->host;
int ret = 0, ret2;
trace_ext4_ordered_write_end(inode, pos, len, copied);
ret = ext4_jbd2_file_inode(handle, inode);
if (ret == 0) {
ret2 = ext4_generic_write_end(file, mapping, pos, len, copied,
page, fsdata);
copied = ret2;
if (pos + len > inode->i_size && ext4_can_truncate(inode))
/* if we have allocated more blocks and copied
* less. We will have blocks allocated outside
* inode->i_size. So truncate them
*/
ext4_orphan_add(handle, inode);
if (ret2 < 0)
ret = ret2;
} else {
unlock_page(page);
page_cache_release(page);
}
ret2 = ext4_journal_stop(handle);
if (!ret)
ret = ret2;
if (pos + len > inode->i_size) {
ext4_truncate_failed_write(inode);
/*
* If truncate failed early the inode might still be
* on the orphan list; we need to make sure the inode
* is removed from the orphan list in that case.
*/
if (inode->i_nlink)
ext4_orphan_del(NULL, inode);
}
return ret ? ret : copied;
}
static int ext4_writeback_write_end(struct file *file,
struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
handle_t *handle = ext4_journal_current_handle();
struct inode *inode = mapping->host;
int ret = 0, ret2;
trace_ext4_writeback_write_end(inode, pos, len, copied);
ret2 = ext4_generic_write_end(file, mapping, pos, len, copied,
page, fsdata);
copied = ret2;
if (pos + len > inode->i_size && ext4_can_truncate(inode))
/* if we have allocated more blocks and copied
* less. We will have blocks allocated outside
* inode->i_size. So truncate them
*/
ext4_orphan_add(handle, inode);
if (ret2 < 0)
ret = ret2;
ret2 = ext4_journal_stop(handle);
if (!ret)
ret = ret2;
if (pos + len > inode->i_size) {
ext4_truncate_failed_write(inode);
/*
* If truncate failed early the inode might still be
* on the orphan list; we need to make sure the inode
* is removed from the orphan list in that case.
*/
if (inode->i_nlink)
ext4_orphan_del(NULL, inode);
}
return ret ? ret : copied;
}
static int ext4_journalled_write_end(struct file *file,
struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
handle_t *handle = ext4_journal_current_handle();
struct inode *inode = mapping->host;
int ret = 0, ret2;
int partial = 0;
unsigned from, to;
loff_t new_i_size;
trace_ext4_journalled_write_end(inode, pos, len, copied);
from = pos & (PAGE_CACHE_SIZE - 1);
to = from + len;
BUG_ON(!ext4_handle_valid(handle));
if (copied < len) {
if (!PageUptodate(page))
copied = 0;
page_zero_new_buffers(page, from+copied, to);
}
ret = ext4_walk_page_buffers(handle, page_buffers(page), from,
to, &partial, write_end_fn);
if (!partial)
SetPageUptodate(page);
new_i_size = pos + copied;
if (new_i_size > inode->i_size)
i_size_write(inode, pos+copied);
ext4_set_inode_state(inode, EXT4_STATE_JDATA);
EXT4_I(inode)->i_datasync_tid = handle->h_transaction->t_tid;
if (new_i_size > EXT4_I(inode)->i_disksize) {
ext4_update_i_disksize(inode, new_i_size);
ret2 = ext4_mark_inode_dirty(handle, inode);
if (!ret)
ret = ret2;
}
unlock_page(page);
page_cache_release(page);
if (pos + len > inode->i_size && ext4_can_truncate(inode))
/* if we have allocated more blocks and copied
* less. We will have blocks allocated outside
* inode->i_size. So truncate them
*/
ext4_orphan_add(handle, inode);
ret2 = ext4_journal_stop(handle);
if (!ret)
ret = ret2;
if (pos + len > inode->i_size) {
ext4_truncate_failed_write(inode);
/*
* If truncate failed early the inode might still be
* on the orphan list; we need to make sure the inode
* is removed from the orphan list in that case.
*/
if (inode->i_nlink)
ext4_orphan_del(NULL, inode);
}
return ret ? ret : copied;
}
/*
* Reserve a single cluster located at lblock
*/
static int ext4_da_reserve_space(struct inode *inode, ext4_lblk_t lblock)
{
int retries = 0;
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
struct ext4_inode_info *ei = EXT4_I(inode);
unsigned int md_needed;
int ret;
ext4_lblk_t save_last_lblock;
int save_len;
/*
* We will charge metadata quota at writeout time; this saves
* us from metadata over-estimation, though we may go over by
* a small amount in the end. Here we just reserve for data.
*/
ret = dquot_reserve_block(inode, EXT4_C2B(sbi, 1));
if (ret)
return ret;
/*
* recalculate the amount of metadata blocks to reserve
* in order to allocate nrblocks
* worse case is one extent per block
*/
repeat:
spin_lock(&ei->i_block_reservation_lock);
/*
* ext4_calc_metadata_amount() has side effects, which we have
* to be prepared undo if we fail to claim space.
*/
save_len = ei->i_da_metadata_calc_len;
save_last_lblock = ei->i_da_metadata_calc_last_lblock;
md_needed = EXT4_NUM_B2C(sbi,
ext4_calc_metadata_amount(inode, lblock));
trace_ext4_da_reserve_space(inode, md_needed);
/*
* We do still charge estimated metadata to the sb though;
* we cannot afford to run out of free blocks.
*/
if (ext4_claim_free_clusters(sbi, md_needed + 1, 0)) {
ei->i_da_metadata_calc_len = save_len;
ei->i_da_metadata_calc_last_lblock = save_last_lblock;
spin_unlock(&ei->i_block_reservation_lock);
if (ext4_should_retry_alloc(inode->i_sb, &retries)) {
yield();
goto repeat;
}
dquot_release_reservation_block(inode, EXT4_C2B(sbi, 1));
return -ENOSPC;
}
ei->i_reserved_data_blocks++;
ei->i_reserved_meta_blocks += md_needed;
spin_unlock(&ei->i_block_reservation_lock);
return 0; /* success */
}
static void ext4_da_release_space(struct inode *inode, int to_free)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
struct ext4_inode_info *ei = EXT4_I(inode);
if (!to_free)
return; /* Nothing to release, exit */
spin_lock(&EXT4_I(inode)->i_block_reservation_lock);
trace_ext4_da_release_space(inode, to_free);
if (unlikely(to_free > ei->i_reserved_data_blocks)) {
/*
* if there aren't enough reserved blocks, then the
* counter is messed up somewhere. Since this
* function is called from invalidate page, it's
* harmless to return without any action.
*/
ext4_msg(inode->i_sb, KERN_NOTICE, "ext4_da_release_space: "
"ino %lu, to_free %d with only %d reserved "
"data blocks", inode->i_ino, to_free,
ei->i_reserved_data_blocks);
WARN_ON(1);
to_free = ei->i_reserved_data_blocks;
}
ei->i_reserved_data_blocks -= to_free;
if (ei->i_reserved_data_blocks == 0) {
/*
* We can release all of the reserved metadata blocks
* only when we have written all of the delayed
* allocation blocks.
* Note that in case of bigalloc, i_reserved_meta_blocks,
* i_reserved_data_blocks, etc. refer to number of clusters.
*/
percpu_counter_sub(&sbi->s_dirtyclusters_counter,
ei->i_reserved_meta_blocks);
ei->i_reserved_meta_blocks = 0;
ei->i_da_metadata_calc_len = 0;
}
/* update fs dirty data blocks counter */
percpu_counter_sub(&sbi->s_dirtyclusters_counter, to_free);
spin_unlock(&EXT4_I(inode)->i_block_reservation_lock);
dquot_release_reservation_block(inode, EXT4_C2B(sbi, to_free));
}
static void ext4_da_page_release_reservation(struct page *page,
unsigned long offset)
{
int to_release = 0;
struct buffer_head *head, *bh;
unsigned int curr_off = 0;
struct inode *inode = page->mapping->host;
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
int num_clusters;
ext4_fsblk_t lblk;
head = page_buffers(page);
bh = head;
do {
unsigned int next_off = curr_off + bh->b_size;
if ((offset <= curr_off) && (buffer_delay(bh))) {
to_release++;
clear_buffer_delay(bh);
}
curr_off = next_off;
} while ((bh = bh->b_this_page) != head);
if (to_release) {
lblk = page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
ext4_es_remove_extent(inode, lblk, to_release);
}
/* If we have released all the blocks belonging to a cluster, then we
* need to release the reserved space for that cluster. */
num_clusters = EXT4_NUM_B2C(sbi, to_release);
while (num_clusters > 0) {
lblk = (page->index << (PAGE_CACHE_SHIFT - inode->i_blkbits)) +
((num_clusters - 1) << sbi->s_cluster_bits);
if (sbi->s_cluster_ratio == 1 ||
!ext4_find_delalloc_cluster(inode, lblk))
ext4_da_release_space(inode, 1);
num_clusters--;
}
}
/*
* Delayed allocation stuff
*/
/*
* mpage_da_submit_io - walks through extent of pages and try to write
* them with writepage() call back
*
* @mpd->inode: inode
* @mpd->first_page: first page of the extent
* @mpd->next_page: page after the last page of the extent
*
* By the time mpage_da_submit_io() is called we expect all blocks
* to be allocated. this may be wrong if allocation failed.
*
* As pages are already locked by write_cache_pages(), we can't use it
*/
static int mpage_da_submit_io(struct mpage_da_data *mpd,
struct ext4_map_blocks *map)
{
struct pagevec pvec;
unsigned long index, end;
int ret = 0, err, nr_pages, i;
struct inode *inode = mpd->inode;
struct address_space *mapping = inode->i_mapping;
loff_t size = i_size_read(inode);
unsigned int len, block_start;
struct buffer_head *bh, *page_bufs = NULL;
int journal_data = ext4_should_journal_data(inode);
sector_t pblock = 0, cur_logical = 0;
struct ext4_io_submit io_submit;
BUG_ON(mpd->next_page <= mpd->first_page);
memset(&io_submit, 0, sizeof(io_submit));
/*
* We need to start from the first_page to the next_page - 1
* to make sure we also write the mapped dirty buffer_heads.
* If we look at mpd->b_blocknr we would only be looking
* at the currently mapped buffer_heads.
*/
index = mpd->first_page;
end = mpd->next_page - 1;
pagevec_init(&pvec, 0);
while (index <= end) {
nr_pages = pagevec_lookup(&pvec, mapping, index, PAGEVEC_SIZE);
if (nr_pages == 0)
break;
for (i = 0; i < nr_pages; i++) {
int commit_write = 0, skip_page = 0;
struct page *page = pvec.pages[i];
index = page->index;
if (index > end)
break;
if (index == size >> PAGE_CACHE_SHIFT)
len = size & ~PAGE_CACHE_MASK;
else
len = PAGE_CACHE_SIZE;
if (map) {
cur_logical = index << (PAGE_CACHE_SHIFT -
inode->i_blkbits);
pblock = map->m_pblk + (cur_logical -
map->m_lblk);
}
index++;
BUG_ON(!PageLocked(page));
BUG_ON(PageWriteback(page));
/*
* If the page does not have buffers (for
* whatever reason), try to create them using
* __block_write_begin. If this fails,
* skip the page and move on.
*/
if (!page_has_buffers(page)) {
if (__block_write_begin(page, 0, len,
noalloc_get_block_write)) {
skip_page:
unlock_page(page);
continue;
}
commit_write = 1;
}
bh = page_bufs = page_buffers(page);
block_start = 0;
do {
if (!bh)
goto skip_page;
if (map && (cur_logical >= map->m_lblk) &&
(cur_logical <= (map->m_lblk +
(map->m_len - 1)))) {
if (buffer_delay(bh)) {
clear_buffer_delay(bh);
bh->b_blocknr = pblock;
}
if (buffer_unwritten(bh) ||
buffer_mapped(bh))
BUG_ON(bh->b_blocknr != pblock);
if (map->m_flags & EXT4_MAP_UNINIT)
set_buffer_uninit(bh);
clear_buffer_unwritten(bh);
}
/*
* skip page if block allocation undone and
* block is dirty
*/
if (ext4_bh_delay_or_unwritten(NULL, bh))
skip_page = 1;
bh = bh->b_this_page;
block_start += bh->b_size;
cur_logical++;
pblock++;
} while (bh != page_bufs);
if (skip_page)
goto skip_page;
if (commit_write)
/* mark the buffer_heads as dirty & uptodate */
block_commit_write(page, 0, len);
clear_page_dirty_for_io(page);
/*
* Delalloc doesn't support data journalling,
* but eventually maybe we'll lift this
* restriction.
*/
if (unlikely(journal_data && PageChecked(page)))
err = __ext4_journalled_writepage(page, len);
else if (test_opt(inode->i_sb, MBLK_IO_SUBMIT))
err = ext4_bio_write_page(&io_submit, page,
len, mpd->wbc);
else if (buffer_uninit(page_bufs)) {
ext4_set_bh_endio(page_bufs, inode);
err = block_write_full_page_endio(page,
noalloc_get_block_write,
mpd->wbc, ext4_end_io_buffer_write);
} else
err = block_write_full_page(page,
noalloc_get_block_write, mpd->wbc);
if (!err)
mpd->pages_written++;
/*
* In error case, we have to continue because
* remaining pages are still locked
*/
if (ret == 0)
ret = err;
}
pagevec_release(&pvec);
}
ext4_io_submit(&io_submit);
return ret;
}
static void ext4_da_block_invalidatepages(struct mpage_da_data *mpd)
{
int nr_pages, i;
pgoff_t index, end;
struct pagevec pvec;
struct inode *inode = mpd->inode;
struct address_space *mapping = inode->i_mapping;
ext4_lblk_t start, last;
index = mpd->first_page;
end = mpd->next_page - 1;
start = index << (PAGE_CACHE_SHIFT - inode->i_blkbits);
last = end << (PAGE_CACHE_SHIFT - inode->i_blkbits);
ext4_es_remove_extent(inode, start, last - start + 1);
pagevec_init(&pvec, 0);
while (index <= end) {
nr_pages = pagevec_lookup(&pvec, mapping, index, PAGEVEC_SIZE);
if (nr_pages == 0)
break;
for (i = 0; i < nr_pages; i++) {
struct page *page = pvec.pages[i];
if (page->index > end)
break;
BUG_ON(!PageLocked(page));
BUG_ON(PageWriteback(page));
block_invalidatepage(page, 0);
ClearPageUptodate(page);
unlock_page(page);
}
index = pvec.pages[nr_pages - 1]->index + 1;
pagevec_release(&pvec);
}
return;
}
static void ext4_print_free_blocks(struct inode *inode)
{
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
struct super_block *sb = inode->i_sb;
ext4_msg(sb, KERN_CRIT, "Total free blocks count %lld",
EXT4_C2B(EXT4_SB(inode->i_sb),
ext4_count_free_clusters(inode->i_sb)));
ext4_msg(sb, KERN_CRIT, "Free/Dirty block details");
ext4_msg(sb, KERN_CRIT, "free_blocks=%lld",
(long long) EXT4_C2B(EXT4_SB(inode->i_sb),
percpu_counter_sum(&sbi->s_freeclusters_counter)));
ext4_msg(sb, KERN_CRIT, "dirty_blocks=%lld",
(long long) EXT4_C2B(EXT4_SB(inode->i_sb),
percpu_counter_sum(&sbi->s_dirtyclusters_counter)));
ext4_msg(sb, KERN_CRIT, "Block reservation details");
ext4_msg(sb, KERN_CRIT, "i_reserved_data_blocks=%u",
EXT4_I(inode)->i_reserved_data_blocks);
ext4_msg(sb, KERN_CRIT, "i_reserved_meta_blocks=%u",
EXT4_I(inode)->i_reserved_meta_blocks);
return;
}
/*
* mpage_da_map_and_submit - go through given space, map them
* if necessary, and then submit them for I/O
*
* @mpd - bh describing space
*
* The function skips space we know is already mapped to disk blocks.
*
*/
static void mpage_da_map_and_submit(struct mpage_da_data *mpd)
{
int err, blks, get_blocks_flags;
struct ext4_map_blocks map, *mapp = NULL;
sector_t next = mpd->b_blocknr;
unsigned max_blocks = mpd->b_size >> mpd->inode->i_blkbits;
loff_t disksize = EXT4_I(mpd->inode)->i_disksize;
handle_t *handle = NULL;
/*
* If the blocks are mapped already, or we couldn't accumulate
* any blocks, then proceed immediately to the submission stage.
*/
if ((mpd->b_size == 0) ||
((mpd->b_state & (1 << BH_Mapped)) &&
!(mpd->b_state & (1 << BH_Delay)) &&
!(mpd->b_state & (1 << BH_Unwritten))))
goto submit_io;
handle = ext4_journal_current_handle();
BUG_ON(!handle);
/*
* Call ext4_map_blocks() to allocate any delayed allocation
* blocks, or to convert an uninitialized extent to be
* initialized (in the case where we have written into
* one or more preallocated blocks).
*
* We pass in the magic EXT4_GET_BLOCKS_DELALLOC_RESERVE to
* indicate that we are on the delayed allocation path. This
* affects functions in many different parts of the allocation
* call path. This flag exists primarily because we don't
* want to change *many* call functions, so ext4_map_blocks()
* will set the EXT4_STATE_DELALLOC_RESERVED flag once the
* inode's allocation semaphore is taken.
*
* If the blocks in questions were delalloc blocks, set
* EXT4_GET_BLOCKS_DELALLOC_RESERVE so the delalloc accounting
* variables are updated after the blocks have been allocated.
*/
map.m_lblk = next;
map.m_len = max_blocks;
get_blocks_flags = EXT4_GET_BLOCKS_CREATE;
if (ext4_should_dioread_nolock(mpd->inode))
get_blocks_flags |= EXT4_GET_BLOCKS_IO_CREATE_EXT;
if (mpd->b_state & (1 << BH_Delay))
get_blocks_flags |= EXT4_GET_BLOCKS_DELALLOC_RESERVE;
blks = ext4_map_blocks(handle, mpd->inode, &map, get_blocks_flags);
if (blks < 0) {
struct super_block *sb = mpd->inode->i_sb;
err = blks;
/*
* If get block returns EAGAIN or ENOSPC and there
* appears to be free blocks we will just let
* mpage_da_submit_io() unlock all of the pages.
*/
if (err == -EAGAIN)
goto submit_io;
if (err == -ENOSPC && ext4_count_free_clusters(sb)) {
mpd->retval = err;
goto submit_io;
}
/*
* get block failure will cause us to loop in
* writepages, because a_ops->writepage won't be able
* to make progress. The page will be redirtied by
* writepage and writepages will again try to write
* the same.
*/
if (!(EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED)) {
ext4_msg(sb, KERN_CRIT,
"delayed block allocation failed for inode %lu "
"at logical offset %llu with max blocks %zd "
"with error %d", mpd->inode->i_ino,
(unsigned long long) next,
mpd->b_size >> mpd->inode->i_blkbits, err);
ext4_msg(sb, KERN_CRIT,
"This should not happen!! Data will be lost\n");
if (err == -ENOSPC)
ext4_print_free_blocks(mpd->inode);
}
/* invalidate all the pages */
ext4_da_block_invalidatepages(mpd);
/* Mark this page range as having been completed */
mpd->io_done = 1;
return;
}
BUG_ON(blks == 0);
mapp = &map;
if (map.m_flags & EXT4_MAP_NEW) {
struct block_device *bdev = mpd->inode->i_sb->s_bdev;
int i;
for (i = 0; i < map.m_len; i++)
unmap_underlying_metadata(bdev, map.m_pblk + i);
}
/*
* Update on-disk size along with block allocation.
*/
disksize = ((loff_t) next + blks) << mpd->inode->i_blkbits;
if (disksize > i_size_read(mpd->inode))
disksize = i_size_read(mpd->inode);
if (disksize > EXT4_I(mpd->inode)->i_disksize) {
ext4_update_i_disksize(mpd->inode, disksize);
err = ext4_mark_inode_dirty(handle, mpd->inode);
if (err)
ext4_error(mpd->inode->i_sb,
"Failed to mark inode %lu dirty",
mpd->inode->i_ino);
}
submit_io:
mpage_da_submit_io(mpd, mapp);
mpd->io_done = 1;
}
#define BH_FLAGS ((1 << BH_Uptodate) | (1 << BH_Mapped) | \
(1 << BH_Delay) | (1 << BH_Unwritten))
/*
* mpage_add_bh_to_extent - try to add one more block to extent of blocks
*
* @mpd->lbh - extent of blocks
* @logical - logical number of the block in the file
* @bh - bh of the block (used to access block's state)
*
* the function is used to collect contig. blocks in same state
*/
static void mpage_add_bh_to_extent(struct mpage_da_data *mpd,
sector_t logical, size_t b_size,
unsigned long b_state)
{
sector_t next;
int nrblocks = mpd->b_size >> mpd->inode->i_blkbits;
ext4: don't scan/accumulate more pages than mballoc will allocate There was a bug reported on RHEL5 that a 10G dd on a 12G box had a very, very slow sync after that. At issue was the loop in write_cache_pages scanning all the way to the end of the 10G file, even though the subsequent call to mpage_da_submit_io would only actually write a smallish amt; then we went back to the write_cache_pages loop ... wasting tons of time in calling __mpage_da_writepage for thousands of pages we would just revisit (many times) later. Upstream it's not such a big issue for sys_sync because we get to the loop with a much smaller nr_to_write, which limits the loop. However, talking with Aneesh he realized that fsync upstream still gets here with a very large nr_to_write and we face the same problem. This patch makes mpage_add_bh_to_extent stop the loop after we've accumulated 2048 pages, by setting mpd->io_done = 1; which ultimately causes the write_cache_pages loop to break. Repeating the test with a dirty_ratio of 80 (to leave something for fsync to do), I don't see huge IO performance gains, but the reduction in cpu usage is striking: 80% usage with stock, and 2% with the below patch. Instrumenting the loop in write_cache_pages clearly shows that we are wasting time here. Eventually we need to change mpage_da_map_pages() also submit its I/O to the block layer, subsuming mpage_da_submit_io(), and then change it call ext4_get_blocks() multiple times. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-05-16 08:00:00 +00:00
/*
* XXX Don't go larger than mballoc is willing to allocate
* This is a stopgap solution. We eventually need to fold
* mpage_da_submit_io() into this function and then call
* ext4_map_blocks() multiple times in a loop
ext4: don't scan/accumulate more pages than mballoc will allocate There was a bug reported on RHEL5 that a 10G dd on a 12G box had a very, very slow sync after that. At issue was the loop in write_cache_pages scanning all the way to the end of the 10G file, even though the subsequent call to mpage_da_submit_io would only actually write a smallish amt; then we went back to the write_cache_pages loop ... wasting tons of time in calling __mpage_da_writepage for thousands of pages we would just revisit (many times) later. Upstream it's not such a big issue for sys_sync because we get to the loop with a much smaller nr_to_write, which limits the loop. However, talking with Aneesh he realized that fsync upstream still gets here with a very large nr_to_write and we face the same problem. This patch makes mpage_add_bh_to_extent stop the loop after we've accumulated 2048 pages, by setting mpd->io_done = 1; which ultimately causes the write_cache_pages loop to break. Repeating the test with a dirty_ratio of 80 (to leave something for fsync to do), I don't see huge IO performance gains, but the reduction in cpu usage is striking: 80% usage with stock, and 2% with the below patch. Instrumenting the loop in write_cache_pages clearly shows that we are wasting time here. Eventually we need to change mpage_da_map_pages() also submit its I/O to the block layer, subsuming mpage_da_submit_io(), and then change it call ext4_get_blocks() multiple times. Signed-off-by: Eric Sandeen <sandeen@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2010-05-16 08:00:00 +00:00
*/
if (nrblocks >= 8*1024*1024/mpd->inode->i_sb->s_blocksize)
goto flush_it;
/* check if thereserved journal credits might overflow */
if (!(ext4_test_inode_flag(mpd->inode, EXT4_INODE_EXTENTS))) {
if (nrblocks >= EXT4_MAX_TRANS_DATA) {
/*
* With non-extent format we are limited by the journal
* credit available. Total credit needed to insert
* nrblocks contiguous blocks is dependent on the
* nrblocks. So limit nrblocks.
*/
goto flush_it;
} else if ((nrblocks + (b_size >> mpd->inode->i_blkbits)) >
EXT4_MAX_TRANS_DATA) {
/*
* Adding the new buffer_head would make it cross the
* allowed limit for which we have journal credit
* reserved. So limit the new bh->b_size
*/
b_size = (EXT4_MAX_TRANS_DATA - nrblocks) <<
mpd->inode->i_blkbits;
/* we will do mpage_da_submit_io in the next loop */
}
}
/*
* First block in the extent
*/
if (mpd->b_size == 0) {
mpd->b_blocknr = logical;
mpd->b_size = b_size;
mpd->b_state = b_state & BH_FLAGS;
return;
}
next = mpd->b_blocknr + nrblocks;
/*
* Can we merge the block to our big extent?
*/
if (logical == next && (b_state & BH_FLAGS) == mpd->b_state) {
mpd->b_size += b_size;
return;
}
flush_it:
/*
* We couldn't merge the block to our extent, so we
* need to flush current extent and start new one
*/
mpage_da_map_and_submit(mpd);
return;
}
static int ext4_bh_delay_or_unwritten(handle_t *handle, struct buffer_head *bh)
{
return (buffer_delay(bh) || buffer_unwritten(bh)) && buffer_dirty(bh);
}
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-09 23:20:51 +00:00
/*
* This function is grabs code from the very beginning of
* ext4_map_blocks, but assumes that the caller is from delayed write
* time. This function looks up the requested blocks and sets the
* buffer delay bit under the protection of i_data_sem.
*/
static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
struct ext4_map_blocks *map,
struct buffer_head *bh)
{
int retval;
sector_t invalid_block = ~((sector_t) 0xffff);
if (invalid_block < ext4_blocks_count(EXT4_SB(inode->i_sb)->s_es))
invalid_block = ~0;
map->m_flags = 0;
ext_debug("ext4_da_map_blocks(): inode %lu, max_blocks %u,"
"logical block %lu\n", inode->i_ino, map->m_len,
(unsigned long) map->m_lblk);
/*
* Try to see if we can get the block without requesting a new
* file system block.
*/
down_read((&EXT4_I(inode)->i_data_sem));
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
retval = ext4_ext_map_blocks(NULL, inode, map, 0);
else
retval = ext4_ind_map_blocks(NULL, inode, map, 0);
if (retval == 0) {
/*
* XXX: __block_prepare_write() unmaps passed block,
* is it OK?
*/
/* If the block was allocated from previously allocated cluster,
* then we dont need to reserve it again. */
if (!(map->m_flags & EXT4_MAP_FROM_CLUSTER)) {
retval = ext4_da_reserve_space(inode, iblock);
if (retval)
/* not enough space to reserve */
goto out_unlock;
}
retval = ext4_es_insert_extent(inode, map->m_lblk, map->m_len);
if (retval)
goto out_unlock;
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-09 23:20:51 +00:00
/* Clear EXT4_MAP_FROM_CLUSTER flag since its purpose is served
* and it should not appear on the bh->b_state.
*/
map->m_flags &= ~EXT4_MAP_FROM_CLUSTER;
map_bh(bh, inode->i_sb, invalid_block);
set_buffer_new(bh);
set_buffer_delay(bh);
}
out_unlock:
up_read((&EXT4_I(inode)->i_data_sem));
return retval;
}
/*
* This is a special get_blocks_t callback which is used by
* ext4_da_write_begin(). It will either return mapped block or
* reserve space for a single block.
*
* For delayed buffer_head we have BH_Mapped, BH_New, BH_Delay set.
* We also have b_blocknr = -1 and b_bdev initialized properly
*
* For unwritten buffer_head we have BH_Mapped, BH_New, BH_Unwritten set.
* We also have b_blocknr = physicalblock mapping unwritten extent and b_bdev
* initialized properly.
*/
static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
struct buffer_head *bh, int create)
{
struct ext4_map_blocks map;
int ret = 0;
BUG_ON(create == 0);
BUG_ON(bh->b_size != inode->i_sb->s_blocksize);
map.m_lblk = iblock;
map.m_len = 1;
/*
* first, we need to know whether the block is allocated already
* preallocated blocks are unmapped but should treated
* the same as allocated blocks.
*/
ext4: attempt to fix race in bigalloc code path Currently, there exists a race between delayed allocated writes and the writeback when bigalloc feature is in use. The race was because we wanted to determine what blocks in a cluster are under delayed allocation and we were using buffer_delayed(bh) check for it. But, the writeback codepath clears this bit without any synchronization which resulted in a race and an ext4 warning similar to: EXT4-fs (ram1): ext4_da_update_reserve_space: ino 13, used 1 with only 0 reserved data blocks The race existed in two places. (1) between ext4_find_delalloc_range() and ext4_map_blocks() when called from writeback code path. (2) between ext4_find_delalloc_range() and ext4_da_get_block_prep() (where buffer_delayed(bh) is set. To fix (1), this patch introduces a new buffer_head state bit - BH_Da_Mapped. This bit is set under the protection of EXT4_I(inode)->i_data_sem when we have actually mapped the delayed allocated blocks during the writeout time. We can now reliably check for this bit inside ext4_find_delalloc_range() to determine whether the reservation for the blocks have already been claimed or not. To fix (2), it was necessary to set buffer_delay(bh) under the protection of i_data_sem. So, I extracted the very beginning of ext4_map_blocks into a new function - ext4_da_map_blocks() - and performed the required setting of bh_delay bit and the quota reservation under the protection of i_data_sem. These two fixes makes the checking of buffer_delay(bh) and buffer_da_mapped(bh) consistent, thus removing the race. Tested: I was able to reproduce the problem by running 'dd' and 'fsync' in parallel. Also, xfstests sometimes used to reproduce this race. After the fix both my test and xfstests were successful and no race (warning message) was observed. Google-Bug-Id: 4997027 Signed-off-by: Aditya Kali <adityakali@google.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-09 23:20:51 +00:00
ret = ext4_da_map_blocks(inode, iblock, &map, bh);
if (ret <= 0)
return ret;
map_bh(bh, inode->i_sb, map.m_pblk);
bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags;
if (buffer_unwritten(bh)) {
/* A delayed write to unwritten bh should be marked
* new and mapped. Mapped ensures that we don't do
* get_block multiple times when we write to the same
* offset and new ensures that we do proper zero out
* for partial write.
*/
set_buffer_new(bh);
set_buffer_mapped(bh);
}
return 0;
}
/*
* This function is used as a standard get_block_t calback function
* when there is no desire to allocate any blocks. It is used as a
* callback function for block_write_begin() and block_write_full_page().
* These functions should only try to map a single block at a time.
*
* Since this function doesn't do block allocations even if the caller
* requests it by passing in create=1, it is critically important that
* any caller checks to make sure that any buffer heads are returned
* by this function are either all already mapped or marked for
* delayed allocation before calling block_write_full_page(). Otherwise,
* b_blocknr could be left unitialized, and the page write functions will
* be taken by surprise.
*/
static int noalloc_get_block_write(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
BUG_ON(bh_result->b_size != inode->i_sb->s_blocksize);
return _ext4_get_block(inode, iblock, bh_result, 0);
}
static int bget_one(handle_t *handle, struct buffer_head *bh)
{
get_bh(bh);
return 0;
}
static int bput_one(handle_t *handle, struct buffer_head *bh)
{
put_bh(bh);
return 0;
}
static int __ext4_journalled_writepage(struct page *page,
unsigned int len)
{
struct address_space *mapping = page->mapping;
struct inode *inode = mapping->host;
struct buffer_head *page_bufs;
handle_t *handle = NULL;
int ret = 0;
int err;
ClearPageChecked(page);
page_bufs = page_buffers(page);
BUG_ON(!page_bufs);
ext4_walk_page_buffers(handle, page_bufs, 0, len, NULL, bget_one);
/* As soon as we unlock the page, it can go away, but we have
* references to buffers so we are safe */
unlock_page(page);
handle = ext4_journal_start(inode, ext4_writepage_trans_blocks(inode));
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
goto out;
}
BUG_ON(!ext4_handle_valid(handle));
ret = ext4_walk_page_buffers(handle, page_bufs, 0, len, NULL,
do_journal_get_write_access);
err = ext4_walk_page_buffers(handle, page_bufs, 0, len, NULL,
write_end_fn);
if (ret == 0)
ret = err;
EXT4_I(inode)->i_datasync_tid = handle->h_transaction->t_tid;
err = ext4_journal_stop(handle);
if (!ret)
ret = err;
ext4_walk_page_buffers(handle, page_bufs, 0, len, NULL, bput_one);
ext4_set_inode_state(inode, EXT4_STATE_JDATA);
out:
return ret;
}
/*
* Note that we don't need to start a transaction unless we're journaling data
* because we should have holes filled from ext4_page_mkwrite(). We even don't
* need to file the inode to the transaction's list in ordered mode because if
* we are writing back data added by write(), the inode is already there and if
* we are writing back data modified via mmap(), no one guarantees in which
* transaction the data will hit the disk. In case we are journaling data, we
* cannot start transaction directly because transaction start ranks above page
* lock so we have to do some magic.
*
* This function can get called via...
* - ext4_da_writepages after taking page lock (have journal handle)
* - journal_submit_inode_data_buffers (no journal handle)
* - shrink_page_list via the kswapd/direct reclaim (no journal handle)
* - grab_page_cache when doing write_begin (have journal handle)
*
* We don't do any block allocation in this function. If we have page with
* multiple blocks we need to write those buffer_heads that are mapped. This
* is important for mmaped based write. So if we do with blocksize 1K
* truncate(f, 1024);
* a = mmap(f, 0, 4096);
* a[0] = 'a';
* truncate(f, 4096);
* we have in the page first buffer_head mapped via page_mkwrite call back
* but other buffer_heads would be unmapped but dirty (dirty done via the
* do_wp_page). So writepage should write the first block. If we modify
* the mmap area beyond 1024 we will again get a page_fault and the
* page_mkwrite callback will do the block allocation and mark the
* buffer_heads mapped.
*
* We redirty the page if we have any buffer_heads that is either delay or
* unwritten in the page.
*
* We can get recursively called as show below.
*
* ext4_writepage() -> kmalloc() -> __alloc_pages() -> page_launder() ->
* ext4_writepage()
*
* But since we don't do any block allocation we should not deadlock.
* Page also have the dirty flag cleared so we don't get recurive page_lock.
*/
static int ext4_writepage(struct page *page,
struct writeback_control *wbc)
{
int ret = 0, commit_write = 0;
loff_t size;
unsigned int len;
struct buffer_head *page_bufs = NULL;
struct inode *inode = page->mapping->host;
trace_ext4_writepage(page);
size = i_size_read(inode);
if (page->index == size >> PAGE_CACHE_SHIFT)
len = size & ~PAGE_CACHE_MASK;
else
len = PAGE_CACHE_SIZE;
/*
* If the page does not have buffers (for whatever reason),
* try to create them using __block_write_begin. If this
* fails, redirty the page and move on.
*/
if (!page_has_buffers(page)) {
if (__block_write_begin(page, 0, len,
noalloc_get_block_write)) {
redirty_page:
redirty_page_for_writepage(wbc, page);
unlock_page(page);
return 0;
}
commit_write = 1;
}
page_bufs = page_buffers(page);
if (ext4_walk_page_buffers(NULL, page_bufs, 0, len, NULL,
ext4_bh_delay_or_unwritten)) {
/*
* We don't want to do block allocation, so redirty
* the page and return. We may reach here when we do
* a journal commit via journal_submit_inode_data_buffers.
* We can also reach here via shrink_page_list but it
* should never be for direct reclaim so warn if that
* happens
*/
WARN_ON_ONCE((current->flags & (PF_MEMALLOC|PF_KSWAPD)) ==
PF_MEMALLOC);
goto redirty_page;
}
if (commit_write)
/* now mark the buffer_heads as dirty and uptodate */
block_commit_write(page, 0, len);
if (PageChecked(page) && ext4_should_journal_data(inode))
/*
* It's mmapped pagecache. Add buffers and journal it. There
* doesn't seem much point in redirtying the page here.
*/
return __ext4_journalled_writepage(page, len);
if (buffer_uninit(page_bufs)) {
ext4_set_bh_endio(page_bufs, inode);
ret = block_write_full_page_endio(page, noalloc_get_block_write,
wbc, ext4_end_io_buffer_write);
} else
ret = block_write_full_page(page, noalloc_get_block_write,
wbc);
return ret;
}
/*
* This is called via ext4_da_writepages() to
* calculate the total number of credits to reserve to fit
* a single extent allocation into a single transaction,
* ext4_da_writpeages() will loop calling this before
* the block allocation.
*/
static int ext4_da_writepages_trans_blocks(struct inode *inode)
{
int max_blocks = EXT4_I(inode)->i_reserved_data_blocks;
/*
* With non-extent format the journal credit needed to
* insert nrblocks contiguous block is dependent on
* number of contiguous block. So we will limit
* number of contiguous block to a sane value
*/
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) &&
(max_blocks > EXT4_MAX_TRANS_DATA))
max_blocks = EXT4_MAX_TRANS_DATA;
return ext4_chunk_trans_blocks(inode, max_blocks);
}
/*
* write_cache_pages_da - walk the list of dirty pages of the given
* address space and accumulate pages that need writing, and call
* mpage_da_map_and_submit to map a single contiguous memory region
* and then write them.
*/
static int write_cache_pages_da(struct address_space *mapping,
struct writeback_control *wbc,
struct mpage_da_data *mpd,
pgoff_t *done_index)
{
struct buffer_head *bh, *head;
struct inode *inode = mapping->host;
struct pagevec pvec;
unsigned int nr_pages;
sector_t logical;
pgoff_t index, end;
long nr_to_write = wbc->nr_to_write;
int i, tag, ret = 0;
memset(mpd, 0, sizeof(struct mpage_da_data));
mpd->wbc = wbc;
mpd->inode = inode;
pagevec_init(&pvec, 0);
index = wbc->range_start >> PAGE_CACHE_SHIFT;
end = wbc->range_end >> PAGE_CACHE_SHIFT;
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
tag = PAGECACHE_TAG_TOWRITE;
else
tag = PAGECACHE_TAG_DIRTY;
*done_index = index;
while (index <= end) {
nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
if (nr_pages == 0)
return 0;
for (i = 0; i < nr_pages; i++) {
struct page *page = pvec.pages[i];
/*
* At this point, the page may be truncated or
* invalidated (changing page->mapping to NULL), or
* even swizzled back from swapper_space to tmpfs file
* mapping. However, page->index will not change
* because we have a reference on the page.
*/
if (page->index > end)
goto out;
*done_index = page->index + 1;
/*
* If we can't merge this page, and we have
* accumulated an contiguous region, write it
*/
if ((mpd->next_page != page->index) &&
(mpd->next_page != mpd->first_page)) {
mpage_da_map_and_submit(mpd);
goto ret_extent_tail;
}
lock_page(page);
/*
* If the page is no longer dirty, or its
* mapping no longer corresponds to inode we
* are writing (which means it has been
* truncated or invalidated), or the page is
* already under writeback and we are not
* doing a data integrity writeback, skip the page
*/
if (!PageDirty(page) ||
(PageWriteback(page) &&
(wbc->sync_mode == WB_SYNC_NONE)) ||
unlikely(page->mapping != mapping)) {
unlock_page(page);
continue;
}
wait_on_page_writeback(page);
BUG_ON(PageWriteback(page));
if (mpd->next_page != page->index)
mpd->first_page = page->index;
mpd->next_page = page->index + 1;
logical = (sector_t) page->index <<
(PAGE_CACHE_SHIFT - inode->i_blkbits);
if (!page_has_buffers(page)) {
mpage_add_bh_to_extent(mpd, logical,
PAGE_CACHE_SIZE,
(1 << BH_Dirty) | (1 << BH_Uptodate));
if (mpd->io_done)
goto ret_extent_tail;
} else {
/*
* Page with regular buffer heads,
* just add all dirty ones
*/
head = page_buffers(page);
bh = head;
do {
BUG_ON(buffer_locked(bh));
/*
* We need to try to allocate
* unmapped blocks in the same page.
* Otherwise we won't make progress
* with the page in ext4_writepage
*/
if (ext4_bh_delay_or_unwritten(NULL, bh)) {
mpage_add_bh_to_extent(mpd, logical,
bh->b_size,
bh->b_state);
if (mpd->io_done)
goto ret_extent_tail;
} else if (buffer_dirty(bh) && (buffer_mapped(bh))) {
/*
* mapped dirty buffer. We need
* to update the b_state
* because we look at b_state
* in mpage_da_map_blocks. We
* don't update b_size because
* if we find an unmapped
* buffer_head later we need to
* use the b_state flag of that
* buffer_head.
*/
if (mpd->b_size == 0)
mpd->b_state = bh->b_state & BH_FLAGS;
}
logical++;
} while ((bh = bh->b_this_page) != head);
}
if (nr_to_write > 0) {
nr_to_write--;
if (nr_to_write == 0 &&
wbc->sync_mode == WB_SYNC_NONE)
/*
* We stop writing back only if we are
* not doing integrity sync. In case of
* integrity sync we have to keep going
* because someone may be concurrently
* dirtying pages, and we might have
* synced a lot of newly appeared dirty
* pages, but have not synced all of the
* old dirty pages.
*/
goto out;
}
}
pagevec_release(&pvec);
cond_resched();
}
return 0;
ret_extent_tail:
ret = MPAGE_DA_EXTENT_TAIL;
out:
pagevec_release(&pvec);
cond_resched();
return ret;
}
static int ext4_da_writepages(struct address_space *mapping,
struct writeback_control *wbc)
{
pgoff_t index;
int range_whole = 0;
handle_t *handle = NULL;
struct mpage_da_data mpd;
struct inode *inode = mapping->host;
int pages_written = 0;
unsigned int max_pages;
int range_cyclic, cycled = 1, io_done = 0;
int needed_blocks, ret = 0;
long desired_nr_to_write, nr_to_writebump = 0;
loff_t range_start = wbc->range_start;
struct ext4_sb_info *sbi = EXT4_SB(mapping->host->i_sb);
pgoff_t done_index = 0;
pgoff_t end;
struct blk_plug plug;
trace_ext4_da_writepages(inode, wbc);
/*
* No pages to write? This is mainly a kludge to avoid starting
* a transaction for special inodes like journal inode on last iput()
* because that could violate lock ordering on umount
*/
if (!mapping->nrpages || !mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
return 0;
/*
* If the filesystem has aborted, it is read-only, so return
* right away instead of dumping stack traces later on that
* will obscure the real source of the problem. We test
* EXT4_MF_FS_ABORTED instead of sb->s_flag's MS_RDONLY because
* the latter could be true if the filesystem is mounted
* read-only, and in that case, ext4_da_writepages should
* *never* be called, so if that ever happens, we would want
* the stack trace.
*/
if (unlikely(sbi->s_mount_flags & EXT4_MF_FS_ABORTED))
return -EROFS;
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
range_whole = 1;
range_cyclic = wbc->range_cyclic;
if (wbc->range_cyclic) {
index = mapping->writeback_index;
if (index)
cycled = 0;
wbc->range_start = index << PAGE_CACHE_SHIFT;
wbc->range_end = LLONG_MAX;
wbc->range_cyclic = 0;
end = -1;
} else {
index = wbc->range_start >> PAGE_CACHE_SHIFT;
end = wbc->range_end >> PAGE_CACHE_SHIFT;
}
/*
* This works around two forms of stupidity. The first is in
* the writeback code, which caps the maximum number of pages
* written to be 1024 pages. This is wrong on multiple
* levels; different architectues have a different page size,
* which changes the maximum amount of data which gets
* written. Secondly, 4 megabytes is way too small. XFS
* forces this value to be 16 megabytes by multiplying
* nr_to_write parameter by four, and then relies on its
* allocator to allocate larger extents to make them
* contiguous. Unfortunately this brings us to the second
* stupidity, which is that ext4's mballoc code only allocates
* at most 2048 blocks. So we force contiguous writes up to
* the number of dirty blocks in the inode, or
* sbi->max_writeback_mb_bump whichever is smaller.
*/
max_pages = sbi->s_max_writeback_mb_bump << (20 - PAGE_CACHE_SHIFT);
if (!range_cyclic && range_whole) {
if (wbc->nr_to_write == LONG_MAX)
desired_nr_to_write = wbc->nr_to_write;
else
desired_nr_to_write = wbc->nr_to_write * 8;
} else
desired_nr_to_write = ext4_num_dirty_pages(inode, index,
max_pages);
if (desired_nr_to_write > max_pages)
desired_nr_to_write = max_pages;
if (wbc->nr_to_write < desired_nr_to_write) {
nr_to_writebump = desired_nr_to_write - wbc->nr_to_write;
wbc->nr_to_write = desired_nr_to_write;
}
retry:
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
tag_pages_for_writeback(mapping, index, end);
blk_start_plug(&plug);
while (!ret && wbc->nr_to_write > 0) {
/*
* we insert one extent at a time. So we need
* credit needed for single extent allocation.
* journalled mode is currently not supported
* by delalloc
*/
BUG_ON(ext4_should_journal_data(inode));
needed_blocks = ext4_da_writepages_trans_blocks(inode);
/* start a new transaction*/
handle = ext4_journal_start(inode, needed_blocks);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
ext4_msg(inode->i_sb, KERN_CRIT, "%s: jbd2_start: "
"%ld pages, ino %lu; err %d", __func__,
wbc->nr_to_write, inode->i_ino, ret);
blk_finish_plug(&plug);
goto out_writepages;
}
/*
* Now call write_cache_pages_da() to find the next
* contiguous region of logical blocks that need
* blocks to be allocated by ext4 and submit them.
*/
ret = write_cache_pages_da(mapping, wbc, &mpd, &done_index);
/*
* If we have a contiguous extent of pages and we
* haven't done the I/O yet, map the blocks and submit
* them for I/O.
*/
if (!mpd.io_done && mpd.next_page != mpd.first_page) {
mpage_da_map_and_submit(&mpd);
ret = MPAGE_DA_EXTENT_TAIL;
}
trace_ext4_da_write_pages(inode, &mpd);
wbc->nr_to_write -= mpd.pages_written;
ext4_journal_stop(handle);
if ((mpd.retval == -ENOSPC) && sbi->s_journal) {
/* commit the transaction which would
* free blocks released in the transaction
* and try again
*/
jbd2_journal_force_commit_nested(sbi->s_journal);
ret = 0;
} else if (ret == MPAGE_DA_EXTENT_TAIL) {
/*
* Got one extent now try with rest of the pages.
* If mpd.retval is set -EIO, journal is aborted.
* So we don't need to write any more.
*/
pages_written += mpd.pages_written;
ret = mpd.retval;
io_done = 1;
} else if (wbc->nr_to_write)
/*
* There is no more writeout needed
* or we requested for a noblocking writeout
* and we found the device congested
*/
break;
}
blk_finish_plug(&plug);
if (!io_done && !cycled) {
cycled = 1;
index = 0;
wbc->range_start = index << PAGE_CACHE_SHIFT;
wbc->range_end = mapping->writeback_index - 1;
goto retry;
}
/* Update index */
wbc->range_cyclic = range_cyclic;
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
/*
* set the writeback_index so that range_cyclic
* mode will write it back later
*/
mapping->writeback_index = done_index;
out_writepages:
wbc->nr_to_write -= nr_to_writebump;
wbc->range_start = range_start;
trace_ext4_da_writepages_result(inode, wbc, ret, pages_written);
return ret;
}
#define FALL_BACK_TO_NONDELALLOC 1
static int ext4_nonda_switch(struct super_block *sb)
{
s64 free_blocks, dirty_blocks;
struct ext4_sb_info *sbi = EXT4_SB(sb);
/*
* switch to non delalloc mode if we are running low
* on free block. The free block accounting via percpu
* counters can get slightly wrong with percpu_counter_batch getting
* accumulated on each CPU without updating global counters
* Delalloc need an accurate free block accounting. So switch
* to non delalloc when we are near to error range.
*/
free_blocks = EXT4_C2B(sbi,
percpu_counter_read_positive(&sbi->s_freeclusters_counter));
dirty_blocks = percpu_counter_read_positive(&sbi->s_dirtyclusters_counter);
ext4: fix potential deadlock in ext4_nonda_switch() In ext4_nonda_switch(), if the file system is getting full we used to call writeback_inodes_sb_if_idle(). The problem is that we can be holding i_mutex already, and this causes a potential deadlock when writeback_inodes_sb_if_idle() when it tries to take s_umount. (See lockdep output below). As it turns out we don't need need to hold s_umount; the fact that we are in the middle of the write(2) system call will keep the superblock pinned. Unfortunately writeback_inodes_sb() checks to make sure s_umount is taken, and the VFS uses a different mechanism for making sure the file system doesn't get unmounted out from under us. The simplest way of dealing with this is to just simply grab s_umount using a trylock, and skip kicking the writeback flusher thread in the very unlikely case that we can't take a read lock on s_umount without blocking. Also, we now check the cirteria for kicking the writeback thread before we decide to whether to fall back to non-delayed writeback, so if there are any outstanding delayed allocation writes, we try to get them resolved as soon as possible. [ INFO: possible circular locking dependency detected ] 3.6.0-rc1-00042-gce894ca #367 Not tainted ------------------------------------------------------- dd/8298 is trying to acquire lock: (&type->s_umount_key#18){++++..}, at: [<c02277d4>] writeback_inodes_sb_if_idle+0x28/0x46 but task is already holding lock: (&sb->s_type->i_mutex_key#8){+.+...}, at: [<c01ddcce>] generic_file_aio_write+0x5f/0xd3 which lock already depends on the new lock. 2 locks held by dd/8298: #0: (sb_writers#2){.+.+.+}, at: [<c01ddcc5>] generic_file_aio_write+0x56/0xd3 #1: (&sb->s_type->i_mutex_key#8){+.+...}, at: [<c01ddcce>] generic_file_aio_write+0x5f/0xd3 stack backtrace: Pid: 8298, comm: dd Not tainted 3.6.0-rc1-00042-gce894ca #367 Call Trace: [<c015b79c>] ? console_unlock+0x345/0x372 [<c06d62a1>] print_circular_bug+0x190/0x19d [<c019906c>] __lock_acquire+0x86d/0xb6c [<c01999db>] ? mark_held_locks+0x5c/0x7b [<c0199724>] lock_acquire+0x66/0xb9 [<c02277d4>] ? writeback_inodes_sb_if_idle+0x28/0x46 [<c06db935>] down_read+0x28/0x58 [<c02277d4>] ? writeback_inodes_sb_if_idle+0x28/0x46 [<c02277d4>] writeback_inodes_sb_if_idle+0x28/0x46 [<c026f3b2>] ext4_nonda_switch+0xe1/0xf4 [<c0271ece>] ext4_da_write_begin+0x27/0x193 [<c01dcdb0>] generic_file_buffered_write+0xc8/0x1bb [<c01ddc47>] __generic_file_aio_write+0x1dd/0x205 [<c01ddce7>] generic_file_aio_write+0x78/0xd3 [<c026d336>] ext4_file_write+0x480/0x4a6 [<c0198c1d>] ? __lock_acquire+0x41e/0xb6c [<c0180944>] ? sched_clock_cpu+0x11a/0x13e [<c01967e9>] ? trace_hardirqs_off+0xb/0xd [<c018099f>] ? local_clock+0x37/0x4e [<c0209f2c>] do_sync_write+0x67/0x9d [<c0209ec5>] ? wait_on_retry_sync_kiocb+0x44/0x44 [<c020a7b9>] vfs_write+0x7b/0xe6 [<c020a9a6>] sys_write+0x3b/0x64 [<c06dd4bd>] syscall_call+0x7/0xb Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-09-20 02:42:36 +00:00
/*
* Start pushing delalloc when 1/2 of free blocks are dirty.
*/
if (dirty_blocks && (free_blocks < 2 * dirty_blocks) &&
!writeback_in_progress(sb->s_bdi) &&
down_read_trylock(&sb->s_umount)) {
writeback_inodes_sb(sb, WB_REASON_FS_FREE_SPACE);
up_read(&sb->s_umount);
}
if (2 * free_blocks < 3 * dirty_blocks ||
free_blocks < (dirty_blocks + EXT4_FREECLUSTERS_WATERMARK)) {
/*
* free block count is less than 150% of dirty blocks
* or free blocks is less than watermark
*/
return 1;
}
return 0;
}
static int ext4_da_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
{
int ret, retries = 0;
struct page *page;
pgoff_t index;
struct inode *inode = mapping->host;
handle_t *handle;
index = pos >> PAGE_CACHE_SHIFT;
if (ext4_nonda_switch(inode->i_sb)) {
*fsdata = (void *)FALL_BACK_TO_NONDELALLOC;
return ext4_write_begin(file, mapping, pos,
len, flags, pagep, fsdata);
}
*fsdata = (void *)0;
trace_ext4_da_write_begin(inode, pos, len, flags);
retry:
/*
* With delayed allocation, we don't log the i_disksize update
* if there is delayed block allocation. But we still need
* to journalling the i_disksize update if writes to the end
* of file which has an already mapped buffer.
*/
handle = ext4_journal_start(inode, 1);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
goto out;
}
/* We cannot recurse into the filesystem as the transaction is already
* started */
flags |= AOP_FLAG_NOFS;
fs: symlink write_begin allocation context fix With the write_begin/write_end aops, page_symlink was broken because it could no longer pass a GFP_NOFS type mask into the point where the allocations happened. They are done in write_begin, which would always assume that the filesystem can be entered from reclaim. This bug could cause filesystem deadlocks. The funny thing with having a gfp_t mask there is that it doesn't really allow the caller to arbitrarily tinker with the context in which it can be called. It couldn't ever be GFP_ATOMIC, for example, because it needs to take the page lock. The only thing any callers care about is __GFP_FS anyway, so turn that into a single flag. Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on this flag in their write_begin function. Change __grab_cache_page to accept a nofs argument as well, to honour that flag (while we're there, change the name to grab_cache_page_write_begin which is more instructive and does away with random leading underscores). This is really a more flexible way to go in the end anyway -- if a filesystem happens to want any extra allocations aside from the pagecache ones in ints write_begin function, it may now use GFP_KERNEL (rather than GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a random example). [kosaki.motohiro@jp.fujitsu.com: fix ubifs] [kosaki.motohiro@jp.fujitsu.com: fix fuse] Signed-off-by: Nick Piggin <npiggin@suse.de> Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: <stable@kernel.org> [2.6.28.x] Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Cleaned up the calling convention: just pass in the AOP flags untouched to the grab_cache_page_write_begin() function. That just simplifies everybody, and may even allow future expansion of the logic. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-04 20:00:53 +00:00
page = grab_cache_page_write_begin(mapping, index, flags);
if (!page) {
ext4_journal_stop(handle);
ret = -ENOMEM;
goto out;
}
*pagep = page;
ret = __block_write_begin(page, pos, len, ext4_da_get_block_prep);
if (ret < 0) {
unlock_page(page);
ext4_journal_stop(handle);
page_cache_release(page);
/*
* block_write_begin 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)
ext4_truncate_failed_write(inode);
}
if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
goto retry;
out:
return ret;
}
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
/*
* Check if we should update i_disksize
* when write to the end of file but not require block allocation
*/
static int ext4_da_should_update_i_disksize(struct page *page,
unsigned long offset)
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
{
struct buffer_head *bh;
struct inode *inode = page->mapping->host;
unsigned int idx;
int i;
bh = page_buffers(page);
idx = offset >> inode->i_blkbits;
for (i = 0; i < idx; i++)
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
bh = bh->b_this_page;
if (!buffer_mapped(bh) || (buffer_delay(bh)) || buffer_unwritten(bh))
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
return 0;
return 1;
}
static int ext4_da_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;
int ret = 0, ret2;
handle_t *handle = ext4_journal_current_handle();
loff_t new_i_size;
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
unsigned long start, end;
int write_mode = (int)(unsigned long)fsdata;
if (write_mode == FALL_BACK_TO_NONDELALLOC) {
ext4: ignore EXT4_INODE_JOURNAL_DATA flag with delalloc Ext4 does not support data journalling with delayed allocation enabled. We even do not allow to mount the file system with delayed allocation and data journalling enabled, however it can be set via FS_IOC_SETFLAGS so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file system mounted with delayed allocation (default) and that's where problem arises. The easies way to reproduce this problem is with the following set of commands: mkfs.ext4 /dev/sdd mount /dev/sdd /mnt/test1 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 chattr +j /mnt/test1/file dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc chattr -j /mnt/test1/file Additionally it can be reproduced quite reliably with xfstests 272 and 269. In fact the above reproducer is a part of test 272. To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if the file system is mounted with delayed allocation. This can be easily done by fixing ext4_should_*_data() functions do ignore data journal flag when delalloc is set (suggested by Ted). We also have to set the appropriate address space operations for the inode (again, ignoring data journal flag if delalloc enabled). Additionally this commit introduces ext4_inode_journal_mode() function because ext4_should_*_data() has already had a lot of common code and this change is putting it all into one function so it is easier to read. Successfully tested with xfstests in following configurations: delalloc + data=ordered delalloc + data=writeback data=journal nodelalloc + data=ordered nodelalloc + data=writeback nodelalloc + data=journal Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-02-20 22:53:00 +00:00
switch (ext4_inode_journal_mode(inode)) {
case EXT4_INODE_ORDERED_DATA_MODE:
return ext4_ordered_write_end(file, mapping, pos,
len, copied, page, fsdata);
ext4: ignore EXT4_INODE_JOURNAL_DATA flag with delalloc Ext4 does not support data journalling with delayed allocation enabled. We even do not allow to mount the file system with delayed allocation and data journalling enabled, however it can be set via FS_IOC_SETFLAGS so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file system mounted with delayed allocation (default) and that's where problem arises. The easies way to reproduce this problem is with the following set of commands: mkfs.ext4 /dev/sdd mount /dev/sdd /mnt/test1 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 chattr +j /mnt/test1/file dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc chattr -j /mnt/test1/file Additionally it can be reproduced quite reliably with xfstests 272 and 269. In fact the above reproducer is a part of test 272. To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if the file system is mounted with delayed allocation. This can be easily done by fixing ext4_should_*_data() functions do ignore data journal flag when delalloc is set (suggested by Ted). We also have to set the appropriate address space operations for the inode (again, ignoring data journal flag if delalloc enabled). Additionally this commit introduces ext4_inode_journal_mode() function because ext4_should_*_data() has already had a lot of common code and this change is putting it all into one function so it is easier to read. Successfully tested with xfstests in following configurations: delalloc + data=ordered delalloc + data=writeback data=journal nodelalloc + data=ordered nodelalloc + data=writeback nodelalloc + data=journal Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-02-20 22:53:00 +00:00
case EXT4_INODE_WRITEBACK_DATA_MODE:
return ext4_writeback_write_end(file, mapping, pos,
len, copied, page, fsdata);
ext4: ignore EXT4_INODE_JOURNAL_DATA flag with delalloc Ext4 does not support data journalling with delayed allocation enabled. We even do not allow to mount the file system with delayed allocation and data journalling enabled, however it can be set via FS_IOC_SETFLAGS so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file system mounted with delayed allocation (default) and that's where problem arises. The easies way to reproduce this problem is with the following set of commands: mkfs.ext4 /dev/sdd mount /dev/sdd /mnt/test1 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 chattr +j /mnt/test1/file dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc chattr -j /mnt/test1/file Additionally it can be reproduced quite reliably with xfstests 272 and 269. In fact the above reproducer is a part of test 272. To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if the file system is mounted with delayed allocation. This can be easily done by fixing ext4_should_*_data() functions do ignore data journal flag when delalloc is set (suggested by Ted). We also have to set the appropriate address space operations for the inode (again, ignoring data journal flag if delalloc enabled). Additionally this commit introduces ext4_inode_journal_mode() function because ext4_should_*_data() has already had a lot of common code and this change is putting it all into one function so it is easier to read. Successfully tested with xfstests in following configurations: delalloc + data=ordered delalloc + data=writeback data=journal nodelalloc + data=ordered nodelalloc + data=writeback nodelalloc + data=journal Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-02-20 22:53:00 +00:00
default:
BUG();
}
}
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
trace_ext4_da_write_end(inode, pos, len, copied);
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
start = pos & (PAGE_CACHE_SIZE - 1);
end = start + copied - 1;
/*
* generic_write_end() will run mark_inode_dirty() if i_size
* changes. So let's piggyback the i_disksize mark_inode_dirty
* into that.
*/
new_i_size = pos + copied;
ext4: avoid hangs in ext4_da_should_update_i_disksize() If the pte mapping in generic_perform_write() is unmapped between iov_iter_fault_in_readable() and iov_iter_copy_from_user_atomic(), the "copied" parameter to ->end_write can be zero. ext4 couldn't cope with it with delayed allocations enabled. This skips the i_disksize enlargement logic if copied is zero and no new data was appeneded to the inode. gdb> bt #0 0xffffffff811afe80 in ext4_da_should_update_i_disksize (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x1\ 08000, len=0x1000, copied=0x0, page=0xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2467 #1 ext4_da_write_end (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x108000, len=0x1000, copied=0x0, page=0\ xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2512 #2 0xffffffff810d97f1 in generic_perform_write (iocb=<value optimized out>, iov=<value optimized out>, nr_segs=<value o\ ptimized out>, pos=0x108000, ppos=0xffff88001e26be40, count=<value optimized out>, written=0x0) at mm/filemap.c:2440 #3 generic_file_buffered_write (iocb=<value optimized out>, iov=<value optimized out>, nr_segs=<value optimized out>, p\ os=0x108000, ppos=0xffff88001e26be40, count=<value optimized out>, written=0x0) at mm/filemap.c:2482 #4 0xffffffff810db5d1 in __generic_file_aio_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=0x1, ppos=0\ xffff88001e26be40) at mm/filemap.c:2600 #5 0xffffffff810db853 in generic_file_aio_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=<value optimi\ zed out>, pos=<value optimized out>) at mm/filemap.c:2632 #6 0xffffffff811a71aa in ext4_file_write (iocb=0xffff88001e26bde8, iov=0xffff88001e26bec8, nr_segs=0x1, pos=0x108000) a\ t fs/ext4/file.c:136 #7 0xffffffff811375aa in do_sync_write (filp=0xffff88003f606a80, buf=<value optimized out>, len=<value optimized out>, \ ppos=0xffff88001e26bf48) at fs/read_write.c:406 #8 0xffffffff81137e56 in vfs_write (file=0xffff88003f606a80, buf=0x1ec2960 <Address 0x1ec2960 out of bounds>, count=0x4\ 000, pos=0xffff88001e26bf48) at fs/read_write.c:435 #9 0xffffffff8113816c in sys_write (fd=<value optimized out>, buf=0x1ec2960 <Address 0x1ec2960 out of bounds>, count=0x\ 4000) at fs/read_write.c:487 #10 <signal handler called> #11 0x00007f120077a390 in __brk_reservation_fn_dmi_alloc__ () #12 0x0000000000000000 in ?? () gdb> print offset $22 = 0xffffffffffffffff gdb> print idx $23 = 0xffffffff gdb> print inode->i_blkbits $24 = 0xc gdb> up #1 ext4_da_write_end (file=0xffff88003f606a80, mapping=0xffff88001d3824e0, pos=0x108000, len=0x1000, copied=0x0, page=0\ xffffea0000d792e8, fsdata=0x0) at fs/ext4/inode.c:2512 2512 if (ext4_da_should_update_i_disksize(page, end)) { gdb> print start $25 = 0x0 gdb> print end $26 = 0xffffffffffffffff gdb> print pos $27 = 0x108000 gdb> print new_i_size $28 = 0x108000 gdb> print ((struct ext4_inode_info *)((char *)inode-((int)(&((struct ext4_inode_info *)0)->vfs_inode))))->i_disksize $29 = 0xd9000 gdb> down 2467 for (i = 0; i < idx; i++) gdb> print i $30 = 0xd44acbee This is 100% reproducible with some autonuma development code tuned in a very aggressive manner (not normal way even for knumad) which does "exotic" changes to the ptes. It wouldn't normally trigger but I don't see why it can't happen normally if the page is added to swap cache in between the two faults leading to "copied" being zero (which then hangs in ext4). So it should be fixed. Especially possible with lumpy reclaim (albeit disabled if compaction is enabled) as that would ignore the young bits in the ptes. Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@kernel.org
2011-12-14 02:41:15 +00:00
if (copied && new_i_size > EXT4_I(inode)->i_disksize) {
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
if (ext4_da_should_update_i_disksize(page, end)) {
down_write(&EXT4_I(inode)->i_data_sem);
if (new_i_size > EXT4_I(inode)->i_disksize)
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
EXT4_I(inode)->i_disksize = new_i_size;
up_write(&EXT4_I(inode)->i_data_sem);
/* We need to mark inode dirty even if
* new_i_size is less that inode->i_size
* bu greater than i_disksize.(hint delalloc)
*/
ext4_mark_inode_dirty(handle, inode);
}
ext4: fix delalloc i_disksize early update issue Ext4_da_write_end() used walk_page_buffers() with a callback function of ext4_bh_unmapped_or_delay() to check if it extended the file size without allocating any blocks (since in this case i_disksize needs to be updated). However, this is didn't work proprely because the buffer head has not been marked dirty yet --- this is done later in block_commit_write() --- which caused ext4_bh_unmapped_or_delay() to always return false. In addition, walk_page_buffers() checks all of the buffer heads covering the page, and the only buffer_head that should be checked is the one covering the end of the write. Otherwise, given a 1k blocksize filesystem and a 4k page size, the buffer head covering the first 1k stripe of the file could be unmapped (because it was a sparse file), and the second or third buffer_head covering that page could be mapped, and using walk_page_buffers() would fail in this case since it would stop at the first unmapped buffer_head and return true. The core problem is that walk_page_buffers() was intended to do work in a callback function, and a non-zero return value indicated a failure, which termined the walk of the buffer heads covering the page. It was not intended to be used with a boolean function, such as ext4_bh_unmapped_or_delay(). Add addtional fix from Aneesh to protect i_disksize update rave with truncate. Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2008-07-11 23:27:31 +00:00
}
ret2 = generic_write_end(file, mapping, pos, len, copied,
page, fsdata);
copied = ret2;
if (ret2 < 0)
ret = ret2;
ret2 = ext4_journal_stop(handle);
if (!ret)
ret = ret2;
return ret ? ret : copied;
}
static void ext4_da_invalidatepage(struct page *page, unsigned long offset)
{
/*
* Drop reserved blocks
*/
BUG_ON(!PageLocked(page));
if (!page_has_buffers(page))
goto out;
ext4_da_page_release_reservation(page, offset);
out:
ext4_invalidatepage(page, offset);
return;
}
/*
* Force all delayed allocation blocks to be allocated for a given inode.
*/
int ext4_alloc_da_blocks(struct inode *inode)
{
trace_ext4_alloc_da_blocks(inode);
if (!EXT4_I(inode)->i_reserved_data_blocks &&
!EXT4_I(inode)->i_reserved_meta_blocks)
return 0;
/*
* We do something simple for now. The filemap_flush() will
* also start triggering a write of the data blocks, which is
* not strictly speaking necessary (and for users of
* laptop_mode, not even desirable). However, to do otherwise
* would require replicating code paths in:
*
* ext4_da_writepages() ->
* write_cache_pages() ---> (via passed in callback function)
* __mpage_da_writepage() -->
* mpage_add_bh_to_extent()
* mpage_da_map_blocks()
*
* The problem is that write_cache_pages(), located in
* mm/page-writeback.c, marks pages clean in preparation for
* doing I/O, which is not desirable if we're not planning on
* doing I/O at all.
*
* We could call write_cache_pages(), and then redirty all of
* the pages by calling redirty_page_for_writepage() but that
* would be ugly in the extreme. So instead we would need to
* replicate parts of the code in the above functions,
* simplifying them because we wouldn't actually intend to
* write out the pages, but rather only collect contiguous
* logical block extents, call the multi-block allocator, and
* then update the buffer heads with the block allocations.
*
* For now, though, we'll cheat by calling filemap_flush(),
* which will map the blocks, and start the I/O, but not
* actually wait for the I/O to complete.
*/
return filemap_flush(inode->i_mapping);
}
/*
* bmap() is special. It gets used by applications such as lilo and by
* the swapper to find the on-disk block of a specific piece of data.
*
* Naturally, this is dangerous if the block concerned is still in the
* journal. If somebody makes a swapfile on an ext4 data-journaling
* filesystem and enables swap, then they may get a nasty shock when the
* data getting swapped to that swapfile suddenly gets overwritten by
* the original zero's written out previously to the journal and
* awaiting writeback in the kernel's buffer cache.
*
* So, if we see any bmap calls here on a modified, data-journaled file,
* take extra steps to flush any blocks which might be in the cache.
*/
static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
{
struct inode *inode = mapping->host;
journal_t *journal;
int err;
/*
* We can get here for an inline file via the FIBMAP ioctl
*/
if (ext4_has_inline_data(inode))
return 0;
if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY) &&
test_opt(inode->i_sb, DELALLOC)) {
/*
* With delalloc we want to sync the file
* so that we can make sure we allocate
* blocks for file
*/
filemap_write_and_wait(mapping);
}
if (EXT4_JOURNAL(inode) &&
ext4_test_inode_state(inode, EXT4_STATE_JDATA)) {
/*
* This is a REALLY heavyweight approach, but the use of
* bmap on dirty files is expected to be extremely rare:
* only if we run lilo or swapon on a freshly made file
* do we expect this to happen.
*
* (bmap requires CAP_SYS_RAWIO so this does not
* represent an unprivileged user DOS attack --- we'd be
* in trouble if mortal users could trigger this path at
* will.)
*
* NB. EXT4_STATE_JDATA is not set on files other than
* regular files. If somebody wants to bmap a directory
* or symlink and gets confused because the buffer
* hasn't yet been flushed to disk, they deserve
* everything they get.
*/
ext4_clear_inode_state(inode, EXT4_STATE_JDATA);
journal = EXT4_JOURNAL(inode);
jbd2_journal_lock_updates(journal);
err = jbd2_journal_flush(journal);
jbd2_journal_unlock_updates(journal);
if (err)
return 0;
}
return generic_block_bmap(mapping, block, ext4_get_block);
}
static int ext4_readpage(struct file *file, struct page *page)
{
int ret = -EAGAIN;
struct inode *inode = page->mapping->host;
trace_ext4_readpage(page);
if (ext4_has_inline_data(inode))
ret = ext4_readpage_inline(inode, page);
if (ret == -EAGAIN)
return mpage_readpage(page, ext4_get_block);
return ret;
}
static int
ext4_readpages(struct file *file, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
struct inode *inode = mapping->host;
/* If the file has inline data, no need to do readpages. */
if (ext4_has_inline_data(inode))
return 0;
return mpage_readpages(mapping, pages, nr_pages, ext4_get_block);
}
static void ext4_invalidatepage_free_endio(struct page *page, unsigned long offset)
{
struct buffer_head *head, *bh;
unsigned int curr_off = 0;
if (!page_has_buffers(page))
return;
head = bh = page_buffers(page);
do {
if (offset <= curr_off && test_clear_buffer_uninit(bh)
&& bh->b_private) {
ext4_free_io_end(bh->b_private);
bh->b_private = NULL;
bh->b_end_io = NULL;
}
curr_off = curr_off + bh->b_size;
bh = bh->b_this_page;
} while (bh != head);
}
static void ext4_invalidatepage(struct page *page, unsigned long offset)
{
journal_t *journal = EXT4_JOURNAL(page->mapping->host);
trace_ext4_invalidatepage(page, offset);
/*
* free any io_end structure allocated for buffers to be discarded
*/
if (ext4_should_dioread_nolock(page->mapping->host))
ext4_invalidatepage_free_endio(page, offset);
/*
* If it's a full truncate we just forget about the pending dirtying
*/
if (offset == 0)
ClearPageChecked(page);
if (journal)
jbd2_journal_invalidatepage(journal, page, offset);
else
block_invalidatepage(page, offset);
}
static int ext4_releasepage(struct page *page, gfp_t wait)
{
journal_t *journal = EXT4_JOURNAL(page->mapping->host);
trace_ext4_releasepage(page);
WARN_ON(PageChecked(page));
if (!page_has_buffers(page))
return 0;
if (journal)
return jbd2_journal_try_to_free_buffers(journal, page, wait);
else
return try_to_free_buffers(page);
}
/*
* ext4_get_block used when preparing for a DIO write or buffer write.
* We allocate an uinitialized extent if blocks haven't been allocated.
* The extent will be converted to initialized after the IO is complete.
*/
int ext4_get_block_write(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
ext4_debug("ext4_get_block_write: inode %lu, create flag %d\n",
inode->i_ino, create);
return _ext4_get_block(inode, iblock, bh_result,
EXT4_GET_BLOCKS_IO_CREATE_EXT);
}
static int ext4_get_block_write_nolock(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
ext4_debug("ext4_get_block_write_nolock: inode %lu, create flag %d\n",
inode->i_ino, create);
return _ext4_get_block(inode, iblock, bh_result,
EXT4_GET_BLOCKS_NO_LOCK);
}
static void ext4_end_io_dio(struct kiocb *iocb, loff_t offset,
ssize_t size, void *private, int ret,
bool is_async)
{
struct inode *inode = iocb->ki_filp->f_path.dentry->d_inode;
ext4_io_end_t *io_end = iocb->private;
/* if not async direct IO or dio with 0 bytes write, just return */
if (!io_end || !size)
goto out;
ext_debug("ext4_end_io_dio(): io_end 0x%p "
"for inode %lu, iocb 0x%p, offset %llu, size %zd\n",
iocb->private, io_end->inode->i_ino, iocb, offset,
size);
iocb->private = NULL;
/* if not aio dio with unwritten extents, just free io and return */
if (!(io_end->flag & EXT4_IO_END_UNWRITTEN)) {
ext4_free_io_end(io_end);
out:
if (is_async)
aio_complete(iocb, ret, 0);
inode_dio_done(inode);
return;
}
io_end->offset = offset;
io_end->size = size;
if (is_async) {
io_end->iocb = iocb;
io_end->result = ret;
}
ext4: completed_io locking cleanup Current unwritten extent conversion state-machine is very fuzzy. - For unknown reason it performs conversion under i_mutex. What for? My diagnosis: We already protect extent tree with i_data_sem, truncate and punch_hole should wait for DIO, so the only data we have to protect is end_io->flags modification, but only flush_completed_IO and end_io_work modified this flags and we can serialize them via i_completed_io_lock. Currently all these games with mutex_trylock result in the following deadlock truncate: kworker: ext4_setattr ext4_end_io_work mutex_lock(i_mutex) inode_dio_wait(inode) ->BLOCK DEADLOCK<- mutex_trylock() inode_dio_done() #TEST_CASE1_BEGIN MNT=/mnt_scrach unlink $MNT/file fallocate -l $((1024*1024*1024)) $MNT/file aio-stress -I 100000 -O -s 100m -n -t 1 -c 10 -o 2 -o 3 $MNT/file sleep 2 truncate -s 0 $MNT/file #TEST_CASE1_END Or use 286's xfstests https://github.com/dmonakhov/xfstests/blob/devel/286 This patch makes state machine simple and clean: (1) xxx_end_io schedule final extent conversion simply by calling ext4_add_complete_io(), which append it to ei->i_completed_io_list NOTE1: because of (2A) work should be queued only if ->i_completed_io_list was empty, otherwise the work is scheduled already. (2) ext4_flush_completed_IO is responsible for handling all pending end_io from ei->i_completed_io_list Flushing sequence consists of following stages: A) LOCKED: Atomically drain completed_io_list to local_list B) Perform extents conversion C) LOCKED: move converted io's to to_free list for final deletion This logic depends on context which we was called from. D) Final end_io context destruction NOTE1: i_mutex is no longer required because end_io->flags modification is protected by ei->ext4_complete_io_lock Full list of changes: - Move all completion end_io related routines to page-io.c in order to improve logic locality - Move open coded logic from various xx_end_xx routines to ext4_add_complete_io() - remove EXT4_IO_END_FSYNC - Improve SMP scalability by removing useless i_mutex which does not protect io->flags anymore. - Reduce lock contention on i_completed_io_lock by optimizing list walk. - Rename ext4_end_io_nolock to end4_end_io and make it static - Check flush completion status to ext4_ext_punch_hole(). Because it is not good idea to punch blocks from corrupted inode. Changes since V3 (in request to Jan's comments): Fall back to active flush_completed_IO() approach in order to prevent performance issues with nolocked DIO reads. Changes since V2: Fix use-after-free caused by race truncate vs end_io_work Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-09-29 04:14:55 +00:00
ext4_add_complete_io(io_end);
}
static void ext4_end_io_buffer_write(struct buffer_head *bh, int uptodate)
{
ext4_io_end_t *io_end = bh->b_private;
struct inode *inode;
if (!test_clear_buffer_uninit(bh) || !io_end)
goto out;
if (!(io_end->inode->i_sb->s_flags & MS_ACTIVE)) {
ext4_msg(io_end->inode->i_sb, KERN_INFO,
"sb umounted, discard end_io request for inode %lu",
io_end->inode->i_ino);
ext4_free_io_end(io_end);
goto out;
}
/*
* It may be over-defensive here to check EXT4_IO_END_UNWRITTEN now,
* but being more careful is always safe for the future change.
*/
inode = io_end->inode;
ext4_set_io_unwritten_flag(inode, io_end);
ext4: completed_io locking cleanup Current unwritten extent conversion state-machine is very fuzzy. - For unknown reason it performs conversion under i_mutex. What for? My diagnosis: We already protect extent tree with i_data_sem, truncate and punch_hole should wait for DIO, so the only data we have to protect is end_io->flags modification, but only flush_completed_IO and end_io_work modified this flags and we can serialize them via i_completed_io_lock. Currently all these games with mutex_trylock result in the following deadlock truncate: kworker: ext4_setattr ext4_end_io_work mutex_lock(i_mutex) inode_dio_wait(inode) ->BLOCK DEADLOCK<- mutex_trylock() inode_dio_done() #TEST_CASE1_BEGIN MNT=/mnt_scrach unlink $MNT/file fallocate -l $((1024*1024*1024)) $MNT/file aio-stress -I 100000 -O -s 100m -n -t 1 -c 10 -o 2 -o 3 $MNT/file sleep 2 truncate -s 0 $MNT/file #TEST_CASE1_END Or use 286's xfstests https://github.com/dmonakhov/xfstests/blob/devel/286 This patch makes state machine simple and clean: (1) xxx_end_io schedule final extent conversion simply by calling ext4_add_complete_io(), which append it to ei->i_completed_io_list NOTE1: because of (2A) work should be queued only if ->i_completed_io_list was empty, otherwise the work is scheduled already. (2) ext4_flush_completed_IO is responsible for handling all pending end_io from ei->i_completed_io_list Flushing sequence consists of following stages: A) LOCKED: Atomically drain completed_io_list to local_list B) Perform extents conversion C) LOCKED: move converted io's to to_free list for final deletion This logic depends on context which we was called from. D) Final end_io context destruction NOTE1: i_mutex is no longer required because end_io->flags modification is protected by ei->ext4_complete_io_lock Full list of changes: - Move all completion end_io related routines to page-io.c in order to improve logic locality - Move open coded logic from various xx_end_xx routines to ext4_add_complete_io() - remove EXT4_IO_END_FSYNC - Improve SMP scalability by removing useless i_mutex which does not protect io->flags anymore. - Reduce lock contention on i_completed_io_lock by optimizing list walk. - Rename ext4_end_io_nolock to end4_end_io and make it static - Check flush completion status to ext4_ext_punch_hole(). Because it is not good idea to punch blocks from corrupted inode. Changes since V3 (in request to Jan's comments): Fall back to active flush_completed_IO() approach in order to prevent performance issues with nolocked DIO reads. Changes since V2: Fix use-after-free caused by race truncate vs end_io_work Signed-off-by: Dmitry Monakhov <dmonakhov@openvz.org> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2012-09-29 04:14:55 +00:00
ext4_add_complete_io(io_end);
out:
bh->b_private = NULL;
bh->b_end_io = NULL;
clear_buffer_uninit(bh);
end_buffer_async_write(bh, uptodate);
}
static int ext4_set_bh_endio(struct buffer_head *bh, struct inode *inode)
{
ext4_io_end_t *io_end;
struct page *page = bh->b_page;
loff_t offset = (sector_t)page->index << PAGE_CACHE_SHIFT;
size_t size = bh->b_size;
retry:
io_end = ext4_init_io_end(inode, GFP_ATOMIC);
if (!io_end) {
pr_warn_ratelimited("%s: allocation fail\n", __func__);
schedule();
goto retry;
}
io_end->offset = offset;
io_end->size = size;
/*
* We need to hold a reference to the page to make sure it
* doesn't get evicted before ext4_end_io_work() has a chance
* to convert the extent from written to unwritten.
*/
io_end->page = page;
get_page(io_end->page);
bh->b_private = io_end;
bh->b_end_io = ext4_end_io_buffer_write;
return 0;
}
/*
* For ext4 extent files, ext4 will do direct-io write to holes,
* preallocated extents, and those write extend the file, no need to
* fall back to buffered IO.
*
* For holes, we fallocate those blocks, mark them as uninitialized
* If those blocks were preallocated, we mark sure they are split, but
* still keep the range to write as uninitialized.
*
* The unwritten extents will be converted to written when DIO is completed.
* For async direct IO, since the IO may still pending when return, we
* set up an end_io call back function, which will do the conversion
* when async direct IO completed.
*
* If the O_DIRECT write will extend the file then add this inode to the
* orphan list. So recovery will truncate it back to the original size
* if the machine crashes during the write.
*
*/
static ssize_t ext4_ext_direct_IO(int rw, struct kiocb *iocb,
const struct iovec *iov, loff_t offset,
unsigned long nr_segs)
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
ssize_t ret;
size_t count = iov_length(iov, nr_segs);
int overwrite = 0;
get_block_t *get_block_func = NULL;
int dio_flags = 0;
loff_t final_size = offset + count;
/* Use the old path for reads and writes beyond i_size. */
if (rw != WRITE || final_size > inode->i_size)
return ext4_ind_direct_IO(rw, iocb, iov, offset, nr_segs);
BUG_ON(iocb->private == NULL);
/* If we do a overwrite dio, i_mutex locking can be released */
overwrite = *((int *)iocb->private);
if (overwrite) {
atomic_inc(&inode->i_dio_count);
down_read(&EXT4_I(inode)->i_data_sem);
mutex_unlock(&inode->i_mutex);
}
/*
* We could direct write to holes and fallocate.
*
* Allocated blocks to fill the hole are marked as
* uninitialized to prevent parallel buffered read to expose
* the stale data before DIO complete the data IO.
*
* As to previously fallocated extents, ext4 get_block will
* just simply mark the buffer mapped but still keep the
* extents uninitialized.
*
* For non AIO case, we will convert those unwritten extents
* to written after return back from blockdev_direct_IO.
*
* For async DIO, the conversion needs to be deferred when the
* IO is completed. The ext4 end_io callback function will be
* called to take care of the conversion work. Here for async
* case, we allocate an io_end structure to hook to the iocb.
*/
iocb->private = NULL;
ext4_inode_aio_set(inode, NULL);
if (!is_sync_kiocb(iocb)) {
ext4_io_end_t *io_end = ext4_init_io_end(inode, GFP_NOFS);
if (!io_end) {
ret = -ENOMEM;
goto retake_lock;
}
io_end->flag |= EXT4_IO_END_DIRECT;
iocb->private = io_end;
/*
* we save the io structure for current async direct
* IO, so that later ext4_map_blocks() could flag the
* io structure whether there is a unwritten extents
* needs to be converted when IO is completed.
*/
ext4_inode_aio_set(inode, io_end);
}
if (overwrite) {
get_block_func = ext4_get_block_write_nolock;
} else {
get_block_func = ext4_get_block_write;
dio_flags = DIO_LOCKING;
}
ret = __blockdev_direct_IO(rw, iocb, inode,
inode->i_sb->s_bdev, iov,
offset, nr_segs,
get_block_func,
ext4_end_io_dio,
NULL,
dio_flags);
if (iocb->private)
ext4_inode_aio_set(inode, NULL);
/*
* The io_end structure takes a reference to the inode, that
* structure needs to be destroyed and the reference to the
* inode need to be dropped, when IO is complete, even with 0
* byte write, or failed.
*
* In the successful AIO DIO case, the io_end structure will
* be destroyed and the reference to the inode will be dropped
* after the end_io call back function is called.
*
* In the case there is 0 byte write, or error case, since VFS
* direct IO won't invoke the end_io call back function, we
* need to free the end_io structure here.
*/
if (ret != -EIOCBQUEUED && ret <= 0 && iocb->private) {
ext4_free_io_end(iocb->private);
iocb->private = NULL;
} else if (ret > 0 && !overwrite && ext4_test_inode_state(inode,
EXT4_STATE_DIO_UNWRITTEN)) {
int err;
/*
* for non AIO case, since the IO is already
* completed, we could do the conversion right here
*/
err = ext4_convert_unwritten_extents(inode,
offset, ret);
if (err < 0)
ret = err;
ext4_clear_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
}
retake_lock:
/* take i_mutex locking again if we do a ovewrite dio */
if (overwrite) {
inode_dio_done(inode);
up_read(&EXT4_I(inode)->i_data_sem);
mutex_lock(&inode->i_mutex);
}
return ret;
}
static ssize_t ext4_direct_IO(int rw, struct kiocb *iocb,
const struct iovec *iov, loff_t offset,
unsigned long nr_segs)
{
struct file *file = iocb->ki_filp;
struct inode *inode = file->f_mapping->host;
ssize_t ret;
/*
* If we are doing data journalling we don't support O_DIRECT
*/
if (ext4_should_journal_data(inode))
return 0;
/* Let buffer I/O handle the inline data case. */
if (ext4_has_inline_data(inode))
return 0;
trace_ext4_direct_IO_enter(inode, offset, iov_length(iov, nr_segs), rw);
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
ret = ext4_ext_direct_IO(rw, iocb, iov, offset, nr_segs);
else
ret = ext4_ind_direct_IO(rw, iocb, iov, offset, nr_segs);
trace_ext4_direct_IO_exit(inode, offset,
iov_length(iov, nr_segs), rw, ret);
return ret;
}
/*
* Pages can be marked dirty completely asynchronously from ext4's journalling
* activity. By filemap_sync_pte(), try_to_unmap_one(), etc. We cannot do
* much here because ->set_page_dirty is called under VFS locks. The page is
* not necessarily locked.
*
* We cannot just dirty the page and leave attached buffers clean, because the
* buffers' dirty state is "definitive". We cannot just set the buffers dirty
* or jbddirty because all the journalling code will explode.
*
* So what we do is to mark the page "pending dirty" and next time writepage
* is called, propagate that into the buffers appropriately.
*/
static int ext4_journalled_set_page_dirty(struct page *page)
{
SetPageChecked(page);
return __set_page_dirty_nobuffers(page);
}
static const struct address_space_operations ext4_ordered_aops = {
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.readpage = ext4_readpage,
.readpages = ext4_readpages,
.writepage = ext4_writepage,
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.write_begin = ext4_write_begin,
.write_end = ext4_ordered_write_end,
.bmap = ext4_bmap,
.invalidatepage = ext4_invalidatepage,
.releasepage = ext4_releasepage,
.direct_IO = ext4_direct_IO,
.migratepage = buffer_migrate_page,
.is_partially_uptodate = block_is_partially_uptodate,
.error_remove_page = generic_error_remove_page,
};
static const struct address_space_operations ext4_writeback_aops = {
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.readpage = ext4_readpage,
.readpages = ext4_readpages,
.writepage = ext4_writepage,
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.write_begin = ext4_write_begin,
.write_end = ext4_writeback_write_end,
.bmap = ext4_bmap,
.invalidatepage = ext4_invalidatepage,
.releasepage = ext4_releasepage,
.direct_IO = ext4_direct_IO,
.migratepage = buffer_migrate_page,
.is_partially_uptodate = block_is_partially_uptodate,
.error_remove_page = generic_error_remove_page,
};
static const struct address_space_operations ext4_journalled_aops = {
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.readpage = ext4_readpage,
.readpages = ext4_readpages,
.writepage = ext4_writepage,
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.write_begin = ext4_write_begin,
.write_end = ext4_journalled_write_end,
.set_page_dirty = ext4_journalled_set_page_dirty,
.bmap = ext4_bmap,
.invalidatepage = ext4_invalidatepage,
.releasepage = ext4_releasepage,
.direct_IO = ext4_direct_IO,
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.is_partially_uptodate = block_is_partially_uptodate,
.error_remove_page = generic_error_remove_page,
};
static const struct address_space_operations ext4_da_aops = {
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.readpage = ext4_readpage,
.readpages = ext4_readpages,
.writepage = ext4_writepage,
vfs: pagecache usage optimization for pagesize!=blocksize When we read some part of a file through pagecache, if there is a pagecache of corresponding index but this page is not uptodate, read IO is issued and this page will be uptodate. I think this is good for pagesize == blocksize environment but there is room for improvement on pagesize != blocksize environment. Because in this case a page can have multiple buffers and even if a page is not uptodate, some buffers can be uptodate. So I suggest that when all buffers which correspond to a part of a file that we want to read are uptodate, use this pagecache and copy data from this pagecache to user buffer even if a page is not uptodate. This can reduce read IO and improve system throughput. I wrote a benchmark program and got result number with this program. This benchmark do: 1: mount and open a test file. 2: create a 512MB file. 3: close a file and umount. 4: mount and again open a test file. 5: pwrite randomly 300000 times on a test file. offset is aligned by IO size(1024bytes). 6: measure time of preading randomly 100000 times on a test file. The result was: 2.6.26 330 sec 2.6.26-patched 226 sec Arch:i386 Filesystem:ext3 Blocksize:1024 bytes Memory: 1GB On ext3/4, a file is written through buffer/block. So random read/write mixed workloads or random read after random write workloads are optimized with this patch under pagesize != blocksize environment. This test result showed this. The benchmark program is as follows: #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <unistd.h> #include <time.h> #include <stdlib.h> #include <string.h> #include <sys/mount.h> #define LEN 1024 #define LOOP 1024*512 /* 512MB */ main(void) { unsigned long i, offset, filesize; int fd; char buf[LEN]; time_t t1, t2; if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } memset(buf, 0, LEN); fd = open("/root/test1/testfile", O_CREAT|O_RDWR|O_TRUNC); if (fd < 0) { perror("cannot open file\n"); exit(1); } for (i = 0; i < LOOP; i++) write(fd, buf, LEN); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } if (mount("/dev/sda1", "/root/test1/", "ext3", 0, 0) < 0) { perror("cannot mount\n"); exit(1); } fd = open("/root/test1/testfile", O_RDWR); if (fd < 0) { perror("cannot open file\n"); exit(1); } filesize = LEN * LOOP; for (i = 0; i < 300000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pwrite(fd, buf, LEN, offset); } printf("start test\n"); time(&t1); for (i = 0; i < 100000; i++){ offset = (random() % filesize) & (~(LEN - 1)); pread(fd, buf, LEN, offset); } time(&t2); printf("%ld sec\n", t2-t1); close(fd); if (umount("/root/test1/") < 0) { perror("cannot umount\n"); exit(1); } } Signed-off-by: Hisashi Hifumi <hifumi.hisashi@oss.ntt.co.jp> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Christoph Hellwig <hch@infradead.org> Cc: Jan Kara <jack@ucw.cz> Cc: <linux-ext4@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-07-28 22:46:36 +00:00
.writepages = ext4_da_writepages,
.write_begin = ext4_da_write_begin,
.write_end = ext4_da_write_end,
.bmap = ext4_bmap,
.invalidatepage = ext4_da_invalidatepage,
.releasepage = ext4_releasepage,
.direct_IO = ext4_direct_IO,
.migratepage = buffer_migrate_page,
.is_partially_uptodate = block_is_partially_uptodate,
.error_remove_page = generic_error_remove_page,
};
void ext4_set_aops(struct inode *inode)
{
ext4: ignore EXT4_INODE_JOURNAL_DATA flag with delalloc Ext4 does not support data journalling with delayed allocation enabled. We even do not allow to mount the file system with delayed allocation and data journalling enabled, however it can be set via FS_IOC_SETFLAGS so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file system mounted with delayed allocation (default) and that's where problem arises. The easies way to reproduce this problem is with the following set of commands: mkfs.ext4 /dev/sdd mount /dev/sdd /mnt/test1 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 chattr +j /mnt/test1/file dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc chattr -j /mnt/test1/file Additionally it can be reproduced quite reliably with xfstests 272 and 269. In fact the above reproducer is a part of test 272. To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if the file system is mounted with delayed allocation. This can be easily done by fixing ext4_should_*_data() functions do ignore data journal flag when delalloc is set (suggested by Ted). We also have to set the appropriate address space operations for the inode (again, ignoring data journal flag if delalloc enabled). Additionally this commit introduces ext4_inode_journal_mode() function because ext4_should_*_data() has already had a lot of common code and this change is putting it all into one function so it is easier to read. Successfully tested with xfstests in following configurations: delalloc + data=ordered delalloc + data=writeback data=journal nodelalloc + data=ordered nodelalloc + data=writeback nodelalloc + data=journal Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-02-20 22:53:00 +00:00
switch (ext4_inode_journal_mode(inode)) {
case EXT4_INODE_ORDERED_DATA_MODE:
if (test_opt(inode->i_sb, DELALLOC))
inode->i_mapping->a_ops = &ext4_da_aops;
else
inode->i_mapping->a_ops = &ext4_ordered_aops;
break;
case EXT4_INODE_WRITEBACK_DATA_MODE:
if (test_opt(inode->i_sb, DELALLOC))
inode->i_mapping->a_ops = &ext4_da_aops;
else
inode->i_mapping->a_ops = &ext4_writeback_aops;
break;
case EXT4_INODE_JOURNAL_DATA_MODE:
inode->i_mapping->a_ops = &ext4_journalled_aops;
ext4: ignore EXT4_INODE_JOURNAL_DATA flag with delalloc Ext4 does not support data journalling with delayed allocation enabled. We even do not allow to mount the file system with delayed allocation and data journalling enabled, however it can be set via FS_IOC_SETFLAGS so we can hit the inode with EXT4_INODE_JOURNAL_DATA set even on file system mounted with delayed allocation (default) and that's where problem arises. The easies way to reproduce this problem is with the following set of commands: mkfs.ext4 /dev/sdd mount /dev/sdd /mnt/test1 dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 chattr +j /mnt/test1/file dd if=/dev/zero of=/mnt/test1/file bs=1M count=4 conv=notrunc chattr -j /mnt/test1/file Additionally it can be reproduced quite reliably with xfstests 272 and 269. In fact the above reproducer is a part of test 272. To fix this we should ignore the EXT4_INODE_JOURNAL_DATA inode flag if the file system is mounted with delayed allocation. This can be easily done by fixing ext4_should_*_data() functions do ignore data journal flag when delalloc is set (suggested by Ted). We also have to set the appropriate address space operations for the inode (again, ignoring data journal flag if delalloc enabled). Additionally this commit introduces ext4_inode_journal_mode() function because ext4_should_*_data() has already had a lot of common code and this change is putting it all into one function so it is easier to read. Successfully tested with xfstests in following configurations: delalloc + data=ordered delalloc + data=writeback data=journal nodelalloc + data=ordered nodelalloc + data=writeback nodelalloc + data=journal Signed-off-by: Lukas Czerner <lczerner@redhat.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: stable@vger.kernel.org
2012-02-20 22:53:00 +00:00
break;
default:
BUG();
}
}
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
/*
* ext4_discard_partial_page_buffers()
* Wrapper function for ext4_discard_partial_page_buffers_no_lock.
* This function finds and locks the page containing the offset
* "from" and passes it to ext4_discard_partial_page_buffers_no_lock.
* Calling functions that already have the page locked should call
* ext4_discard_partial_page_buffers_no_lock directly.
*/
int ext4_discard_partial_page_buffers(handle_t *handle,
struct address_space *mapping, loff_t from,
loff_t length, int flags)
{
struct inode *inode = mapping->host;
struct page *page;
int err = 0;
page = find_or_create_page(mapping, from >> PAGE_CACHE_SHIFT,
mapping_gfp_mask(mapping) & ~__GFP_FS);
if (!page)
return -ENOMEM;
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
err = ext4_discard_partial_page_buffers_no_lock(handle, inode, page,
from, length, flags);
unlock_page(page);
page_cache_release(page);
return err;
}
/*
* ext4_discard_partial_page_buffers_no_lock()
* Zeros a page range of length 'length' starting from offset 'from'.
* Buffer heads that correspond to the block aligned regions of the
* zeroed range will be unmapped. Unblock aligned regions
* will have the corresponding buffer head mapped if needed so that
* that region of the page can be updated with the partial zero out.
*
* This function assumes that the page has already been locked. The
* The range to be discarded must be contained with in the given page.
* If the specified range exceeds the end of the page it will be shortened
* to the end of the page that corresponds to 'from'. This function is
* appropriate for updating a page and it buffer heads to be unmapped and
* zeroed for blocks that have been either released, or are going to be
* released.
*
* handle: The journal handle
* inode: The files inode
* page: A locked page that contains the offset "from"
* from: The starting byte offset (from the beginning of the file)
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
* to begin discarding
* len: The length of bytes to discard
* flags: Optional flags that may be used:
*
* EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED
* Only zero the regions of the page whose buffer heads
* have already been unmapped. This flag is appropriate
* for updating the contents of a page whose blocks may
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
* have already been released, and we only want to zero
* out the regions that correspond to those released blocks.
*
* Returns zero on success or negative on failure.
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
*/
static int ext4_discard_partial_page_buffers_no_lock(handle_t *handle,
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
struct inode *inode, struct page *page, loff_t from,
loff_t length, int flags)
{
ext4_fsblk_t index = from >> PAGE_CACHE_SHIFT;
unsigned int offset = from & (PAGE_CACHE_SIZE-1);
unsigned int blocksize, max, pos;
ext4_lblk_t iblock;
struct buffer_head *bh;
int err = 0;
blocksize = inode->i_sb->s_blocksize;
max = PAGE_CACHE_SIZE - offset;
if (index != page->index)
return -EINVAL;
/*
* correct length if it does not fall between
* 'from' and the end of the page
*/
if (length > max || length < 0)
length = max;
iblock = index << (PAGE_CACHE_SHIFT - inode->i_sb->s_blocksize_bits);
if (!page_has_buffers(page))
create_empty_buffers(page, blocksize, 0);
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
/* Find the buffer that contains "offset" */
bh = page_buffers(page);
pos = blocksize;
while (offset >= pos) {
bh = bh->b_this_page;
iblock++;
pos += blocksize;
}
pos = offset;
while (pos < offset + length) {
unsigned int end_of_block, range_to_discard;
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
err = 0;
/* The length of space left to zero and unmap */
range_to_discard = offset + length - pos;
/* The length of space until the end of the block */
end_of_block = blocksize - (pos & (blocksize-1));
/*
* Do not unmap or zero past end of block
* for this buffer head
*/
if (range_to_discard > end_of_block)
range_to_discard = end_of_block;
/*
* Skip this buffer head if we are only zeroing unampped
* regions of the page
*/
if (flags & EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED &&
buffer_mapped(bh))
goto next;
/* If the range is block aligned, unmap */
if (range_to_discard == blocksize) {
clear_buffer_dirty(bh);
bh->b_bdev = NULL;
clear_buffer_mapped(bh);
clear_buffer_req(bh);
clear_buffer_new(bh);
clear_buffer_delay(bh);
clear_buffer_unwritten(bh);
clear_buffer_uptodate(bh);
zero_user(page, pos, range_to_discard);
BUFFER_TRACE(bh, "Buffer discarded");
goto next;
}
/*
* If this block is not completely contained in the range
* to be discarded, then it is not going to be released. Because
* we need to keep this block, we need to make sure this part
* of the page is uptodate before we modify it by writeing
* partial zeros on it.
*/
if (!buffer_mapped(bh)) {
/*
* Buffer head must be mapped before we can read
* from the block
*/
BUFFER_TRACE(bh, "unmapped");
ext4_get_block(inode, iblock, bh, 0);
/* unmapped? It's a hole - nothing to do */
if (!buffer_mapped(bh)) {
BUFFER_TRACE(bh, "still unmapped");
goto next;
}
}
/* Ok, it's mapped. Make sure it's up-to-date */
if (PageUptodate(page))
set_buffer_uptodate(bh);
if (!buffer_uptodate(bh)) {
err = -EIO;
ll_rw_block(READ, 1, &bh);
wait_on_buffer(bh);
/* Uhhuh. Read error. Complain and punt.*/
if (!buffer_uptodate(bh))
goto next;
}
if (ext4_should_journal_data(inode)) {
BUFFER_TRACE(bh, "get write access");
err = ext4_journal_get_write_access(handle, bh);
if (err)
goto next;
}
zero_user(page, pos, range_to_discard);
err = 0;
if (ext4_should_journal_data(inode)) {
err = ext4_handle_dirty_metadata(handle, inode, bh);
ext4: only call ext4_jbd2_file_inode when an inode has been extended In delayed allocation mode, it's important to only call ext4_jbd2_file_inode when the file has been extended. This is necessary to avoid a race which first got introduced in commit 678aaf481, but which was made much more common with the introduction of the "punch hole" functionality. (Especially when dioread_nolock was enabled; when I could reliably reproduce this problem with xfstests #74.) The race is this: If while trying to writeback a delayed allocation inode, there is a need to map delalloc blocks, and we run out of space in the journal, *and* at the same time the inode is already on the committing transaction's t_inode_list (because for example while doing the punch hole operation, ext4_jbd2_file_inode() is called), then the commit operation will wait for the inode to finish all of its pending writebacks by calling filemap_fdatawait(), but since that inode has one or more pages with the PageWriteback flag set, the commit operation will wait forever, and the so the writeback of the inode can never take place, and the kjournald thread and the writeback thread end up waiting for each other --- forever. It's important at this point to recall why an inode is placed on the t_inode_list; it is to provide the data=ordered guarantees that we don't end up exposing stale data. In the case where we are truncating or punching a hole in the inode, there is no possibility that stale data could be exposed in the first place, so we don't need to put the inode on the t_inode_list! The right long-term fix is to get rid of data=ordered mode altogether, and only update the extent tree or indirect blocks after the data has been written. Until then, this change will also avoid some unnecessary waiting in the commit operation. Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Cc: Allison Henderson <achender@linux.vnet.ibm.com> Cc: Jan Kara <jack@suse.cz>
2011-09-06 06:37:06 +00:00
} else
ext4: Add new ext4_discard_partial_page_buffers routines This patch adds two new routines: ext4_discard_partial_page_buffers and ext4_discard_partial_page_buffers_no_lock. The ext4_discard_partial_page_buffers routine is a wrapper function to ext4_discard_partial_page_buffers_no_lock. The wrapper function locks the page and passes it to ext4_discard_partial_page_buffers_no_lock. Calling functions that already have the page locked can call ext4_discard_partial_page_buffers_no_lock directly. The ext4_discard_partial_page_buffers_no_lock function zeros a specified range in a page, and unmaps the corresponding buffer heads. Only block aligned regions of the page will have their buffer heads unmapped. Unblock aligned regions will be mapped if needed so that they can be updated with the partial zero out. This function is meant to be used to update a page and its buffer heads to be zeroed and unmapped when the corresponding blocks have been released or will be released. This routine is used in the following scenarios: * A hole is punched and the non page aligned regions of the head and tail of the hole need to be discarded * The file is truncated and the partial page beyond EOF needs to be discarded * The end of a hole is in the same page as EOF. After the page is flushed, the partial page beyond EOF needs to be discarded. * A write operation begins or ends inside a hole and the partial page appearing before or after the write needs to be discarded * A write operation extends EOF and the partial page beyond EOF needs to be discarded This function takes a flag EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED which is used when a write operation begins or ends in a hole. When the EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED flag is used, only buffer heads that are already unmapped will have the corresponding regions of the page zeroed. Signed-off-by: Allison Henderson <achender@linux.vnet.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-09-03 15:51:09 +00:00
mark_buffer_dirty(bh);
BUFFER_TRACE(bh, "Partial buffer zeroed");
next:
bh = bh->b_this_page;
iblock++;
pos += range_to_discard;
}
return err;
}
int ext4_can_truncate(struct inode *inode)
{
if (S_ISREG(inode->i_mode))
return 1;
if (S_ISDIR(inode->i_mode))
return 1;
if (S_ISLNK(inode->i_mode))
return !ext4_inode_is_fast_symlink(inode);
return 0;
}
/*
* ext4_punch_hole: punches a hole in a file by releaseing the blocks
* associated with the given offset and length
*
* @inode: File inode
* @offset: The offset where the hole will begin
* @len: The length of the hole
*
* Returns: 0 on success or negative on failure
*/
int ext4_punch_hole(struct file *file, loff_t offset, loff_t length)
{
struct inode *inode = file->f_path.dentry->d_inode;
if (!S_ISREG(inode->i_mode))
return -EOPNOTSUPP;
if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
/* TODO: Add support for non extent hole punching */
return -EOPNOTSUPP;
}
if (EXT4_SB(inode->i_sb)->s_cluster_ratio > 1) {
/* TODO: Add support for bigalloc file systems */
return -EOPNOTSUPP;
}
return ext4_ext_punch_hole(file, offset, length);
}
/*
* ext4_truncate()
*
* We block out ext4_get_block() block instantiations across the entire
* transaction, and VFS/VM ensures that ext4_truncate() cannot run
* simultaneously on behalf of the same inode.
*
* As we work through the truncate and commit bits of it to the journal there
* is one core, guiding principle: the file's tree must always be consistent on
* disk. We must be able to restart the truncate after a crash.
*
* The file's tree may be transiently inconsistent in memory (although it
* probably isn't), but whenever we close off and commit a journal transaction,
* the contents of (the filesystem + the journal) must be consistent and
* restartable. It's pretty simple, really: bottom up, right to left (although
* left-to-right works OK too).
*
* Note that at recovery time, journal replay occurs *before* the restart of
* truncate against the orphan inode list.
*
* The committed inode has the new, desired i_size (which is the same as
* i_disksize in this case). After a crash, ext4_orphan_cleanup() will see
* that this inode's truncate did not complete and it will again call
* ext4_truncate() to have another go. So there will be instantiated blocks
* to the right of the truncation point in a crashed ext4 filesystem. But
* that's fine - as long as they are linked from the inode, the post-crash
* ext4_truncate() run will find them and release them.
*/
void ext4_truncate(struct inode *inode)
{
trace_ext4_truncate_enter(inode);
if (!ext4_can_truncate(inode))
return;
ext4_clear_inode_flag(inode, EXT4_INODE_EOFBLOCKS);
if (inode->i_size == 0 && !test_opt(inode->i_sb, NO_AUTO_DA_ALLOC))
ext4_set_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE);
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
ext4_ext_truncate(inode);
else
ext4_ind_truncate(inode);
trace_ext4_truncate_exit(inode);
}
/*
* ext4_get_inode_loc returns with an extra refcount against the inode's
* underlying buffer_head on success. If 'in_mem' is true, we have all
* data in memory that is needed to recreate the on-disk version of this
* inode.
*/
static int __ext4_get_inode_loc(struct inode *inode,
struct ext4_iloc *iloc, int in_mem)
{
struct ext4_group_desc *gdp;
struct buffer_head *bh;
struct super_block *sb = inode->i_sb;
ext4_fsblk_t block;
int inodes_per_block, inode_offset;
iloc->bh = NULL;
if (!ext4_valid_inum(sb, inode->i_ino))
return -EIO;
iloc->block_group = (inode->i_ino - 1) / EXT4_INODES_PER_GROUP(sb);
gdp = ext4_get_group_desc(sb, iloc->block_group, NULL);
if (!gdp)
return -EIO;
/*
* Figure out the offset within the block group inode table
*/
inodes_per_block = EXT4_SB(sb)->s_inodes_per_block;
inode_offset = ((inode->i_ino - 1) %
EXT4_INODES_PER_GROUP(sb));
block = ext4_inode_table(sb, gdp) + (inode_offset / inodes_per_block);
iloc->offset = (inode_offset % inodes_per_block) * EXT4_INODE_SIZE(sb);
bh = sb_getblk(sb, block);
if (!bh) {
EXT4_ERROR_INODE_BLOCK(inode, block,
"unable to read itable block");
return -EIO;
}
if (!buffer_uptodate(bh)) {
lock_buffer(bh);
/*
* If the buffer has the write error flag, we have failed
* to write out another inode in the same block. In this
* case, we don't have to read the block because we may
* read the old inode data successfully.
*/
if (buffer_write_io_error(bh) && !buffer_uptodate(bh))
set_buffer_uptodate(bh);
if (buffer_uptodate(bh)) {
/* someone brought it uptodate while we waited */
unlock_buffer(bh);
goto has_buffer;
}
/*
* If we have all information of the inode in memory and this
* is the only valid inode in the block, we need not read the
* block.
*/
if (in_mem) {
struct buffer_head *bitmap_bh;
int i, start;
start = inode_offset & ~(inodes_per_block - 1);
/* Is the inode bitmap in cache? */
bitmap_bh = sb_getblk(sb, ext4_inode_bitmap(sb, gdp));
if (!bitmap_bh)
goto make_io;
/*
* If the inode bitmap isn't in cache then the
* optimisation may end up performing two reads instead
* of one, so skip it.
*/
if (!buffer_uptodate(bitmap_bh)) {
brelse(bitmap_bh);
goto make_io;
}
for (i = start; i < start + inodes_per_block; i++) {
if (i == inode_offset)
continue;
if (ext4_test_bit(i, bitmap_bh->b_data))
break;
}
brelse(bitmap_bh);
if (i == start + inodes_per_block) {
/* all other inodes are free, so skip I/O */
memset(bh->b_data, 0, bh->b_size);
set_buffer_uptodate(bh);
unlock_buffer(bh);
goto has_buffer;
}
}
make_io:
/*
* If we need to do any I/O, try to pre-readahead extra
* blocks from the inode table.
*/
if (EXT4_SB(sb)->s_inode_readahead_blks) {
ext4_fsblk_t b, end, table;
unsigned num;
table = ext4_inode_table(sb, gdp);
/* s_inode_readahead_blks is always a power of 2 */
b = block & ~(EXT4_SB(sb)->s_inode_readahead_blks-1);
if (table > b)
b = table;
end = b + EXT4_SB(sb)->s_inode_readahead_blks;
num = EXT4_INODES_PER_GROUP(sb);
if (ext4_has_group_desc_csum(sb))
num -= ext4_itable_unused_count(sb, gdp);
table += num / inodes_per_block;
if (end > table)
end = table;
while (b <= end)
sb_breadahead(sb, b++);
}
/*
* There are other valid inodes in the buffer, this inode
* has in-inode xattrs, or we don't have this inode in memory.
* Read the block from disk.
*/
trace_ext4_load_inode(inode);
get_bh(bh);
bh->b_end_io = end_buffer_read_sync;
submit_bh(READ | REQ_META | REQ_PRIO, bh);
wait_on_buffer(bh);
if (!buffer_uptodate(bh)) {
EXT4_ERROR_INODE_BLOCK(inode, block,
"unable to read itable block");
brelse(bh);
return -EIO;
}
}
has_buffer:
iloc->bh = bh;
return 0;
}
int ext4_get_inode_loc(struct inode *inode, struct ext4_iloc *iloc)
{
/* We have all inode data except xattrs in memory here. */
return __ext4_get_inode_loc(inode, iloc,
!ext4_test_inode_state(inode, EXT4_STATE_XATTR));
}
void ext4_set_inode_flags(struct inode *inode)
{
unsigned int flags = EXT4_I(inode)->i_flags;
inode->i_flags &= ~(S_SYNC|S_APPEND|S_IMMUTABLE|S_NOATIME|S_DIRSYNC);
if (flags & EXT4_SYNC_FL)
inode->i_flags |= S_SYNC;
if (flags & EXT4_APPEND_FL)
inode->i_flags |= S_APPEND;
if (flags & EXT4_IMMUTABLE_FL)
inode->i_flags |= S_IMMUTABLE;
if (flags & EXT4_NOATIME_FL)
inode->i_flags |= S_NOATIME;
if (flags & EXT4_DIRSYNC_FL)
inode->i_flags |= S_DIRSYNC;
}
/* Propagate flags from i_flags to EXT4_I(inode)->i_flags */
void ext4_get_inode_flags(struct ext4_inode_info *ei)
{
unsigned int vfs_fl;
unsigned long old_fl, new_fl;
do {
vfs_fl = ei->vfs_inode.i_flags;
old_fl = ei->i_flags;
new_fl = old_fl & ~(EXT4_SYNC_FL|EXT4_APPEND_FL|
EXT4_IMMUTABLE_FL|EXT4_NOATIME_FL|
EXT4_DIRSYNC_FL);
if (vfs_fl & S_SYNC)
new_fl |= EXT4_SYNC_FL;
if (vfs_fl & S_APPEND)
new_fl |= EXT4_APPEND_FL;
if (vfs_fl & S_IMMUTABLE)
new_fl |= EXT4_IMMUTABLE_FL;
if (vfs_fl & S_NOATIME)
new_fl |= EXT4_NOATIME_FL;
if (vfs_fl & S_DIRSYNC)
new_fl |= EXT4_DIRSYNC_FL;
} while (cmpxchg(&ei->i_flags, old_fl, new_fl) != old_fl);
}
static blkcnt_t ext4_inode_blocks(struct ext4_inode *raw_inode,
struct ext4_inode_info *ei)
{
blkcnt_t i_blocks ;
struct inode *inode = &(ei->vfs_inode);
struct super_block *sb = inode->i_sb;
if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
EXT4_FEATURE_RO_COMPAT_HUGE_FILE)) {
/* we are using combined 48 bit field */
i_blocks = ((u64)le16_to_cpu(raw_inode->i_blocks_high)) << 32 |
le32_to_cpu(raw_inode->i_blocks_lo);
if (ext4_test_inode_flag(inode, EXT4_INODE_HUGE_FILE)) {
/* i_blocks represent file system block size */
return i_blocks << (inode->i_blkbits - 9);
} else {
return i_blocks;
}
} else {
return le32_to_cpu(raw_inode->i_blocks_lo);
}
}
static inline void ext4_iget_extra_inode(struct inode *inode,
struct ext4_inode *raw_inode,
struct ext4_inode_info *ei)
{
__le32 *magic = (void *)raw_inode +
EXT4_GOOD_OLD_INODE_SIZE + ei->i_extra_isize;
if (*magic == cpu_to_le32(EXT4_XATTR_MAGIC)) {
ext4_set_inode_state(inode, EXT4_STATE_XATTR);
ext4_find_inline_data_nolock(inode);
} else
EXT4_I(inode)->i_inline_off = 0;
}
struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
{
struct ext4_iloc iloc;
struct ext4_inode *raw_inode;
struct ext4_inode_info *ei;
struct inode *inode;
journal_t *journal = EXT4_SB(sb)->s_journal;
long ret;
int block;
uid_t i_uid;
gid_t i_gid;
inode = iget_locked(sb, ino);
if (!inode)
return ERR_PTR(-ENOMEM);
if (!(inode->i_state & I_NEW))
return inode;
ei = EXT4_I(inode);
iloc.bh = NULL;
ret = __ext4_get_inode_loc(inode, &iloc, 0);
if (ret < 0)
goto bad_inode;
raw_inode = ext4_raw_inode(&iloc);
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE) {
ei->i_extra_isize = le16_to_cpu(raw_inode->i_extra_isize);
if (EXT4_GOOD_OLD_INODE_SIZE + ei->i_extra_isize >
EXT4_INODE_SIZE(inode->i_sb)) {
EXT4_ERROR_INODE(inode, "bad extra_isize (%u != %u)",
EXT4_GOOD_OLD_INODE_SIZE + ei->i_extra_isize,
EXT4_INODE_SIZE(inode->i_sb));
ret = -EIO;
goto bad_inode;
}
} else
ei->i_extra_isize = 0;
/* Precompute checksum seed for inode metadata */
if (EXT4_HAS_RO_COMPAT_FEATURE(sb,
EXT4_FEATURE_RO_COMPAT_METADATA_CSUM)) {
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
__u32 csum;
__le32 inum = cpu_to_le32(inode->i_ino);
__le32 gen = raw_inode->i_generation;
csum = ext4_chksum(sbi, sbi->s_csum_seed, (__u8 *)&inum,
sizeof(inum));
ei->i_csum_seed = ext4_chksum(sbi, csum, (__u8 *)&gen,
sizeof(gen));
}
if (!ext4_inode_csum_verify(inode, raw_inode, ei)) {
EXT4_ERROR_INODE(inode, "checksum invalid");
ret = -EIO;
goto bad_inode;
}
inode->i_mode = le16_to_cpu(raw_inode->i_mode);
i_uid = (uid_t)le16_to_cpu(raw_inode->i_uid_low);
i_gid = (gid_t)le16_to_cpu(raw_inode->i_gid_low);
if (!(test_opt(inode->i_sb, NO_UID32))) {
i_uid |= le16_to_cpu(raw_inode->i_uid_high) << 16;
i_gid |= le16_to_cpu(raw_inode->i_gid_high) << 16;
}
i_uid_write(inode, i_uid);
i_gid_write(inode, i_gid);
set_nlink(inode, le16_to_cpu(raw_inode->i_links_count));
ext4_clear_state_flags(ei); /* Only relevant on 32-bit archs */
ei->i_inline_off = 0;
ei->i_dir_start_lookup = 0;
ei->i_dtime = le32_to_cpu(raw_inode->i_dtime);
/* We now have enough fields to check if the inode was active or not.
* This is needed because nfsd might try to access dead inodes
* the test is that same one that e2fsck uses
* NeilBrown 1999oct15
*/
if (inode->i_nlink == 0) {
if (inode->i_mode == 0 ||
!(EXT4_SB(inode->i_sb)->s_mount_state & EXT4_ORPHAN_FS)) {
/* this inode is deleted */
ret = -ESTALE;
goto bad_inode;
}
/* The only unlinked inodes we let through here have
* valid i_mode and are being read by the orphan
* recovery code: that's fine, we're about to complete
* the process of deleting those. */
}
ei->i_flags = le32_to_cpu(raw_inode->i_flags);
inode->i_blocks = ext4_inode_blocks(raw_inode, ei);
ei->i_file_acl = le32_to_cpu(raw_inode->i_file_acl_lo);
if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_64BIT))
ei->i_file_acl |=
((__u64)le16_to_cpu(raw_inode->i_file_acl_high)) << 32;
inode->i_size = ext4_isize(raw_inode);
ei->i_disksize = inode->i_size;
#ifdef CONFIG_QUOTA
ei->i_reserved_quota = 0;
#endif
inode->i_generation = le32_to_cpu(raw_inode->i_generation);
ei->i_block_group = iloc.block_group;
ei->i_last_alloc_group = ~0;
/*
* NOTE! The in-memory inode i_data array is in little-endian order
* even on big-endian machines: we do NOT byteswap the block numbers!
*/
for (block = 0; block < EXT4_N_BLOCKS; block++)
ei->i_data[block] = raw_inode->i_block[block];
INIT_LIST_HEAD(&ei->i_orphan);
/*
* Set transaction id's of transactions that have to be committed
* to finish f[data]sync. We set them to currently running transaction
* as we cannot be sure that the inode or some of its metadata isn't
* part of the transaction - the inode could have been reclaimed and
* now it is reread from disk.
*/
if (journal) {
transaction_t *transaction;
tid_t tid;
read_lock(&journal->j_state_lock);
if (journal->j_running_transaction)
transaction = journal->j_running_transaction;
else
transaction = journal->j_committing_transaction;
if (transaction)
tid = transaction->t_tid;
else
tid = journal->j_commit_sequence;
read_unlock(&journal->j_state_lock);
ei->i_sync_tid = tid;
ei->i_datasync_tid = tid;
}
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE) {
if (ei->i_extra_isize == 0) {
/* The extra space is currently unused. Use it. */
ei->i_extra_isize = sizeof(struct ext4_inode) -
EXT4_GOOD_OLD_INODE_SIZE;
} else {
ext4_iget_extra_inode(inode, raw_inode, ei);
}
}
EXT4_INODE_GET_XTIME(i_ctime, inode, raw_inode);
EXT4_INODE_GET_XTIME(i_mtime, inode, raw_inode);
EXT4_INODE_GET_XTIME(i_atime, inode, raw_inode);
EXT4_EINODE_GET_XTIME(i_crtime, ei, raw_inode);
inode->i_version = le32_to_cpu(raw_inode->i_disk_version);
if (EXT4_INODE_SIZE(inode->i_sb) > EXT4_GOOD_OLD_INODE_SIZE) {
if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi))
inode->i_version |=
(__u64)(le32_to_cpu(raw_inode->i_version_hi)) << 32;
}
ret = 0;
if (ei->i_file_acl &&
!ext4_data_block_valid(EXT4_SB(sb), ei->i_file_acl, 1)) {
EXT4_ERROR_INODE(inode, "bad extended attribute block %llu",
ei->i_file_acl);
ret = -EIO;
goto bad_inode;
} else if (!ext4_has_inline_data(inode)) {
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
if ((S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
(S_ISLNK(inode->i_mode) &&
!ext4_inode_is_fast_symlink(inode))))
/* Validate extent which is part of inode */
ret = ext4_ext_check_inode(inode);
} else if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
(S_ISLNK(inode->i_mode) &&
!ext4_inode_is_fast_symlink(inode))) {
/* Validate block references which are part of inode */
ret = ext4_ind_check_inode(inode);
}
}
if (ret)
goto bad_inode;
if (S_ISREG(inode->i_mode)) {
inode->i_op = &ext4_file_inode_operations;
inode->i_fop = &ext4_file_operations;
ext4_set_aops(inode);
} else if (S_ISDIR(inode->i_mode)) {
inode->i_op = &ext4_dir_inode_operations;
inode->i_fop = &ext4_dir_operations;
} else if (S_ISLNK(inode->i_mode)) {
if (ext4_inode_is_fast_symlink(inode)) {
inode->i_op = &ext4_fast_symlink_inode_operations;
nd_terminate_link(ei->i_data, inode->i_size,
sizeof(ei->i_data) - 1);
} else {
inode->i_op = &ext4_symlink_inode_operations;
ext4_set_aops(inode);
}
} else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode) ||
S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
inode->i_op = &ext4_special_inode_operations;
if (raw_inode->i_block[0])
init_special_inode(inode, inode->i_mode,
old_decode_dev(le32_to_cpu(raw_inode->i_block[0])));
else
init_special_inode(inode, inode->i_mode,
new_decode_dev(le32_to_cpu(raw_inode->i_block[1])));
} else {
ret = -EIO;
EXT4_ERROR_INODE(inode, "bogus i_mode (%o)", inode->i_mode);
goto bad_inode;
}
brelse(iloc.bh);
ext4_set_inode_flags(inode);
unlock_new_inode(inode);
return inode;
bad_inode:
brelse(iloc.bh);
iget_failed(inode);
return ERR_PTR(ret);
}
static int ext4_inode_blocks_set(handle_t *handle,
struct ext4_inode *raw_inode,
struct ext4_inode_info *ei)
{
struct inode *inode = &(ei->vfs_inode);
u64 i_blocks = inode->i_blocks;
struct super_block *sb = inode->i_sb;
if (i_blocks <= ~0U) {
/*
* i_blocks can be represented in a 32 bit variable
* as multiple of 512 bytes
*/
raw_inode->i_blocks_lo = cpu_to_le32(i_blocks);
raw_inode->i_blocks_high = 0;
ext4_clear_inode_flag(inode, EXT4_INODE_HUGE_FILE);
return 0;
}
if (!EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_HUGE_FILE))
return -EFBIG;
if (i_blocks <= 0xffffffffffffULL) {
/*
* i_blocks can be represented in a 48 bit variable
* as multiple of 512 bytes
*/
raw_inode->i_blocks_lo = cpu_to_le32(i_blocks);
raw_inode->i_blocks_high = cpu_to_le16(i_blocks >> 32);
ext4_clear_inode_flag(inode, EXT4_INODE_HUGE_FILE);
} else {
ext4_set_inode_flag(inode, EXT4_INODE_HUGE_FILE);
/* i_block is stored in file system block size */
i_blocks = i_blocks >> (inode->i_blkbits - 9);
raw_inode->i_blocks_lo = cpu_to_le32(i_blocks);
raw_inode->i_blocks_high = cpu_to_le16(i_blocks >> 32);
}
return 0;
}
/*
* Post the struct inode info into an on-disk inode location in the
* buffer-cache. This gobbles the caller's reference to the
* buffer_head in the inode location struct.
*
* The caller must have write access to iloc->bh.
*/
static int ext4_do_update_inode(handle_t *handle,
struct inode *inode,
struct ext4_iloc *iloc)
{
struct ext4_inode *raw_inode = ext4_raw_inode(iloc);
struct ext4_inode_info *ei = EXT4_I(inode);
struct buffer_head *bh = iloc->bh;
int err = 0, rc, block;
int need_datasync = 0;
uid_t i_uid;
gid_t i_gid;
/* For fields not not tracking in the in-memory inode,
* initialise them to zero for new inodes. */
if (ext4_test_inode_state(inode, EXT4_STATE_NEW))
memset(raw_inode, 0, EXT4_SB(inode->i_sb)->s_inode_size);
ext4_get_inode_flags(ei);
raw_inode->i_mode = cpu_to_le16(inode->i_mode);
i_uid = i_uid_read(inode);
i_gid = i_gid_read(inode);
if (!(test_opt(inode->i_sb, NO_UID32))) {
raw_inode->i_uid_low = cpu_to_le16(low_16_bits(i_uid));
raw_inode->i_gid_low = cpu_to_le16(low_16_bits(i_gid));
/*
* Fix up interoperability with old kernels. Otherwise, old inodes get
* re-used with the upper 16 bits of the uid/gid intact
*/
if (!ei->i_dtime) {
raw_inode->i_uid_high =
cpu_to_le16(high_16_bits(i_uid));
raw_inode->i_gid_high =
cpu_to_le16(high_16_bits(i_gid));
} else {
raw_inode->i_uid_high = 0;
raw_inode->i_gid_high = 0;
}
} else {
raw_inode->i_uid_low = cpu_to_le16(fs_high2lowuid(i_uid));
raw_inode->i_gid_low = cpu_to_le16(fs_high2lowgid(i_gid));
raw_inode->i_uid_high = 0;
raw_inode->i_gid_high = 0;
}
raw_inode->i_links_count = cpu_to_le16(inode->i_nlink);
EXT4_INODE_SET_XTIME(i_ctime, inode, raw_inode);
EXT4_INODE_SET_XTIME(i_mtime, inode, raw_inode);
EXT4_INODE_SET_XTIME(i_atime, inode, raw_inode);
EXT4_EINODE_SET_XTIME(i_crtime, ei, raw_inode);
if (ext4_inode_blocks_set(handle, raw_inode, ei))
goto out_brelse;
raw_inode->i_dtime = cpu_to_le32(ei->i_dtime);
raw_inode->i_flags = cpu_to_le32(ei->i_flags & 0xFFFFFFFF);
if (EXT4_SB(inode->i_sb)->s_es->s_creator_os !=
cpu_to_le32(EXT4_OS_HURD))
raw_inode->i_file_acl_high =
cpu_to_le16(ei->i_file_acl >> 32);
raw_inode->i_file_acl_lo = cpu_to_le32(ei->i_file_acl);
if (ei->i_disksize != ext4_isize(raw_inode)) {
ext4_isize_set(raw_inode, ei->i_disksize);
need_datasync = 1;
}
if (ei->i_disksize > 0x7fffffffULL) {
struct super_block *sb = inode->i_sb;
if (!EXT4_HAS_RO_COMPAT_FEATURE(sb,
EXT4_FEATURE_RO_COMPAT_LARGE_FILE) ||
EXT4_SB(sb)->s_es->s_rev_level ==
cpu_to_le32(EXT4_GOOD_OLD_REV)) {
/* If this is the first large file
* created, add a flag to the superblock.
*/
err = ext4_journal_get_write_access(handle,
EXT4_SB(sb)->s_sbh);
if (err)
goto out_brelse;
ext4_update_dynamic_rev(sb);
EXT4_SET_RO_COMPAT_FEATURE(sb,
EXT4_FEATURE_RO_COMPAT_LARGE_FILE);
ext4_handle_sync(handle);
err = ext4_handle_dirty_super(handle, sb);
}
}
raw_inode->i_generation = cpu_to_le32(inode->i_generation);
if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) {
if (old_valid_dev(inode->i_rdev)) {
raw_inode->i_block[0] =
cpu_to_le32(old_encode_dev(inode->i_rdev));
raw_inode->i_block[1] = 0;
} else {
raw_inode->i_block[0] = 0;
raw_inode->i_block[1] =
cpu_to_le32(new_encode_dev(inode->i_rdev));
raw_inode->i_block[2] = 0;
}
} else if (!ext4_has_inline_data(inode)) {
for (block = 0; block < EXT4_N_BLOCKS; block++)
raw_inode->i_block[block] = ei->i_data[block];
}
raw_inode->i_disk_version = cpu_to_le32(inode->i_version);
if (ei->i_extra_isize) {
if (EXT4_FITS_IN_INODE(raw_inode, ei, i_version_hi))
raw_inode->i_version_hi =
cpu_to_le32(inode->i_version >> 32);
raw_inode->i_extra_isize = cpu_to_le16(ei->i_extra_isize);
}
ext4_inode_csum_set(inode, raw_inode, ei);
BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
rc = ext4_handle_dirty_metadata(handle, NULL, bh);
if (!err)
err = rc;
ext4_clear_inode_state(inode, EXT4_STATE_NEW);
ext4_update_inode_fsync_trans(handle, inode, need_datasync);
out_brelse:
brelse(bh);
ext4_std_error(inode->i_sb, err);
return err;
}
/*
* ext4_write_inode()
*
* We are called from a few places:
*
* - Within generic_file_write() for O_SYNC files.
* Here, there will be no transaction running. We wait for any running
* transaction to commit.
*
* - Within sys_sync(), kupdate and such.
* We wait on commit, if tol to.
*
* - Within prune_icache() (PF_MEMALLOC == true)
* Here we simply return. We can't afford to block kswapd on the
* journal commit.
*
* In all cases it is actually safe for us to return without doing anything,
* because the inode has been copied into a raw inode buffer in
* ext4_mark_inode_dirty(). This is a correctness thing for O_SYNC and for
* knfsd.
*
* Note that we are absolutely dependent upon all inode dirtiers doing the
* right thing: they *must* call mark_inode_dirty() after dirtying info in
* which we are interested.
*
* It would be a bug for them to not do this. The code:
*
* mark_inode_dirty(inode)
* stuff();
* inode->i_size = expr;
*
* is in error because a kswapd-driven write_inode() could occur while
* `stuff()' is running, and the new i_size will be lost. Plus the inode
* will no longer be on the superblock's dirty inode list.
*/
int ext4_write_inode(struct inode *inode, struct writeback_control *wbc)
{
int err;
if (current->flags & PF_MEMALLOC)
return 0;
if (EXT4_SB(inode->i_sb)->s_journal) {
if (ext4_journal_current_handle()) {
jbd_debug(1, "called recursively, non-PF_MEMALLOC!\n");
dump_stack();
return -EIO;
}
if (wbc->sync_mode != WB_SYNC_ALL)
return 0;
err = ext4_force_commit(inode->i_sb);
} else {
struct ext4_iloc iloc;
err = __ext4_get_inode_loc(inode, &iloc, 0);
if (err)
return err;
if (wbc->sync_mode == WB_SYNC_ALL)
sync_dirty_buffer(iloc.bh);
if (buffer_req(iloc.bh) && !buffer_uptodate(iloc.bh)) {
EXT4_ERROR_INODE_BLOCK(inode, iloc.bh->b_blocknr,
"IO error syncing inode");
err = -EIO;
}
brelse(iloc.bh);
}
return err;
}
/*
* ext4_setattr()
*
* Called from notify_change.
*
* We want to trap VFS attempts to truncate the file as soon as
* possible. In particular, we want to make sure that when the VFS
* shrinks i_size, we put the inode on the orphan list and modify
* i_disksize immediately, so that during the subsequent flushing of
* dirty pages and freeing of disk blocks, we can guarantee that any
* commit will leave the blocks being flushed in an unused state on
* disk. (On recovery, the inode will get truncated and the blocks will
* be freed, so we have a strong guarantee that no future commit will
* leave these blocks visible to the user.)
*
* Another thing we have to assure is that if we are in ordered mode
* and inode is still attached to the committing transaction, we must
* we start writeout of all the dirty pages which are being truncated.
* This way we are sure that all the data written in the previous
* transaction are already on disk (truncate waits for pages under
* writeback).
*
* Called with inode->i_mutex down.
*/
int ext4_setattr(struct dentry *dentry, struct iattr *attr)
{
struct inode *inode = dentry->d_inode;
int error, rc = 0;
int orphan = 0;
const unsigned int ia_valid = attr->ia_valid;
error = inode_change_ok(inode, attr);
if (error)
return error;
if (is_quota_modification(inode, attr))
dquot_initialize(inode);
if ((ia_valid & ATTR_UID && !uid_eq(attr->ia_uid, inode->i_uid)) ||
(ia_valid & ATTR_GID && !gid_eq(attr->ia_gid, inode->i_gid))) {
handle_t *handle;
/* (user+group)*(old+new) structure, inode write (sb,
* inode block, ? - but truncate inode update has it) */
handle = ext4_journal_start(inode, (EXT4_MAXQUOTAS_INIT_BLOCKS(inode->i_sb)+
EXT4_MAXQUOTAS_DEL_BLOCKS(inode->i_sb))+3);
if (IS_ERR(handle)) {
error = PTR_ERR(handle);
goto err_out;
}
error = dquot_transfer(inode, attr);
if (error) {
ext4_journal_stop(handle);
return error;
}
/* Update corresponding info in inode so that everything is in
* one transaction */
if (attr->ia_valid & ATTR_UID)
inode->i_uid = attr->ia_uid;
if (attr->ia_valid & ATTR_GID)
inode->i_gid = attr->ia_gid;
error = ext4_mark_inode_dirty(handle, inode);
ext4_journal_stop(handle);
}
if (attr->ia_valid & ATTR_SIZE) {
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))) {
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
if (attr->ia_size > sbi->s_bitmap_maxbytes)
return -EFBIG;
}
}
if (S_ISREG(inode->i_mode) &&
attr->ia_valid & ATTR_SIZE &&
(attr->ia_size < inode->i_size)) {
handle_t *handle;
handle = ext4_journal_start(inode, 3);
if (IS_ERR(handle)) {
error = PTR_ERR(handle);
goto err_out;
}
if (ext4_handle_valid(handle)) {
error = ext4_orphan_add(handle, inode);
orphan = 1;
}
EXT4_I(inode)->i_disksize = attr->ia_size;
rc = ext4_mark_inode_dirty(handle, inode);
if (!error)
error = rc;
ext4_journal_stop(handle);
if (ext4_should_order_data(inode)) {
error = ext4_begin_ordered_truncate(inode,
attr->ia_size);
if (error) {
/* Do as much error cleanup as possible */
handle = ext4_journal_start(inode, 3);
if (IS_ERR(handle)) {
ext4_orphan_del(NULL, inode);
goto err_out;
}
ext4_orphan_del(handle, inode);
orphan = 0;
ext4_journal_stop(handle);
goto err_out;
}
}
}
if (attr->ia_valid & ATTR_SIZE) {
if (attr->ia_size != i_size_read(inode)) {
truncate_setsize(inode, attr->ia_size);
/* Inode size will be reduced, wait for dio in flight.
* Temporarily disable dioread_nolock to prevent
* livelock. */
if (orphan) {
ext4_inode_block_unlocked_dio(inode);
inode_dio_wait(inode);
ext4_inode_resume_unlocked_dio(inode);
}
}
ext4_truncate(inode);
}
if (!rc) {
setattr_copy(inode, attr);
mark_inode_dirty(inode);
}
/*
* If the call to ext4_truncate failed to get a transaction handle at
* all, we need to clean up the in-core orphan list manually.
*/
if (orphan && inode->i_nlink)
ext4_orphan_del(NULL, inode);
if (!rc && (ia_valid & ATTR_MODE))
rc = ext4_acl_chmod(inode);
err_out:
ext4_std_error(inode->i_sb, error);
if (!error)
error = rc;
return error;
}
int ext4_getattr(struct vfsmount *mnt, struct dentry *dentry,
struct kstat *stat)
{
struct inode *inode;
unsigned long delalloc_blocks;
inode = dentry->d_inode;
generic_fillattr(inode, stat);
/*
* We can't update i_blocks if the block allocation is delayed
* otherwise in the case of system crash before the real block
* allocation is done, we will have i_blocks inconsistent with
* on-disk file blocks.
* We always keep i_blocks updated together with real
* allocation. But to not confuse with user, stat
* will return the blocks that include the delayed allocation
* blocks for this file.
*/
delalloc_blocks = EXT4_C2B(EXT4_SB(inode->i_sb),
EXT4_I(inode)->i_reserved_data_blocks);
stat->blocks += (delalloc_blocks << inode->i_sb->s_blocksize_bits)>>9;
return 0;
}
static int ext4_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
{
if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
return ext4_ind_trans_blocks(inode, nrblocks, chunk);
return ext4_ext_index_trans_blocks(inode, nrblocks, chunk);
}
/*
* Account for index blocks, block groups bitmaps and block group
* descriptor blocks if modify datablocks and index blocks
* worse case, the indexs blocks spread over different block groups
*
* If datablocks are discontiguous, they are possible to spread over
* different block groups too. If they are contiguous, with flexbg,
* they could still across block group boundary.
*
* Also account for superblock, inode, quota and xattr blocks
*/
static int ext4_meta_trans_blocks(struct inode *inode, int nrblocks, int chunk)
{
ext4_group_t groups, ngroups = ext4_get_groups_count(inode->i_sb);
int gdpblocks;
int idxblocks;
int ret = 0;
/*
* How many index blocks need to touch to modify nrblocks?
* The "Chunk" flag indicating whether the nrblocks is
* physically contiguous on disk
*
* For Direct IO and fallocate, they calls get_block to allocate
* one single extent at a time, so they could set the "Chunk" flag
*/
idxblocks = ext4_index_trans_blocks(inode, nrblocks, chunk);
ret = idxblocks;
/*
* Now let's see how many group bitmaps and group descriptors need
* to account
*/
groups = idxblocks;
if (chunk)
groups += 1;
else
groups += nrblocks;
gdpblocks = groups;
if (groups > ngroups)
groups = ngroups;
if (groups > EXT4_SB(inode->i_sb)->s_gdb_count)
gdpblocks = EXT4_SB(inode->i_sb)->s_gdb_count;
/* bitmaps and block group descriptor blocks */
ret += groups + gdpblocks;
/* Blocks for super block, inode, quota and xattr blocks */
ret += EXT4_META_TRANS_BLOCKS(inode->i_sb);
return ret;
}
/*
* Calculate the total number of credits to reserve to fit
* the modification of a single pages into a single transaction,
* which may include multiple chunks of block allocations.
*
* This could be called via ext4_write_begin()
*
* We need to consider the worse case, when
* one new block per extent.
*/
int ext4_writepage_trans_blocks(struct inode *inode)
{
int bpp = ext4_journal_blocks_per_page(inode);
int ret;
ret = ext4_meta_trans_blocks(inode, bpp, 0);
/* Account for data blocks for journalled mode */
if (ext4_should_journal_data(inode))
ret += bpp;
return ret;
}
/*
* Calculate the journal credits for a chunk of data modification.
*
* This is called from DIO, fallocate or whoever calling
* ext4_map_blocks() to map/allocate a chunk of contiguous disk blocks.
*
* journal buffers for data blocks are not included here, as DIO
* and fallocate do no need to journal data buffers.
*/
int ext4_chunk_trans_blocks(struct inode *inode, int nrblocks)
{
return ext4_meta_trans_blocks(inode, nrblocks, 1);
}
/*
* The caller must have previously called ext4_reserve_inode_write().
* Give this, we know that the caller already has write access to iloc->bh.
*/
int ext4_mark_iloc_dirty(handle_t *handle,
struct inode *inode, struct ext4_iloc *iloc)
{
int err = 0;
if (IS_I_VERSION(inode))
inode_inc_iversion(inode);
/* the do_update_inode consumes one bh->b_count */
get_bh(iloc->bh);
/* ext4_do_update_inode() does jbd2_journal_dirty_metadata */
err = ext4_do_update_inode(handle, inode, iloc);
put_bh(iloc->bh);
return err;
}
/*
* On success, We end up with an outstanding reference count against
* iloc->bh. This _must_ be cleaned up later.
*/
int
ext4_reserve_inode_write(handle_t *handle, struct inode *inode,
struct ext4_iloc *iloc)
{
int err;
err = ext4_get_inode_loc(inode, iloc);
if (!err) {
BUFFER_TRACE(iloc->bh, "get_write_access");
err = ext4_journal_get_write_access(handle, iloc->bh);
if (err) {
brelse(iloc->bh);
iloc->bh = NULL;
}
}
ext4_std_error(inode->i_sb, err);
return err;
}
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
/*
* Expand an inode by new_extra_isize bytes.
* Returns 0 on success or negative error number on failure.
*/
static int ext4_expand_extra_isize(struct inode *inode,
unsigned int new_extra_isize,
struct ext4_iloc iloc,
handle_t *handle)
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
{
struct ext4_inode *raw_inode;
struct ext4_xattr_ibody_header *header;
if (EXT4_I(inode)->i_extra_isize >= new_extra_isize)
return 0;
raw_inode = ext4_raw_inode(&iloc);
header = IHDR(inode, raw_inode);
/* No extended attributes present */
if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR) ||
header->h_magic != cpu_to_le32(EXT4_XATTR_MAGIC)) {
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
memset((void *)raw_inode + EXT4_GOOD_OLD_INODE_SIZE, 0,
new_extra_isize);
EXT4_I(inode)->i_extra_isize = new_extra_isize;
return 0;
}
/* try to expand with EAs present */
return ext4_expand_extra_isize_ea(inode, new_extra_isize,
raw_inode, handle);
}
/*
* What we do here is to mark the in-core inode as clean with respect to inode
* dirtiness (it may still be data-dirty).
* This means that the in-core inode may be reaped by prune_icache
* without having to perform any I/O. This is a very good thing,
* because *any* task may call prune_icache - even ones which
* have a transaction open against a different journal.
*
* Is this cheating? Not really. Sure, we haven't written the
* inode out, but prune_icache isn't a user-visible syncing function.
* Whenever the user wants stuff synced (sys_sync, sys_msync, sys_fsync)
* we start and wait on commits.
*/
int ext4_mark_inode_dirty(handle_t *handle, struct inode *inode)
{
struct ext4_iloc iloc;
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
static unsigned int mnt_count;
int err, ret;
might_sleep();
trace_ext4_mark_inode_dirty(inode, _RET_IP_);
err = ext4_reserve_inode_write(handle, inode, &iloc);
if (ext4_handle_valid(handle) &&
EXT4_I(inode)->i_extra_isize < sbi->s_want_extra_isize &&
!ext4_test_inode_state(inode, EXT4_STATE_NO_EXPAND)) {
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
/*
* We need extra buffer credits since we may write into EA block
* with this same handle. If journal_extend fails, then it will
* only result in a minor loss of functionality for that inode.
* If this is felt to be critical, then e2fsck should be run to
* force a large enough s_min_extra_isize.
*/
if ((jbd2_journal_extend(handle,
EXT4_DATA_TRANS_BLOCKS(inode->i_sb))) == 0) {
ret = ext4_expand_extra_isize(inode,
sbi->s_want_extra_isize,
iloc, handle);
if (ret) {
ext4_set_inode_state(inode,
EXT4_STATE_NO_EXPAND);
if (mnt_count !=
le16_to_cpu(sbi->s_es->s_mnt_count)) {
ext4_warning(inode->i_sb,
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
"Unable to expand inode %lu. Delete"
" some EAs or run e2fsck.",
inode->i_ino);
mnt_count =
le16_to_cpu(sbi->s_es->s_mnt_count);
ext4: Expand extra_inodes space per the s_{want,min}_extra_isize fields We need to make sure that existing ext3 filesystems can also avail the new fields that have been added to the ext4 inode. We use s_want_extra_isize and s_min_extra_isize to decide by how much we should expand the inode. If EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE feature is set then we expand the inode by max(s_want_extra_isize, s_min_extra_isize , sizeof(ext4_inode) - EXT4_GOOD_OLD_INODE_SIZE) bytes. Actually it is still an open question about whether users should be able to set s_*_extra_isize smaller than the known fields or not. This patch also adds the functionality to expand inodes to include the newly added fields. We start by trying to expand by s_want_extra_isize bytes and if its fails we try to expand by s_min_extra_isize bytes. This is done by changing the i_extra_isize if enough space is available in the inode and no EAs are present. If EAs are present and there is enough space in the inode then the EAs in the inode are shifted to make space. If enough space is not available in the inode due to the EAs then 1 or more EAs are shifted to the external EA block. In the worst case when even the external EA block does not have enough space we inform the user that some EA would need to be deleted or s_min_extra_isize would have to be reduced. Signed-off-by: Andreas Dilger <adilger@clusterfs.com> Signed-off-by: Kalpak Shah <kalpak@clusterfs.com> Signed-off-by: Mingming Cao <cmm@us.ibm.com> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2007-07-18 13:19:57 +00:00
}
}
}
}
if (!err)
err = ext4_mark_iloc_dirty(handle, inode, &iloc);
return err;
}
/*
* ext4_dirty_inode() is called from __mark_inode_dirty()
*
* We're really interested in the case where a file is being extended.
* i_size has been changed by generic_commit_write() and we thus need
* to include the updated inode in the current transaction.
*
* Also, dquot_alloc_block() will always dirty the inode when blocks
* are allocated to the file.
*
* If the inode is marked synchronous, we don't honour that here - doing
* so would cause a commit on atime updates, which we don't bother doing.
* We handle synchronous inodes at the highest possible level.
*/
void ext4_dirty_inode(struct inode *inode, int flags)
{
handle_t *handle;
handle = ext4_journal_start(inode, 2);
if (IS_ERR(handle))
goto out;
ext4_mark_inode_dirty(handle, inode);
ext4_journal_stop(handle);
out:
return;
}
#if 0
/*
* Bind an inode's backing buffer_head into this transaction, to prevent
* it from being flushed to disk early. Unlike
* ext4_reserve_inode_write, this leaves behind no bh reference and
* returns no iloc structure, so the caller needs to repeat the iloc
* lookup to mark the inode dirty later.
*/
static int ext4_pin_inode(handle_t *handle, struct inode *inode)
{
struct ext4_iloc iloc;
int err = 0;
if (handle) {
err = ext4_get_inode_loc(inode, &iloc);
if (!err) {
BUFFER_TRACE(iloc.bh, "get_write_access");
err = jbd2_journal_get_write_access(handle, iloc.bh);
if (!err)
err = ext4_handle_dirty_metadata(handle,
NULL,
iloc.bh);
brelse(iloc.bh);
}
}
ext4_std_error(inode->i_sb, err);
return err;
}
#endif
int ext4_change_inode_journal_flag(struct inode *inode, int val)
{
journal_t *journal;
handle_t *handle;
int err;
/*
* We have to be very careful here: changing a data block's
* journaling status dynamically is dangerous. If we write a
* data block to the journal, change the status and then delete
* that block, we risk forgetting to revoke the old log record
* from the journal and so a subsequent replay can corrupt data.
* So, first we make sure that the journal is empty and that
* nobody is changing anything.
*/
journal = EXT4_JOURNAL(inode);
if (!journal)
return 0;
if (is_journal_aborted(journal))
return -EROFS;
/* We have to allocate physical blocks for delalloc blocks
* before flushing journal. otherwise delalloc blocks can not
* be allocated any more. even more truncate on delalloc blocks
* could trigger BUG by flushing delalloc blocks in journal.
* There is no delalloc block in non-journal data mode.
*/
if (val && test_opt(inode->i_sb, DELALLOC)) {
err = ext4_alloc_da_blocks(inode);
if (err < 0)
return err;
}
/* Wait for all existing dio workers */
ext4_inode_block_unlocked_dio(inode);
inode_dio_wait(inode);
jbd2_journal_lock_updates(journal);
/*
* OK, there are no updates running now, and all cached data is
* synced to disk. We are now in a completely consistent state
* which doesn't have anything in the journal, and we know that
* no filesystem updates are running, so it is safe to modify
* the inode's in-core data-journaling state flag now.
*/
if (val)
ext4_set_inode_flag(inode, EXT4_INODE_JOURNAL_DATA);
else {
jbd2_journal_flush(journal);
ext4_clear_inode_flag(inode, EXT4_INODE_JOURNAL_DATA);
}
ext4_set_aops(inode);
jbd2_journal_unlock_updates(journal);
ext4_inode_resume_unlocked_dio(inode);
/* Finally we can mark the inode as dirty. */
handle = ext4_journal_start(inode, 1);
if (IS_ERR(handle))
return PTR_ERR(handle);
err = ext4_mark_inode_dirty(handle, inode);
ext4_handle_sync(handle);
ext4_journal_stop(handle);
ext4_std_error(inode->i_sb, err);
return err;
}
static int ext4_bh_unmapped(handle_t *handle, struct buffer_head *bh)
{
return !buffer_mapped(bh);
}
int ext4_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
{
struct page *page = vmf->page;
loff_t size;
unsigned long len;
int ret;
struct file *file = vma->vm_file;
struct inode *inode = file->f_path.dentry->d_inode;
struct address_space *mapping = inode->i_mapping;
handle_t *handle;
get_block_t *get_block;
int retries = 0;
sb_start_pagefault(inode->i_sb);
file_update_time(vma->vm_file);
/* Delalloc case is easy... */
if (test_opt(inode->i_sb, DELALLOC) &&
!ext4_should_journal_data(inode) &&
!ext4_nonda_switch(inode->i_sb)) {
do {
ret = __block_page_mkwrite(vma, vmf,
ext4_da_get_block_prep);
} while (ret == -ENOSPC &&
ext4_should_retry_alloc(inode->i_sb, &retries));
goto out_ret;
}
lock_page(page);
size = i_size_read(inode);
/* Page got truncated from under us? */
if (page->mapping != mapping || page_offset(page) > size) {
unlock_page(page);
ret = VM_FAULT_NOPAGE;
goto out;
}
if (page->index == size >> PAGE_CACHE_SHIFT)
len = size & ~PAGE_CACHE_MASK;
else
len = PAGE_CACHE_SIZE;
/*
* Return if we have all the buffers mapped. This avoids the need to do
* journal_start/journal_stop which can block and take a long time
*/
if (page_has_buffers(page)) {
if (!ext4_walk_page_buffers(NULL, page_buffers(page),
0, len, NULL,
ext4_bh_unmapped)) {
/* Wait so that we don't change page under IO */
wait_on_page_writeback(page);
ret = VM_FAULT_LOCKED;
goto out;
}
}
unlock_page(page);
/* OK, we need to fill the hole... */
if (ext4_should_dioread_nolock(inode))
get_block = ext4_get_block_write;
else
get_block = ext4_get_block;
retry_alloc:
handle = ext4_journal_start(inode, ext4_writepage_trans_blocks(inode));
if (IS_ERR(handle)) {
ret = VM_FAULT_SIGBUS;
goto out;
}
ret = __block_page_mkwrite(vma, vmf, get_block);
if (!ret && ext4_should_journal_data(inode)) {
if (ext4_walk_page_buffers(handle, page_buffers(page), 0,
PAGE_CACHE_SIZE, NULL, do_journal_get_write_access)) {
unlock_page(page);
ret = VM_FAULT_SIGBUS;
ext4_journal_stop(handle);
goto out;
}
ext4_set_inode_state(inode, EXT4_STATE_JDATA);
}
ext4_journal_stop(handle);
if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
goto retry_alloc;
out_ret:
ret = block_page_mkwrite_return(ret);
out:
sb_end_pagefault(inode->i_sb);
return ret;
}