linux/fs/f2fs/data.c

2408 lines
57 KiB
C
Raw Normal View History

/*
* fs/f2fs/data.c
*
* Copyright (c) 2012 Samsung Electronics Co., Ltd.
* http://www.samsung.com/
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
#include <linux/fs.h>
#include <linux/f2fs_fs.h>
#include <linux/buffer_head.h>
#include <linux/mpage.h>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
#include <linux/pagevec.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
#include <linux/prefetch.h>
#include <linux/uio.h>
#include <linux/mm.h>
#include <linux/memcontrol.h>
#include <linux/cleancache.h>
#include <linux/sched/signal.h>
#include "f2fs.h"
#include "node.h"
#include "segment.h"
#include "trace.h"
#include <trace/events/f2fs.h>
static bool __is_cp_guaranteed(struct page *page)
{
struct address_space *mapping = page->mapping;
struct inode *inode;
struct f2fs_sb_info *sbi;
if (!mapping)
return false;
inode = mapping->host;
sbi = F2FS_I_SB(inode);
if (inode->i_ino == F2FS_META_INO(sbi) ||
inode->i_ino == F2FS_NODE_INO(sbi) ||
S_ISDIR(inode->i_mode) ||
is_cold_data(page))
return true;
return false;
}
static void f2fs_read_end_io(struct bio *bio)
{
Merge branch 'for-3.14/core' of git://git.kernel.dk/linux-block Pull core block IO changes from Jens Axboe: "The major piece in here is the immutable bio_ve series from Kent, the rest is fairly minor. It was supposed to go in last round, but various issues pushed it to this release instead. The pull request contains: - Various smaller blk-mq fixes from different folks. Nothing major here, just minor fixes and cleanups. - Fix for a memory leak in the error path in the block ioctl code from Christian Engelmayer. - Header export fix from CaiZhiyong. - Finally the immutable biovec changes from Kent Overstreet. This enables some nice future work on making arbitrarily sized bios possible, and splitting more efficient. Related fixes to immutable bio_vecs: - dm-cache immutable fixup from Mike Snitzer. - btrfs immutable fixup from Muthu Kumar. - bio-integrity fix from Nic Bellinger, which is also going to stable" * 'for-3.14/core' of git://git.kernel.dk/linux-block: (44 commits) xtensa: fixup simdisk driver to work with immutable bio_vecs block/blk-mq-cpu.c: use hotcpu_notifier() blk-mq: for_each_* macro correctness block: Fix memory leak in rw_copy_check_uvector() handling bio-integrity: Fix bio_integrity_verify segment start bug block: remove unrelated header files and export symbol blk-mq: uses page->list incorrectly blk-mq: use __smp_call_function_single directly btrfs: fix missing increment of bi_remaining Revert "block: Warn and free bio if bi_end_io is not set" block: Warn and free bio if bi_end_io is not set blk-mq: fix initializing request's start time block: blk-mq: don't export blk_mq_free_queue() block: blk-mq: make blk_sync_queue support mq block: blk-mq: support draining mq queue dm cache: increment bi_remaining when bi_end_io is restored block: fixup for generic bio chaining block: Really silence spurious compiler warnings block: Silence spurious compiler warnings block: Kill bio_pair_split() ...
2014-01-30 19:19:05 +00:00
struct bio_vec *bvec;
int i;
#ifdef CONFIG_F2FS_FAULT_INJECTION
if (time_to_inject(F2FS_P_SB(bio->bi_io_vec->bv_page), FAULT_IO)) {
f2fs_show_injection_info(FAULT_IO);
bio->bi_status = BLK_STS_IOERR;
}
#endif
if (f2fs_bio_encrypted(bio)) {
if (bio->bi_status) {
fscrypt_release_ctx(bio->bi_private);
} else {
fscrypt_decrypt_bio_pages(bio->bi_private, bio);
return;
}
}
bio_for_each_segment_all(bvec, bio, i) {
struct page *page = bvec->bv_page;
if (!bio->bi_status) {
if (!PageUptodate(page))
SetPageUptodate(page);
} else {
ClearPageUptodate(page);
SetPageError(page);
}
unlock_page(page);
}
bio_put(bio);
}
static void f2fs_write_end_io(struct bio *bio)
{
struct f2fs_sb_info *sbi = bio->bi_private;
Merge branch 'for-3.14/core' of git://git.kernel.dk/linux-block Pull core block IO changes from Jens Axboe: "The major piece in here is the immutable bio_ve series from Kent, the rest is fairly minor. It was supposed to go in last round, but various issues pushed it to this release instead. The pull request contains: - Various smaller blk-mq fixes from different folks. Nothing major here, just minor fixes and cleanups. - Fix for a memory leak in the error path in the block ioctl code from Christian Engelmayer. - Header export fix from CaiZhiyong. - Finally the immutable biovec changes from Kent Overstreet. This enables some nice future work on making arbitrarily sized bios possible, and splitting more efficient. Related fixes to immutable bio_vecs: - dm-cache immutable fixup from Mike Snitzer. - btrfs immutable fixup from Muthu Kumar. - bio-integrity fix from Nic Bellinger, which is also going to stable" * 'for-3.14/core' of git://git.kernel.dk/linux-block: (44 commits) xtensa: fixup simdisk driver to work with immutable bio_vecs block/blk-mq-cpu.c: use hotcpu_notifier() blk-mq: for_each_* macro correctness block: Fix memory leak in rw_copy_check_uvector() handling bio-integrity: Fix bio_integrity_verify segment start bug block: remove unrelated header files and export symbol blk-mq: uses page->list incorrectly blk-mq: use __smp_call_function_single directly btrfs: fix missing increment of bi_remaining Revert "block: Warn and free bio if bi_end_io is not set" block: Warn and free bio if bi_end_io is not set blk-mq: fix initializing request's start time block: blk-mq: don't export blk_mq_free_queue() block: blk-mq: make blk_sync_queue support mq block: blk-mq: support draining mq queue dm cache: increment bi_remaining when bi_end_io is restored block: fixup for generic bio chaining block: Really silence spurious compiler warnings block: Silence spurious compiler warnings block: Kill bio_pair_split() ...
2014-01-30 19:19:05 +00:00
struct bio_vec *bvec;
int i;
Merge branch 'for-3.14/core' of git://git.kernel.dk/linux-block Pull core block IO changes from Jens Axboe: "The major piece in here is the immutable bio_ve series from Kent, the rest is fairly minor. It was supposed to go in last round, but various issues pushed it to this release instead. The pull request contains: - Various smaller blk-mq fixes from different folks. Nothing major here, just minor fixes and cleanups. - Fix for a memory leak in the error path in the block ioctl code from Christian Engelmayer. - Header export fix from CaiZhiyong. - Finally the immutable biovec changes from Kent Overstreet. This enables some nice future work on making arbitrarily sized bios possible, and splitting more efficient. Related fixes to immutable bio_vecs: - dm-cache immutable fixup from Mike Snitzer. - btrfs immutable fixup from Muthu Kumar. - bio-integrity fix from Nic Bellinger, which is also going to stable" * 'for-3.14/core' of git://git.kernel.dk/linux-block: (44 commits) xtensa: fixup simdisk driver to work with immutable bio_vecs block/blk-mq-cpu.c: use hotcpu_notifier() blk-mq: for_each_* macro correctness block: Fix memory leak in rw_copy_check_uvector() handling bio-integrity: Fix bio_integrity_verify segment start bug block: remove unrelated header files and export symbol blk-mq: uses page->list incorrectly blk-mq: use __smp_call_function_single directly btrfs: fix missing increment of bi_remaining Revert "block: Warn and free bio if bi_end_io is not set" block: Warn and free bio if bi_end_io is not set blk-mq: fix initializing request's start time block: blk-mq: don't export blk_mq_free_queue() block: blk-mq: make blk_sync_queue support mq block: blk-mq: support draining mq queue dm cache: increment bi_remaining when bi_end_io is restored block: fixup for generic bio chaining block: Really silence spurious compiler warnings block: Silence spurious compiler warnings block: Kill bio_pair_split() ...
2014-01-30 19:19:05 +00:00
bio_for_each_segment_all(bvec, bio, i) {
struct page *page = bvec->bv_page;
enum count_type type = WB_DATA_TYPE(page);
if (IS_DUMMY_WRITTEN_PAGE(page)) {
set_page_private(page, (unsigned long)NULL);
ClearPagePrivate(page);
unlock_page(page);
mempool_free(page, sbi->write_io_dummy);
if (unlikely(bio->bi_status))
f2fs_stop_checkpoint(sbi, true);
continue;
}
fscrypt_pullback_bio_page(&page, true);
if (unlikely(bio->bi_status)) {
mapping_set_error(page->mapping, -EIO);
if (type == F2FS_WB_CP_DATA)
f2fs_stop_checkpoint(sbi, true);
}
f2fs_bug_on(sbi, page->mapping == NODE_MAPPING(sbi) &&
page->index != nid_of_node(page));
dec_page_count(sbi, type);
clear_cold_data(page);
end_page_writeback(page);
Merge branch 'for-3.14/core' of git://git.kernel.dk/linux-block Pull core block IO changes from Jens Axboe: "The major piece in here is the immutable bio_ve series from Kent, the rest is fairly minor. It was supposed to go in last round, but various issues pushed it to this release instead. The pull request contains: - Various smaller blk-mq fixes from different folks. Nothing major here, just minor fixes and cleanups. - Fix for a memory leak in the error path in the block ioctl code from Christian Engelmayer. - Header export fix from CaiZhiyong. - Finally the immutable biovec changes from Kent Overstreet. This enables some nice future work on making arbitrarily sized bios possible, and splitting more efficient. Related fixes to immutable bio_vecs: - dm-cache immutable fixup from Mike Snitzer. - btrfs immutable fixup from Muthu Kumar. - bio-integrity fix from Nic Bellinger, which is also going to stable" * 'for-3.14/core' of git://git.kernel.dk/linux-block: (44 commits) xtensa: fixup simdisk driver to work with immutable bio_vecs block/blk-mq-cpu.c: use hotcpu_notifier() blk-mq: for_each_* macro correctness block: Fix memory leak in rw_copy_check_uvector() handling bio-integrity: Fix bio_integrity_verify segment start bug block: remove unrelated header files and export symbol blk-mq: uses page->list incorrectly blk-mq: use __smp_call_function_single directly btrfs: fix missing increment of bi_remaining Revert "block: Warn and free bio if bi_end_io is not set" block: Warn and free bio if bi_end_io is not set blk-mq: fix initializing request's start time block: blk-mq: don't export blk_mq_free_queue() block: blk-mq: make blk_sync_queue support mq block: blk-mq: support draining mq queue dm cache: increment bi_remaining when bi_end_io is restored block: fixup for generic bio chaining block: Really silence spurious compiler warnings block: Silence spurious compiler warnings block: Kill bio_pair_split() ...
2014-01-30 19:19:05 +00:00
}
if (!get_pages(sbi, F2FS_WB_CP_DATA) &&
wq_has_sleeper(&sbi->cp_wait))
wake_up(&sbi->cp_wait);
bio_put(bio);
}
/*
* Return true, if pre_bio's bdev is same as its target device.
*/
struct block_device *f2fs_target_device(struct f2fs_sb_info *sbi,
block_t blk_addr, struct bio *bio)
{
struct block_device *bdev = sbi->sb->s_bdev;
int i;
for (i = 0; i < sbi->s_ndevs; i++) {
if (FDEV(i).start_blk <= blk_addr &&
FDEV(i).end_blk >= blk_addr) {
blk_addr -= FDEV(i).start_blk;
bdev = FDEV(i).bdev;
break;
}
}
if (bio) {
bio_set_dev(bio, bdev);
bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(blk_addr);
}
return bdev;
}
int f2fs_target_device_index(struct f2fs_sb_info *sbi, block_t blkaddr)
{
int i;
for (i = 0; i < sbi->s_ndevs; i++)
if (FDEV(i).start_blk <= blkaddr && FDEV(i).end_blk >= blkaddr)
return i;
return 0;
}
static bool __same_bdev(struct f2fs_sb_info *sbi,
block_t blk_addr, struct bio *bio)
{
struct block_device *b = f2fs_target_device(sbi, blk_addr, NULL);
return bio->bi_disk == b->bd_disk && bio->bi_partno == b->bd_partno;
}
/*
* Low-level block read/write IO operations.
*/
static struct bio *__bio_alloc(struct f2fs_sb_info *sbi, block_t blk_addr,
struct writeback_control *wbc,
int npages, bool is_read)
{
struct bio *bio;
bio = f2fs_bio_alloc(sbi, npages, true);
f2fs_target_device(sbi, blk_addr, bio);
bio->bi_end_io = is_read ? f2fs_read_end_io : f2fs_write_end_io;
bio->bi_private = is_read ? NULL : sbi;
if (wbc)
wbc_init_bio(wbc, bio);
return bio;
}
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
static inline void __submit_bio(struct f2fs_sb_info *sbi,
struct bio *bio, enum page_type type)
{
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
if (!is_read_io(bio_op(bio))) {
unsigned int start;
if (f2fs_sb_mounted_blkzoned(sbi->sb) &&
current->plug && (type == DATA || type == NODE))
blk_finish_plug(current->plug);
if (type != DATA && type != NODE)
goto submit_io;
start = bio->bi_iter.bi_size >> F2FS_BLKSIZE_BITS;
start %= F2FS_IO_SIZE(sbi);
if (start == 0)
goto submit_io;
/* fill dummy pages */
for (; start < F2FS_IO_SIZE(sbi); start++) {
struct page *page =
mempool_alloc(sbi->write_io_dummy,
GFP_NOIO | __GFP_ZERO | __GFP_NOFAIL);
f2fs_bug_on(sbi, !page);
SetPagePrivate(page);
set_page_private(page, (unsigned long)DUMMY_WRITTEN_PAGE);
lock_page(page);
if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE)
f2fs_bug_on(sbi, 1);
}
/*
* In the NODE case, we lose next block address chain. So, we
* need to do checkpoint in f2fs_sync_file.
*/
if (type == NODE)
set_sbi_flag(sbi, SBI_NEED_CP);
}
submit_io:
if (is_read_io(bio_op(bio)))
trace_f2fs_submit_read_bio(sbi->sb, type, bio);
else
trace_f2fs_submit_write_bio(sbi->sb, type, bio);
submit_bio(bio);
}
static void __submit_merged_bio(struct f2fs_bio_info *io)
{
struct f2fs_io_info *fio = &io->fio;
if (!io->bio)
return;
bio_set_op_attrs(io->bio, fio->op, fio->op_flags);
if (is_read_io(fio->op))
trace_f2fs_prepare_read_bio(io->sbi->sb, fio->type, io->bio);
else
trace_f2fs_prepare_write_bio(io->sbi->sb, fio->type, io->bio);
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
__submit_bio(io->sbi, io->bio, fio->type);
io->bio = NULL;
}
static bool __has_merged_page(struct f2fs_bio_info *io,
struct inode *inode, nid_t ino, pgoff_t idx)
{
struct bio_vec *bvec;
struct page *target;
int i;
if (!io->bio)
return false;
if (!inode && !ino)
return true;
bio_for_each_segment_all(bvec, io->bio, i) {
if (bvec->bv_page->mapping)
target = bvec->bv_page;
else
target = fscrypt_control_page(bvec->bv_page);
if (idx != target->index)
continue;
if (inode && inode == target->mapping->host)
return true;
if (ino && ino == ino_of_node(target))
return true;
}
return false;
}
static bool has_merged_page(struct f2fs_sb_info *sbi, struct inode *inode,
nid_t ino, pgoff_t idx, enum page_type type)
{
enum page_type btype = PAGE_TYPE_OF_BIO(type);
enum temp_type temp;
struct f2fs_bio_info *io;
bool ret = false;
for (temp = HOT; temp < NR_TEMP_TYPE; temp++) {
io = sbi->write_io[btype] + temp;
down_read(&io->io_rwsem);
ret = __has_merged_page(io, inode, ino, idx);
up_read(&io->io_rwsem);
/* TODO: use HOT temp only for meta pages now. */
if (ret || btype == META)
break;
}
return ret;
}
static void __f2fs_submit_merged_write(struct f2fs_sb_info *sbi,
enum page_type type, enum temp_type temp)
{
enum page_type btype = PAGE_TYPE_OF_BIO(type);
struct f2fs_bio_info *io = sbi->write_io[btype] + temp;
down_write(&io->io_rwsem);
/* change META to META_FLUSH in the checkpoint procedure */
if (type >= META_FLUSH) {
io->fio.type = META_FLUSH;
io->fio.op = REQ_OP_WRITE;
io->fio.op_flags = REQ_META | REQ_PRIO | REQ_SYNC;
if (!test_opt(sbi, NOBARRIER))
io->fio.op_flags |= REQ_PREFLUSH | REQ_FUA;
}
__submit_merged_bio(io);
up_write(&io->io_rwsem);
}
static void __submit_merged_write_cond(struct f2fs_sb_info *sbi,
struct inode *inode, nid_t ino, pgoff_t idx,
enum page_type type, bool force)
{
enum temp_type temp;
if (!force && !has_merged_page(sbi, inode, ino, idx, type))
return;
for (temp = HOT; temp < NR_TEMP_TYPE; temp++) {
__f2fs_submit_merged_write(sbi, type, temp);
/* TODO: use HOT temp only for meta pages now. */
if (type >= META)
break;
}
}
void f2fs_submit_merged_write(struct f2fs_sb_info *sbi, enum page_type type)
{
__submit_merged_write_cond(sbi, NULL, 0, 0, type, true);
}
void f2fs_submit_merged_write_cond(struct f2fs_sb_info *sbi,
struct inode *inode, nid_t ino, pgoff_t idx,
enum page_type type)
{
__submit_merged_write_cond(sbi, inode, ino, idx, type, false);
}
void f2fs_flush_merged_writes(struct f2fs_sb_info *sbi)
{
f2fs_submit_merged_write(sbi, DATA);
f2fs_submit_merged_write(sbi, NODE);
f2fs_submit_merged_write(sbi, META);
}
/*
* Fill the locked page with data located in the block address.
* A caller needs to unlock the page on failure.
*/
int f2fs_submit_page_bio(struct f2fs_io_info *fio)
{
struct bio *bio;
struct page *page = fio->encrypted_page ?
fio->encrypted_page : fio->page;
trace_f2fs_submit_page_bio(page, fio);
f2fs_trace_ios(fio, 0);
/* Allocate a new bio */
bio = __bio_alloc(fio->sbi, fio->new_blkaddr, fio->io_wbc,
1, is_read_io(fio->op));
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) {
bio_put(bio);
return -EFAULT;
}
bio_set_op_attrs(bio, fio->op, fio->op_flags);
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
__submit_bio(fio->sbi, bio, fio->type);
if (!is_read_io(fio->op))
inc_page_count(fio->sbi, WB_DATA_TYPE(fio->page));
return 0;
}
int f2fs_submit_page_write(struct f2fs_io_info *fio)
{
struct f2fs_sb_info *sbi = fio->sbi;
enum page_type btype = PAGE_TYPE_OF_BIO(fio->type);
struct f2fs_bio_info *io = sbi->write_io[btype] + fio->temp;
struct page *bio_page;
int err = 0;
f2fs_bug_on(sbi, is_read_io(fio->op));
down_write(&io->io_rwsem);
next:
if (fio->in_list) {
spin_lock(&io->io_lock);
if (list_empty(&io->io_list)) {
spin_unlock(&io->io_lock);
goto out_fail;
}
fio = list_first_entry(&io->io_list,
struct f2fs_io_info, list);
list_del(&fio->list);
spin_unlock(&io->io_lock);
}
if (fio->old_blkaddr != NEW_ADDR)
verify_block_addr(sbi, fio->old_blkaddr);
verify_block_addr(sbi, fio->new_blkaddr);
bio_page = fio->encrypted_page ? fio->encrypted_page : fio->page;
/* set submitted = true as a return value */
fio->submitted = true;
inc_page_count(sbi, WB_DATA_TYPE(bio_page));
if (io->bio && (io->last_block_in_bio != fio->new_blkaddr - 1 ||
(io->fio.op != fio->op || io->fio.op_flags != fio->op_flags) ||
!__same_bdev(sbi, fio->new_blkaddr, io->bio)))
__submit_merged_bio(io);
alloc_new:
if (io->bio == NULL) {
if ((fio->type == DATA || fio->type == NODE) &&
fio->new_blkaddr & F2FS_IO_SIZE_MASK(sbi)) {
err = -EAGAIN;
dec_page_count(sbi, WB_DATA_TYPE(bio_page));
goto out_fail;
}
io->bio = __bio_alloc(sbi, fio->new_blkaddr, fio->io_wbc,
BIO_MAX_PAGES, false);
io->fio = *fio;
}
if (bio_add_page(io->bio, bio_page, PAGE_SIZE, 0) < PAGE_SIZE) {
__submit_merged_bio(io);
goto alloc_new;
}
if (fio->io_wbc)
wbc_account_io(fio->io_wbc, bio_page, PAGE_SIZE);
io->last_block_in_bio = fio->new_blkaddr;
f2fs_trace_ios(fio, 0);
trace_f2fs_submit_page_write(fio->page, fio);
if (fio->in_list)
goto next;
out_fail:
up_write(&io->io_rwsem);
return err;
}
static struct bio *f2fs_grab_read_bio(struct inode *inode, block_t blkaddr,
unsigned nr_pages)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
struct fscrypt_ctx *ctx = NULL;
struct bio *bio;
if (f2fs_encrypted_file(inode)) {
ctx = fscrypt_get_ctx(inode, GFP_NOFS);
if (IS_ERR(ctx))
return ERR_CAST(ctx);
/* wait the page to be moved by cleaning */
f2fs_wait_on_block_writeback(sbi, blkaddr);
}
bio = f2fs_bio_alloc(sbi, min_t(int, nr_pages, BIO_MAX_PAGES), false);
if (!bio) {
if (ctx)
fscrypt_release_ctx(ctx);
return ERR_PTR(-ENOMEM);
}
f2fs_target_device(sbi, blkaddr, bio);
bio->bi_end_io = f2fs_read_end_io;
bio->bi_private = ctx;
bio_set_op_attrs(bio, REQ_OP_READ, 0);
return bio;
}
/* This can handle encryption stuffs */
static int f2fs_submit_page_read(struct inode *inode, struct page *page,
block_t blkaddr)
{
struct bio *bio = f2fs_grab_read_bio(inode, blkaddr, 1);
if (IS_ERR(bio))
return PTR_ERR(bio);
if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) {
bio_put(bio);
return -EFAULT;
}
__submit_bio(F2FS_I_SB(inode), bio, DATA);
return 0;
}
static void __set_data_blkaddr(struct dnode_of_data *dn)
{
struct f2fs_node *rn = F2FS_NODE(dn->node_page);
__le32 *addr_array;
f2fs: enhance on-disk inode structure scalability This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline to indicate that on-disk structure of current inode is extended. In order to extend, we changed the inode structure a bit: Original one: struct f2fs_inode { ... struct f2fs_extent i_ext; __le32 i_addr[DEF_ADDRS_PER_INODE]; __le32 i_nid[DEF_NIDS_PER_INODE]; } Extended one: struct f2fs_inode { ... struct f2fs_extent i_ext; union { struct { __le16 i_extra_isize; __le16 i_padding; __le32 i_extra_end[0]; }; __le32 i_addr[DEF_ADDRS_PER_INODE]; }; __le32 i_nid[DEF_NIDS_PER_INODE]; } Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of i_addr field for storing i_extra_isize and i_padding. with i_extra_isize, we can calculate actual size of reserved space in i_addr, available attribute fields included in total extra attribute fields for current inode can be described as below: +--------------------+ | .i_mode | | ... | | .i_ext | +--------------------+ | .i_extra_isize |-----+ | .i_padding | | | .i_prjid | | | .i_atime_extra | | | .i_ctime_extra | | | .i_mtime_extra |<----+ | .i_inode_cs |<----- store blkaddr/inline from here | .i_xattr_cs | | ... | +--------------------+ | | | block address | | | +--------------------+ | .i_nid | +--------------------+ | node_footer | | (nid, ino, offset) | +--------------------+ Hence, with this patch, we would enhance scalability of f2fs inode for storing more newly added attribute. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-07-18 16:19:06 +00:00
int base = 0;
if (IS_INODE(dn->node_page) && f2fs_has_extra_attr(dn->inode))
base = get_extra_isize(dn->inode);
/* Get physical address of data block */
addr_array = blkaddr_in_node(rn);
f2fs: enhance on-disk inode structure scalability This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline to indicate that on-disk structure of current inode is extended. In order to extend, we changed the inode structure a bit: Original one: struct f2fs_inode { ... struct f2fs_extent i_ext; __le32 i_addr[DEF_ADDRS_PER_INODE]; __le32 i_nid[DEF_NIDS_PER_INODE]; } Extended one: struct f2fs_inode { ... struct f2fs_extent i_ext; union { struct { __le16 i_extra_isize; __le16 i_padding; __le32 i_extra_end[0]; }; __le32 i_addr[DEF_ADDRS_PER_INODE]; }; __le32 i_nid[DEF_NIDS_PER_INODE]; } Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of i_addr field for storing i_extra_isize and i_padding. with i_extra_isize, we can calculate actual size of reserved space in i_addr, available attribute fields included in total extra attribute fields for current inode can be described as below: +--------------------+ | .i_mode | | ... | | .i_ext | +--------------------+ | .i_extra_isize |-----+ | .i_padding | | | .i_prjid | | | .i_atime_extra | | | .i_ctime_extra | | | .i_mtime_extra |<----+ | .i_inode_cs |<----- store blkaddr/inline from here | .i_xattr_cs | | ... | +--------------------+ | | | block address | | | +--------------------+ | .i_nid | +--------------------+ | node_footer | | (nid, ino, offset) | +--------------------+ Hence, with this patch, we would enhance scalability of f2fs inode for storing more newly added attribute. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-07-18 16:19:06 +00:00
addr_array[base + dn->ofs_in_node] = cpu_to_le32(dn->data_blkaddr);
}
/*
* Lock ordering for the change of data block address:
* ->data_page
* ->node_page
* update block addresses in the node page
*/
void set_data_blkaddr(struct dnode_of_data *dn)
{
f2fs_wait_on_page_writeback(dn->node_page, NODE, true);
__set_data_blkaddr(dn);
if (set_page_dirty(dn->node_page))
dn->node_changed = true;
}
void f2fs_update_data_blkaddr(struct dnode_of_data *dn, block_t blkaddr)
{
dn->data_blkaddr = blkaddr;
set_data_blkaddr(dn);
f2fs_update_extent_cache(dn);
}
/* dn->ofs_in_node will be returned with up-to-date last block pointer */
int reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
int err;
if (!count)
return 0;
if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
return -EPERM;
if (unlikely((err = inc_valid_block_count(sbi, dn->inode, &count))))
return err;
trace_f2fs_reserve_new_blocks(dn->inode, dn->nid,
dn->ofs_in_node, count);
f2fs_wait_on_page_writeback(dn->node_page, NODE, true);
for (; count > 0; dn->ofs_in_node++) {
f2fs: enhance on-disk inode structure scalability This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline to indicate that on-disk structure of current inode is extended. In order to extend, we changed the inode structure a bit: Original one: struct f2fs_inode { ... struct f2fs_extent i_ext; __le32 i_addr[DEF_ADDRS_PER_INODE]; __le32 i_nid[DEF_NIDS_PER_INODE]; } Extended one: struct f2fs_inode { ... struct f2fs_extent i_ext; union { struct { __le16 i_extra_isize; __le16 i_padding; __le32 i_extra_end[0]; }; __le32 i_addr[DEF_ADDRS_PER_INODE]; }; __le32 i_nid[DEF_NIDS_PER_INODE]; } Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of i_addr field for storing i_extra_isize and i_padding. with i_extra_isize, we can calculate actual size of reserved space in i_addr, available attribute fields included in total extra attribute fields for current inode can be described as below: +--------------------+ | .i_mode | | ... | | .i_ext | +--------------------+ | .i_extra_isize |-----+ | .i_padding | | | .i_prjid | | | .i_atime_extra | | | .i_ctime_extra | | | .i_mtime_extra |<----+ | .i_inode_cs |<----- store blkaddr/inline from here | .i_xattr_cs | | ... | +--------------------+ | | | block address | | | +--------------------+ | .i_nid | +--------------------+ | node_footer | | (nid, ino, offset) | +--------------------+ Hence, with this patch, we would enhance scalability of f2fs inode for storing more newly added attribute. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-07-18 16:19:06 +00:00
block_t blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
if (blkaddr == NULL_ADDR) {
dn->data_blkaddr = NEW_ADDR;
__set_data_blkaddr(dn);
count--;
}
}
if (set_page_dirty(dn->node_page))
dn->node_changed = true;
return 0;
}
/* Should keep dn->ofs_in_node unchanged */
int reserve_new_block(struct dnode_of_data *dn)
{
unsigned int ofs_in_node = dn->ofs_in_node;
int ret;
ret = reserve_new_blocks(dn, 1);
dn->ofs_in_node = ofs_in_node;
return ret;
}
int f2fs_reserve_block(struct dnode_of_data *dn, pgoff_t index)
{
bool need_put = dn->inode_page ? false : true;
int err;
err = get_dnode_of_data(dn, index, ALLOC_NODE);
if (err)
return err;
if (dn->data_blkaddr == NULL_ADDR)
err = reserve_new_block(dn);
if (err || need_put)
f2fs_put_dnode(dn);
return err;
}
int f2fs_get_block(struct dnode_of_data *dn, pgoff_t index)
{
struct extent_info ei = {0,0,0};
struct inode *inode = dn->inode;
if (f2fs_lookup_extent_cache(inode, index, &ei)) {
dn->data_blkaddr = ei.blk + index - ei.fofs;
return 0;
}
return f2fs_reserve_block(dn, index);
}
struct page *get_read_data_page(struct inode *inode, pgoff_t index,
int op_flags, bool for_write)
{
struct address_space *mapping = inode->i_mapping;
struct dnode_of_data dn;
struct page *page;
struct extent_info ei = {0,0,0};
int err;
page = f2fs_grab_cache_page(mapping, index, for_write);
if (!page)
return ERR_PTR(-ENOMEM);
if (f2fs_lookup_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
goto got_it;
}
set_new_dnode(&dn, inode, NULL, NULL, 0);
err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
if (err)
goto put_err;
f2fs_put_dnode(&dn);
if (unlikely(dn.data_blkaddr == NULL_ADDR)) {
err = -ENOENT;
goto put_err;
}
got_it:
if (PageUptodate(page)) {
unlock_page(page);
return page;
}
/*
* A new dentry page is allocated but not able to be written, since its
* new inode page couldn't be allocated due to -ENOSPC.
* In such the case, its blkaddr can be remained as NEW_ADDR.
* see, f2fs_add_link -> get_new_data_page -> init_inode_metadata.
*/
if (dn.data_blkaddr == NEW_ADDR) {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, 0, PAGE_SIZE);
if (!PageUptodate(page))
SetPageUptodate(page);
unlock_page(page);
return page;
}
err = f2fs_submit_page_read(inode, page, dn.data_blkaddr);
if (err)
goto put_err;
return page;
put_err:
f2fs_put_page(page, 1);
return ERR_PTR(err);
}
struct page *find_data_page(struct inode *inode, pgoff_t index)
{
struct address_space *mapping = inode->i_mapping;
struct page *page;
page = find_get_page(mapping, index);
if (page && PageUptodate(page))
return page;
f2fs_put_page(page, 0);
page = get_read_data_page(inode, index, 0, false);
if (IS_ERR(page))
return page;
if (PageUptodate(page))
return page;
wait_on_page_locked(page);
if (unlikely(!PageUptodate(page))) {
f2fs_put_page(page, 0);
return ERR_PTR(-EIO);
}
return page;
}
/*
* If it tries to access a hole, return an error.
* Because, the callers, functions in dir.c and GC, should be able to know
* whether this page exists or not.
*/
struct page *get_lock_data_page(struct inode *inode, pgoff_t index,
bool for_write)
{
struct address_space *mapping = inode->i_mapping;
struct page *page;
repeat:
page = get_read_data_page(inode, index, 0, for_write);
if (IS_ERR(page))
return page;
/* wait for read completion */
lock_page(page);
if (unlikely(page->mapping != mapping)) {
f2fs_put_page(page, 1);
goto repeat;
}
if (unlikely(!PageUptodate(page))) {
f2fs_put_page(page, 1);
return ERR_PTR(-EIO);
}
return page;
}
/*
* Caller ensures that this data page is never allocated.
* A new zero-filled data page is allocated in the page cache.
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
*
* Also, caller should grab and release a rwsem by calling f2fs_lock_op() and
* f2fs_unlock_op().
* Note that, ipage is set only by make_empty_dir, and if any error occur,
* ipage should be released by this function.
*/
struct page *get_new_data_page(struct inode *inode,
struct page *ipage, pgoff_t index, bool new_i_size)
{
struct address_space *mapping = inode->i_mapping;
struct page *page;
struct dnode_of_data dn;
int err;
page = f2fs_grab_cache_page(mapping, index, true);
if (!page) {
/*
* before exiting, we should make sure ipage will be released
* if any error occur.
*/
f2fs_put_page(ipage, 1);
return ERR_PTR(-ENOMEM);
}
set_new_dnode(&dn, inode, ipage, NULL, 0);
err = f2fs_reserve_block(&dn, index);
if (err) {
f2fs_put_page(page, 1);
return ERR_PTR(err);
}
if (!ipage)
f2fs_put_dnode(&dn);
if (PageUptodate(page))
goto got_it;
if (dn.data_blkaddr == NEW_ADDR) {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, 0, PAGE_SIZE);
if (!PageUptodate(page))
SetPageUptodate(page);
} else {
f2fs_put_page(page, 1);
/* if ipage exists, blkaddr should be NEW_ADDR */
f2fs_bug_on(F2FS_I_SB(inode), ipage);
page = get_lock_data_page(inode, index, true);
if (IS_ERR(page))
return page;
}
got_it:
if (new_i_size && i_size_read(inode) <
((loff_t)(index + 1) << PAGE_SHIFT))
f2fs_i_size_write(inode, ((loff_t)(index + 1) << PAGE_SHIFT));
return page;
}
static int __allocate_data_block(struct dnode_of_data *dn, int seg_type)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode);
struct f2fs_summary sum;
struct node_info ni;
pgoff_t fofs;
blkcnt_t count = 1;
int err;
if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
return -EPERM;
f2fs: enhance on-disk inode structure scalability This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline to indicate that on-disk structure of current inode is extended. In order to extend, we changed the inode structure a bit: Original one: struct f2fs_inode { ... struct f2fs_extent i_ext; __le32 i_addr[DEF_ADDRS_PER_INODE]; __le32 i_nid[DEF_NIDS_PER_INODE]; } Extended one: struct f2fs_inode { ... struct f2fs_extent i_ext; union { struct { __le16 i_extra_isize; __le16 i_padding; __le32 i_extra_end[0]; }; __le32 i_addr[DEF_ADDRS_PER_INODE]; }; __le32 i_nid[DEF_NIDS_PER_INODE]; } Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of i_addr field for storing i_extra_isize and i_padding. with i_extra_isize, we can calculate actual size of reserved space in i_addr, available attribute fields included in total extra attribute fields for current inode can be described as below: +--------------------+ | .i_mode | | ... | | .i_ext | +--------------------+ | .i_extra_isize |-----+ | .i_padding | | | .i_prjid | | | .i_atime_extra | | | .i_ctime_extra | | | .i_mtime_extra |<----+ | .i_inode_cs |<----- store blkaddr/inline from here | .i_xattr_cs | | ... | +--------------------+ | | | block address | | | +--------------------+ | .i_nid | +--------------------+ | node_footer | | (nid, ino, offset) | +--------------------+ Hence, with this patch, we would enhance scalability of f2fs inode for storing more newly added attribute. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-07-18 16:19:06 +00:00
dn->data_blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
if (dn->data_blkaddr == NEW_ADDR)
goto alloc;
if (unlikely((err = inc_valid_block_count(sbi, dn->inode, &count))))
return err;
alloc:
get_node_info(sbi, dn->nid, &ni);
set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version);
allocate_data_block(sbi, NULL, dn->data_blkaddr, &dn->data_blkaddr,
&sum, seg_type, NULL, false);
set_data_blkaddr(dn);
/* update i_size */
fofs = start_bidx_of_node(ofs_of_node(dn->node_page), dn->inode) +
dn->ofs_in_node;
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
if (i_size_read(dn->inode) < ((loff_t)(fofs + 1) << PAGE_SHIFT))
f2fs_i_size_write(dn->inode,
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
((loff_t)(fofs + 1) << PAGE_SHIFT));
return 0;
}
static inline bool __force_buffered_io(struct inode *inode, int rw)
{
return (f2fs_encrypted_file(inode) ||
(rw == WRITE && test_opt(F2FS_I_SB(inode), LFS)) ||
F2FS_I_SB(inode)->s_ndevs);
}
int f2fs_preallocate_blocks(struct kiocb *iocb, struct iov_iter *from)
{
struct inode *inode = file_inode(iocb->ki_filp);
struct f2fs_map_blocks map;
int flag;
int err = 0;
bool direct_io = iocb->ki_flags & IOCB_DIRECT;
/* convert inline data for Direct I/O*/
if (direct_io) {
err = f2fs_convert_inline_inode(inode);
if (err)
return err;
}
if (is_inode_flag_set(inode, FI_NO_PREALLOC))
return 0;
map.m_lblk = F2FS_BLK_ALIGN(iocb->ki_pos);
map.m_len = F2FS_BYTES_TO_BLK(iocb->ki_pos + iov_iter_count(from));
if (map.m_len > map.m_lblk)
map.m_len -= map.m_lblk;
else
map.m_len = 0;
map.m_next_pgofs = NULL;
map.m_seg_type = NO_CHECK_TYPE;
if (direct_io) {
map.m_seg_type = rw_hint_to_seg_type(iocb->ki_hint);
flag = __force_buffered_io(inode, WRITE) ?
F2FS_GET_BLOCK_PRE_AIO :
F2FS_GET_BLOCK_PRE_DIO;
goto map_blocks;
}
if (iocb->ki_pos + iov_iter_count(from) > MAX_INLINE_DATA(inode)) {
err = f2fs_convert_inline_inode(inode);
if (err)
return err;
}
if (f2fs_has_inline_data(inode))
return err;
flag = F2FS_GET_BLOCK_PRE_AIO;
map_blocks:
err = f2fs_map_blocks(inode, &map, 1, flag);
if (map.m_len > 0 && err == -ENOSPC) {
if (!direct_io)
set_inode_flag(inode, FI_NO_PREALLOC);
err = 0;
}
return err;
}
static inline void __do_map_lock(struct f2fs_sb_info *sbi, int flag, bool lock)
{
if (flag == F2FS_GET_BLOCK_PRE_AIO) {
if (lock)
down_read(&sbi->node_change);
else
up_read(&sbi->node_change);
} else {
if (lock)
f2fs_lock_op(sbi);
else
f2fs_unlock_op(sbi);
}
}
/*
* f2fs_map_blocks() now supported readahead/bmap/rw direct_IO with
* f2fs_map_blocks structure.
* If original data blocks are allocated, then give them to blockdev.
* Otherwise,
* a. preallocate requested block addresses
* b. do not use extent cache for better performance
* c. give the block addresses to blockdev
*/
int f2fs_map_blocks(struct inode *inode, struct f2fs_map_blocks *map,
int create, int flag)
{
unsigned int maxblocks = map->m_len;
struct dnode_of_data dn;
f2fs: fix to handle io error in ->direct_IO Here is a oops reported as following message when testing generic/019 of xfstest: ------------[ cut here ]------------ kernel BUG at /home/yuchao/git/f2fs-dev/segment.c:882! invalid opcode: 0000 [#1] SMP Modules linked in: zram lz4_compress lz4_decompress f2fs(O) ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_def CPU: 2 PID: 25441 Comm: fio Tainted: G O 4.3.0-rc1+ #6 Hardware name: Hewlett-Packard HP Z220 CMT Workstation/1790, BIOS K51 v01.61 05/16/2013 task: ffff8803f4e85580 ti: ffff8803fd61c000 task.ti: ffff8803fd61c000 RIP: 0010:[<ffffffffa0784981>] [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP: 0018:ffff8803fd61f918 EFLAGS: 00010246 RAX: 00000000000007ed RBX: 0000000000000224 RCX: 000000000000001f RDX: 0000000000000800 RSI: ffffffffffffffff RDI: ffff8803f56f4300 RBP: ffff8803fd61f978 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000024 R11: ffff8800d23bbd78 R12: ffff8800d0ef0000 R13: 0000000000000224 R14: 0000000000000000 R15: 0000000000000001 FS: 00007f827ff85700(0000) GS:ffff88041ea80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffff600000 CR3: 00000003fef17000 CR4: 00000000001406e0 Stack: 000007ea00000002 0000000100000001 ffff8803f6456248 000007ed0000002b 0000000000000224 ffff880404d1aa20 ffff8803fd61f9c8 ffff8800d0ef0000 ffff8803f6456248 0000000000000001 00000000ffffffff ffffffffa078f358 Call Trace: [<ffffffffa0785b87>] allocate_segment_by_default+0x1a7/0x1f0 [f2fs] [<ffffffffa078322c>] allocate_data_block+0x17c/0x360 [f2fs] [<ffffffffa0779521>] __allocate_data_block+0x131/0x1d0 [f2fs] [<ffffffffa077a995>] f2fs_direct_IO+0x4b5/0x580 [f2fs] [<ffffffff811510ae>] generic_file_direct_write+0xae/0x160 [<ffffffff811518f5>] __generic_file_write_iter+0xd5/0x1f0 [<ffffffff81151e07>] generic_file_write_iter+0xf7/0x200 [<ffffffff81319e38>] ? apparmor_file_permission+0x18/0x20 [<ffffffffa0768480>] ? f2fs_fallocate+0x1190/0x1190 [f2fs] [<ffffffffa07684c6>] f2fs_file_write_iter+0x46/0x90 [f2fs] [<ffffffff8120b4fe>] aio_run_iocb+0x1ee/0x290 [<ffffffff81700f7e>] ? mutex_lock+0x1e/0x50 [<ffffffff8120a1d7>] ? aio_read_events+0x207/0x2b0 [<ffffffff8120b913>] do_io_submit+0x373/0x630 [<ffffffff8120a4f6>] ? SyS_io_getevents+0x56/0xb0 [<ffffffff8120bbe0>] SyS_io_submit+0x10/0x20 [<ffffffff81703857>] entry_SYSCALL_64_fastpath+0x12/0x6a Code: 45 c8 48 8b 78 10 e8 9f 23 bf e0 41 8b 8c 24 cc 03 00 00 89 c7 31 d2 89 c6 89 d8 29 df f7 f1 29 d1 39 cf 0f 83 be fd ff ff eb RIP [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP <ffff8803fd61f918> ---[ end trace 2e577d7f711ddb86 ]--- The reason is that: in the test of generic/019, we will trigger a manmade IO error in block layer through debugfs, after that, prefree segment will no longer be freed, because we always skip doing gc or checkpoint when there occurs an IO error. Meanwhile fio with aio engine generated a large number of direct IOs, which continue allocating spaces in free segment until we run out of them, eventually, results in panic in new_curseg as no more free segment was found. So, this patch changes to return EIO in direct_IO for this condition. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2015-09-21 12:17:52 +00:00
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
int mode = create ? ALLOC_NODE : LOOKUP_NODE;
pgoff_t pgofs, end_offset, end;
int err = 0, ofs = 1;
unsigned int ofs_in_node, last_ofs_in_node;
blkcnt_t prealloc;
struct extent_info ei = {0,0,0};
block_t blkaddr;
if (!maxblocks)
return 0;
map->m_len = 0;
map->m_flags = 0;
/* it only supports block size == page size */
pgofs = (pgoff_t)map->m_lblk;
end = pgofs + maxblocks;
if (!create && f2fs_lookup_extent_cache(inode, pgofs, &ei)) {
map->m_pblk = ei.blk + pgofs - ei.fofs;
map->m_len = min((pgoff_t)maxblocks, ei.fofs + ei.len - pgofs);
map->m_flags = F2FS_MAP_MAPPED;
goto out;
}
next_dnode:
if (create)
__do_map_lock(sbi, flag, true);
/* When reading holes, we need its node page */
set_new_dnode(&dn, inode, NULL, NULL, 0);
err = get_dnode_of_data(&dn, pgofs, mode);
if (err) {
if (flag == F2FS_GET_BLOCK_BMAP)
map->m_pblk = 0;
if (err == -ENOENT) {
err = 0;
if (map->m_next_pgofs)
*map->m_next_pgofs =
get_next_page_offset(&dn, pgofs);
}
goto unlock_out;
}
prealloc = 0;
last_ofs_in_node = ofs_in_node = dn.ofs_in_node;
end_offset = ADDRS_PER_PAGE(dn.node_page, inode);
next_block:
f2fs: enhance on-disk inode structure scalability This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline to indicate that on-disk structure of current inode is extended. In order to extend, we changed the inode structure a bit: Original one: struct f2fs_inode { ... struct f2fs_extent i_ext; __le32 i_addr[DEF_ADDRS_PER_INODE]; __le32 i_nid[DEF_NIDS_PER_INODE]; } Extended one: struct f2fs_inode { ... struct f2fs_extent i_ext; union { struct { __le16 i_extra_isize; __le16 i_padding; __le32 i_extra_end[0]; }; __le32 i_addr[DEF_ADDRS_PER_INODE]; }; __le32 i_nid[DEF_NIDS_PER_INODE]; } Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of i_addr field for storing i_extra_isize and i_padding. with i_extra_isize, we can calculate actual size of reserved space in i_addr, available attribute fields included in total extra attribute fields for current inode can be described as below: +--------------------+ | .i_mode | | ... | | .i_ext | +--------------------+ | .i_extra_isize |-----+ | .i_padding | | | .i_prjid | | | .i_atime_extra | | | .i_ctime_extra | | | .i_mtime_extra |<----+ | .i_inode_cs |<----- store blkaddr/inline from here | .i_xattr_cs | | ... | +--------------------+ | | | block address | | | +--------------------+ | .i_nid | +--------------------+ | node_footer | | (nid, ino, offset) | +--------------------+ Hence, with this patch, we would enhance scalability of f2fs inode for storing more newly added attribute. Signed-off-by: Chao Yu <yuchao0@huawei.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-07-18 16:19:06 +00:00
blkaddr = datablock_addr(dn.inode, dn.node_page, dn.ofs_in_node);
if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR) {
if (create) {
f2fs: fix to handle io error in ->direct_IO Here is a oops reported as following message when testing generic/019 of xfstest: ------------[ cut here ]------------ kernel BUG at /home/yuchao/git/f2fs-dev/segment.c:882! invalid opcode: 0000 [#1] SMP Modules linked in: zram lz4_compress lz4_decompress f2fs(O) ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_def CPU: 2 PID: 25441 Comm: fio Tainted: G O 4.3.0-rc1+ #6 Hardware name: Hewlett-Packard HP Z220 CMT Workstation/1790, BIOS K51 v01.61 05/16/2013 task: ffff8803f4e85580 ti: ffff8803fd61c000 task.ti: ffff8803fd61c000 RIP: 0010:[<ffffffffa0784981>] [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP: 0018:ffff8803fd61f918 EFLAGS: 00010246 RAX: 00000000000007ed RBX: 0000000000000224 RCX: 000000000000001f RDX: 0000000000000800 RSI: ffffffffffffffff RDI: ffff8803f56f4300 RBP: ffff8803fd61f978 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000024 R11: ffff8800d23bbd78 R12: ffff8800d0ef0000 R13: 0000000000000224 R14: 0000000000000000 R15: 0000000000000001 FS: 00007f827ff85700(0000) GS:ffff88041ea80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffff600000 CR3: 00000003fef17000 CR4: 00000000001406e0 Stack: 000007ea00000002 0000000100000001 ffff8803f6456248 000007ed0000002b 0000000000000224 ffff880404d1aa20 ffff8803fd61f9c8 ffff8800d0ef0000 ffff8803f6456248 0000000000000001 00000000ffffffff ffffffffa078f358 Call Trace: [<ffffffffa0785b87>] allocate_segment_by_default+0x1a7/0x1f0 [f2fs] [<ffffffffa078322c>] allocate_data_block+0x17c/0x360 [f2fs] [<ffffffffa0779521>] __allocate_data_block+0x131/0x1d0 [f2fs] [<ffffffffa077a995>] f2fs_direct_IO+0x4b5/0x580 [f2fs] [<ffffffff811510ae>] generic_file_direct_write+0xae/0x160 [<ffffffff811518f5>] __generic_file_write_iter+0xd5/0x1f0 [<ffffffff81151e07>] generic_file_write_iter+0xf7/0x200 [<ffffffff81319e38>] ? apparmor_file_permission+0x18/0x20 [<ffffffffa0768480>] ? f2fs_fallocate+0x1190/0x1190 [f2fs] [<ffffffffa07684c6>] f2fs_file_write_iter+0x46/0x90 [f2fs] [<ffffffff8120b4fe>] aio_run_iocb+0x1ee/0x290 [<ffffffff81700f7e>] ? mutex_lock+0x1e/0x50 [<ffffffff8120a1d7>] ? aio_read_events+0x207/0x2b0 [<ffffffff8120b913>] do_io_submit+0x373/0x630 [<ffffffff8120a4f6>] ? SyS_io_getevents+0x56/0xb0 [<ffffffff8120bbe0>] SyS_io_submit+0x10/0x20 [<ffffffff81703857>] entry_SYSCALL_64_fastpath+0x12/0x6a Code: 45 c8 48 8b 78 10 e8 9f 23 bf e0 41 8b 8c 24 cc 03 00 00 89 c7 31 d2 89 c6 89 d8 29 df f7 f1 29 d1 39 cf 0f 83 be fd ff ff eb RIP [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP <ffff8803fd61f918> ---[ end trace 2e577d7f711ddb86 ]--- The reason is that: in the test of generic/019, we will trigger a manmade IO error in block layer through debugfs, after that, prefree segment will no longer be freed, because we always skip doing gc or checkpoint when there occurs an IO error. Meanwhile fio with aio engine generated a large number of direct IOs, which continue allocating spaces in free segment until we run out of them, eventually, results in panic in new_curseg as no more free segment was found. So, this patch changes to return EIO in direct_IO for this condition. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2015-09-21 12:17:52 +00:00
if (unlikely(f2fs_cp_error(sbi))) {
err = -EIO;
goto sync_out;
f2fs: fix to handle io error in ->direct_IO Here is a oops reported as following message when testing generic/019 of xfstest: ------------[ cut here ]------------ kernel BUG at /home/yuchao/git/f2fs-dev/segment.c:882! invalid opcode: 0000 [#1] SMP Modules linked in: zram lz4_compress lz4_decompress f2fs(O) ip6table_filter ip6_tables ebtable_nat ebtables nf_conntrack_ipv4 nf_def CPU: 2 PID: 25441 Comm: fio Tainted: G O 4.3.0-rc1+ #6 Hardware name: Hewlett-Packard HP Z220 CMT Workstation/1790, BIOS K51 v01.61 05/16/2013 task: ffff8803f4e85580 ti: ffff8803fd61c000 task.ti: ffff8803fd61c000 RIP: 0010:[<ffffffffa0784981>] [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP: 0018:ffff8803fd61f918 EFLAGS: 00010246 RAX: 00000000000007ed RBX: 0000000000000224 RCX: 000000000000001f RDX: 0000000000000800 RSI: ffffffffffffffff RDI: ffff8803f56f4300 RBP: ffff8803fd61f978 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000024 R11: ffff8800d23bbd78 R12: ffff8800d0ef0000 R13: 0000000000000224 R14: 0000000000000000 R15: 0000000000000001 FS: 00007f827ff85700(0000) GS:ffff88041ea80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffffffff600000 CR3: 00000003fef17000 CR4: 00000000001406e0 Stack: 000007ea00000002 0000000100000001 ffff8803f6456248 000007ed0000002b 0000000000000224 ffff880404d1aa20 ffff8803fd61f9c8 ffff8800d0ef0000 ffff8803f6456248 0000000000000001 00000000ffffffff ffffffffa078f358 Call Trace: [<ffffffffa0785b87>] allocate_segment_by_default+0x1a7/0x1f0 [f2fs] [<ffffffffa078322c>] allocate_data_block+0x17c/0x360 [f2fs] [<ffffffffa0779521>] __allocate_data_block+0x131/0x1d0 [f2fs] [<ffffffffa077a995>] f2fs_direct_IO+0x4b5/0x580 [f2fs] [<ffffffff811510ae>] generic_file_direct_write+0xae/0x160 [<ffffffff811518f5>] __generic_file_write_iter+0xd5/0x1f0 [<ffffffff81151e07>] generic_file_write_iter+0xf7/0x200 [<ffffffff81319e38>] ? apparmor_file_permission+0x18/0x20 [<ffffffffa0768480>] ? f2fs_fallocate+0x1190/0x1190 [f2fs] [<ffffffffa07684c6>] f2fs_file_write_iter+0x46/0x90 [f2fs] [<ffffffff8120b4fe>] aio_run_iocb+0x1ee/0x290 [<ffffffff81700f7e>] ? mutex_lock+0x1e/0x50 [<ffffffff8120a1d7>] ? aio_read_events+0x207/0x2b0 [<ffffffff8120b913>] do_io_submit+0x373/0x630 [<ffffffff8120a4f6>] ? SyS_io_getevents+0x56/0xb0 [<ffffffff8120bbe0>] SyS_io_submit+0x10/0x20 [<ffffffff81703857>] entry_SYSCALL_64_fastpath+0x12/0x6a Code: 45 c8 48 8b 78 10 e8 9f 23 bf e0 41 8b 8c 24 cc 03 00 00 89 c7 31 d2 89 c6 89 d8 29 df f7 f1 29 d1 39 cf 0f 83 be fd ff ff eb RIP [<ffffffffa0784981>] new_curseg+0x321/0x330 [f2fs] RSP <ffff8803fd61f918> ---[ end trace 2e577d7f711ddb86 ]--- The reason is that: in the test of generic/019, we will trigger a manmade IO error in block layer through debugfs, after that, prefree segment will no longer be freed, because we always skip doing gc or checkpoint when there occurs an IO error. Meanwhile fio with aio engine generated a large number of direct IOs, which continue allocating spaces in free segment until we run out of them, eventually, results in panic in new_curseg as no more free segment was found. So, this patch changes to return EIO in direct_IO for this condition. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2015-09-21 12:17:52 +00:00
}
if (flag == F2FS_GET_BLOCK_PRE_AIO) {
if (blkaddr == NULL_ADDR) {
prealloc++;
last_ofs_in_node = dn.ofs_in_node;
}
} else {
err = __allocate_data_block(&dn,
map->m_seg_type);
if (!err)
set_inode_flag(inode, FI_APPEND_WRITE);
}
if (err)
goto sync_out;
map->m_flags |= F2FS_MAP_NEW;
blkaddr = dn.data_blkaddr;
} else {
if (flag == F2FS_GET_BLOCK_BMAP) {
map->m_pblk = 0;
goto sync_out;
}
if (flag == F2FS_GET_BLOCK_FIEMAP &&
blkaddr == NULL_ADDR) {
if (map->m_next_pgofs)
*map->m_next_pgofs = pgofs + 1;
goto sync_out;
}
if (flag != F2FS_GET_BLOCK_FIEMAP)
goto sync_out;
}
}
if (flag == F2FS_GET_BLOCK_PRE_AIO)
goto skip;
if (map->m_len == 0) {
/* preallocated unwritten block should be mapped for fiemap. */
if (blkaddr == NEW_ADDR)
map->m_flags |= F2FS_MAP_UNWRITTEN;
map->m_flags |= F2FS_MAP_MAPPED;
map->m_pblk = blkaddr;
map->m_len = 1;
} else if ((map->m_pblk != NEW_ADDR &&
blkaddr == (map->m_pblk + ofs)) ||
(map->m_pblk == NEW_ADDR && blkaddr == NEW_ADDR) ||
flag == F2FS_GET_BLOCK_PRE_DIO) {
ofs++;
map->m_len++;
} else {
goto sync_out;
}
skip:
dn.ofs_in_node++;
pgofs++;
/* preallocate blocks in batch for one dnode page */
if (flag == F2FS_GET_BLOCK_PRE_AIO &&
(pgofs == end || dn.ofs_in_node == end_offset)) {
dn.ofs_in_node = ofs_in_node;
err = reserve_new_blocks(&dn, prealloc);
if (err)
goto sync_out;
map->m_len += dn.ofs_in_node - ofs_in_node;
if (prealloc && dn.ofs_in_node != last_ofs_in_node + 1) {
err = -ENOSPC;
goto sync_out;
}
dn.ofs_in_node = end_offset;
}
if (pgofs >= end)
goto sync_out;
else if (dn.ofs_in_node < end_offset)
goto next_block;
f2fs_put_dnode(&dn);
if (create) {
__do_map_lock(sbi, flag, false);
f2fs_balance_fs(sbi, dn.node_changed);
}
goto next_dnode;
sync_out:
f2fs_put_dnode(&dn);
unlock_out:
if (create) {
__do_map_lock(sbi, flag, false);
f2fs_balance_fs(sbi, dn.node_changed);
}
out:
trace_f2fs_map_blocks(inode, map, err);
return err;
}
static int __get_data_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh, int create, int flag,
pgoff_t *next_pgofs, int seg_type)
{
struct f2fs_map_blocks map;
int err;
map.m_lblk = iblock;
map.m_len = bh->b_size >> inode->i_blkbits;
map.m_next_pgofs = next_pgofs;
map.m_seg_type = seg_type;
err = f2fs_map_blocks(inode, &map, create, flag);
if (!err) {
map_bh(bh, inode->i_sb, map.m_pblk);
bh->b_state = (bh->b_state & ~F2FS_MAP_FLAGS) | map.m_flags;
bh->b_size = (u64)map.m_len << inode->i_blkbits;
}
return err;
}
static int get_data_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create, int flag,
pgoff_t *next_pgofs)
{
return __get_data_block(inode, iblock, bh_result, create,
flag, next_pgofs,
NO_CHECK_TYPE);
}
static int get_data_block_dio(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
return __get_data_block(inode, iblock, bh_result, create,
F2FS_GET_BLOCK_DEFAULT, NULL,
rw_hint_to_seg_type(
inode->i_write_hint));
}
static int get_data_block_bmap(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
/* Block number less than F2FS MAX BLOCKS */
if (unlikely(iblock >= F2FS_I_SB(inode)->max_file_blocks))
return -EFBIG;
return __get_data_block(inode, iblock, bh_result, create,
F2FS_GET_BLOCK_BMAP, NULL,
NO_CHECK_TYPE);
}
static inline sector_t logical_to_blk(struct inode *inode, loff_t offset)
{
return (offset >> inode->i_blkbits);
}
static inline loff_t blk_to_logical(struct inode *inode, sector_t blk)
{
return (blk << inode->i_blkbits);
}
static int f2fs_xattr_fiemap(struct inode *inode,
struct fiemap_extent_info *fieinfo)
{
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
struct page *page;
struct node_info ni;
__u64 phys = 0, len;
__u32 flags;
nid_t xnid = F2FS_I(inode)->i_xattr_nid;
int err = 0;
if (f2fs_has_inline_xattr(inode)) {
int offset;
page = f2fs_grab_cache_page(NODE_MAPPING(sbi),
inode->i_ino, false);
if (!page)
return -ENOMEM;
get_node_info(sbi, inode->i_ino, &ni);
phys = (__u64)blk_to_logical(inode, ni.blk_addr);
offset = offsetof(struct f2fs_inode, i_addr) +
sizeof(__le32) * (DEF_ADDRS_PER_INODE -
F2FS_INLINE_XATTR_ADDRS(inode));
phys += offset;
len = inline_xattr_size(inode);
f2fs_put_page(page, 1);
flags = FIEMAP_EXTENT_DATA_INLINE | FIEMAP_EXTENT_NOT_ALIGNED;
if (!xnid)
flags |= FIEMAP_EXTENT_LAST;
err = fiemap_fill_next_extent(fieinfo, 0, phys, len, flags);
if (err || err == 1)
return err;
}
if (xnid) {
page = f2fs_grab_cache_page(NODE_MAPPING(sbi), xnid, false);
if (!page)
return -ENOMEM;
get_node_info(sbi, xnid, &ni);
phys = (__u64)blk_to_logical(inode, ni.blk_addr);
len = inode->i_sb->s_blocksize;
f2fs_put_page(page, 1);
flags = FIEMAP_EXTENT_LAST;
}
if (phys)
err = fiemap_fill_next_extent(fieinfo, 0, phys, len, flags);
return (err < 0 ? err : 0);
}
int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
u64 start, u64 len)
{
struct buffer_head map_bh;
sector_t start_blk, last_blk;
pgoff_t next_pgofs;
u64 logical = 0, phys = 0, size = 0;
u32 flags = 0;
int ret = 0;
ret = fiemap_check_flags(fieinfo, FIEMAP_FLAG_SYNC | FIEMAP_FLAG_XATTR);
if (ret)
return ret;
inode_lock(inode);
if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
ret = f2fs_xattr_fiemap(inode, fieinfo);
goto out;
}
if (f2fs_has_inline_data(inode)) {
ret = f2fs_inline_data_fiemap(inode, fieinfo, start, len);
if (ret != -EAGAIN)
goto out;
}
if (logical_to_blk(inode, len) == 0)
len = blk_to_logical(inode, 1);
start_blk = logical_to_blk(inode, start);
last_blk = logical_to_blk(inode, start + len - 1);
next:
memset(&map_bh, 0, sizeof(struct buffer_head));
map_bh.b_size = len;
ret = get_data_block(inode, start_blk, &map_bh, 0,
F2FS_GET_BLOCK_FIEMAP, &next_pgofs);
if (ret)
goto out;
/* HOLE */
if (!buffer_mapped(&map_bh)) {
start_blk = next_pgofs;
if (blk_to_logical(inode, start_blk) < blk_to_logical(inode,
F2FS_I_SB(inode)->max_file_blocks))
goto prep_next;
flags |= FIEMAP_EXTENT_LAST;
}
if (size) {
if (f2fs_encrypted_inode(inode))
flags |= FIEMAP_EXTENT_DATA_ENCRYPTED;
ret = fiemap_fill_next_extent(fieinfo, logical,
phys, size, flags);
}
if (start_blk > last_blk || ret)
goto out;
logical = blk_to_logical(inode, start_blk);
phys = blk_to_logical(inode, map_bh.b_blocknr);
size = map_bh.b_size;
flags = 0;
if (buffer_unwritten(&map_bh))
flags = FIEMAP_EXTENT_UNWRITTEN;
start_blk += logical_to_blk(inode, size);
prep_next:
cond_resched();
if (fatal_signal_pending(current))
ret = -EINTR;
else
goto next;
out:
if (ret == 1)
ret = 0;
inode_unlock(inode);
return ret;
}
/*
* This function was originally taken from fs/mpage.c, and customized for f2fs.
* Major change was from block_size == page_size in f2fs by default.
*/
static int f2fs_mpage_readpages(struct address_space *mapping,
struct list_head *pages, struct page *page,
unsigned nr_pages)
{
struct bio *bio = NULL;
sector_t last_block_in_bio = 0;
struct inode *inode = mapping->host;
const unsigned blkbits = inode->i_blkbits;
const unsigned blocksize = 1 << blkbits;
sector_t block_in_file;
sector_t last_block;
sector_t last_block_in_file;
sector_t block_nr;
struct f2fs_map_blocks map;
map.m_pblk = 0;
map.m_lblk = 0;
map.m_len = 0;
map.m_flags = 0;
map.m_next_pgofs = NULL;
map.m_seg_type = NO_CHECK_TYPE;
for (; nr_pages; nr_pages--) {
if (pages) {
page = list_last_entry(pages, struct page, lru);
prefetchw(&page->flags);
list_del(&page->lru);
if (add_to_page_cache_lru(page, mapping,
mm, memcg: use consistent gfp flags during readahead Vladimir has noticed that we might declare memcg oom even during readahead because read_pages only uses GFP_KERNEL (with mapping_gfp restriction) while __do_page_cache_readahead uses page_cache_alloc_readahead which adds __GFP_NORETRY to prevent from OOMs. This gfp mask discrepancy is really unfortunate and easily fixable. Drop page_cache_alloc_readahead() which only has one user and outsource the gfp_mask logic into readahead_gfp_mask and propagate this mask from __do_page_cache_readahead down to read_pages. This alone would have only very limited impact as most filesystems are implementing ->readpages and the common implementation mpage_readpages does GFP_KERNEL (with mapping_gfp restriction) again. We can tell it to use readahead_gfp_mask instead as this function is called only during readahead as well. The same applies to read_cache_pages. ext4 has its own ext4_mpage_readpages but the path which has pages != NULL can use the same gfp mask. Btrfs, cifs, f2fs and orangefs are doing a very similar pattern to mpage_readpages so the same can be applied to them as well. [akpm@linux-foundation.org: coding-style fixes] [mhocko@suse.com: restrict gfp mask in mpage_alloc] Link: http://lkml.kernel.org/r/20160610074223.GC32285@dhcp22.suse.cz Link: http://lkml.kernel.org/r/1465301556-26431-1-git-send-email-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Chris Mason <clm@fb.com> Cc: Steve French <sfrench@samba.org> Cc: Theodore Ts'o <tytso@mit.edu> Cc: Jan Kara <jack@suse.cz> Cc: Mike Marshall <hubcap@omnibond.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Changman Lee <cm224.lee@samsung.com> Cc: Chao Yu <yuchao0@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-26 22:24:53 +00:00
page->index,
readahead_gfp_mask(mapping)))
goto next_page;
}
block_in_file = (sector_t)page->index;
last_block = block_in_file + nr_pages;
last_block_in_file = (i_size_read(inode) + blocksize - 1) >>
blkbits;
if (last_block > last_block_in_file)
last_block = last_block_in_file;
/*
* Map blocks using the previous result first.
*/
if ((map.m_flags & F2FS_MAP_MAPPED) &&
block_in_file > map.m_lblk &&
block_in_file < (map.m_lblk + map.m_len))
goto got_it;
/*
* Then do more f2fs_map_blocks() calls until we are
* done with this page.
*/
map.m_flags = 0;
if (block_in_file < last_block) {
map.m_lblk = block_in_file;
map.m_len = last_block - block_in_file;
if (f2fs_map_blocks(inode, &map, 0,
F2FS_GET_BLOCK_DEFAULT))
goto set_error_page;
}
got_it:
if ((map.m_flags & F2FS_MAP_MAPPED)) {
block_nr = map.m_pblk + block_in_file - map.m_lblk;
SetPageMappedToDisk(page);
if (!PageUptodate(page) && !cleancache_get_page(page)) {
SetPageUptodate(page);
goto confused;
}
} else {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, 0, PAGE_SIZE);
if (!PageUptodate(page))
SetPageUptodate(page);
unlock_page(page);
goto next_page;
}
/*
* This page will go to BIO. Do we need to send this
* BIO off first?
*/
if (bio && (last_block_in_bio != block_nr - 1 ||
!__same_bdev(F2FS_I_SB(inode), block_nr, bio))) {
submit_and_realloc:
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
__submit_bio(F2FS_I_SB(inode), bio, DATA);
bio = NULL;
}
if (bio == NULL) {
bio = f2fs_grab_read_bio(inode, block_nr, nr_pages);
if (IS_ERR(bio)) {
bio = NULL;
goto set_error_page;
}
}
if (bio_add_page(bio, page, blocksize, 0) < blocksize)
goto submit_and_realloc;
last_block_in_bio = block_nr;
goto next_page;
set_error_page:
SetPageError(page);
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, 0, PAGE_SIZE);
unlock_page(page);
goto next_page;
confused:
if (bio) {
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
__submit_bio(F2FS_I_SB(inode), bio, DATA);
bio = NULL;
}
unlock_page(page);
next_page:
if (pages)
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
put_page(page);
}
BUG_ON(pages && !list_empty(pages));
if (bio)
The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. = Enhancement = - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime = Bug fixes = - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert -----BEGIN PGP SIGNATURE----- Version: GnuPG v1 iQIcBAABAgAGBQJXmDJFAAoJEEAUqH6CSFDSJeYP/0ru8+5/ui5VTCdNPQB9KxYD DIUaDGpeoLvmn3ZdrMEdyNr6kWbgjCE9JjOGPQ7l1/apErOGVPyaBwflKcCDwloU pAlEqVM1Q9j4qH4i9SWTlvPtsHBHB7G7YSe3vDB9fJGSTqumubIlnaBm+Wfjx31U p53WcPn9LpOyzfmvZf2tOHmvZ7bWLkE/a07x9kPC6XHUFb9C17jLRFFGeuhZQHv1 Yo7HgokBnPExa8TnEILYyX/x+eecFS/1Cp/cN0STsebSu8pStTHTcAP7qEpKQB88 Cc51Lf+d5gFeydxKDFxwdH3VWOGIr9Ppako+lHW83gJcHP0zw8zdxULab+HJMa4n MOByRRiafwu1sL0dl7TCfsYNIHdEnXhWbhcRhMVZbb5C2Q6+Htuac8ZrKSOWExNN DUqRkzeTib9u+cHxUTFFPgOGdUjDLmg3XHU7mvb+2hViluVjIImC4tqD5XPpv7vt WnaDJxLCGD/6DF2yhiVY9NysuxInLTNFFCF06LworZ4L24hlg5TvN0UeUNRO9954 ux6f+lSORCzV3TmrsHP5vwjSAW26FviPXV1q1HHJeTpWKMlhsZtHmOAJOtZKKmxP WFnHT0aiWF+sQf4qfxVQL+lLqtgRKJAI9zqGRyfDJWJp5aXdRuVsZs9pWNQF7lCo 5gVnCYk3ULjXG3b23j2S =tKTR -----END PGP SIGNATURE----- Merge tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs Pull f2fs updates from Jaegeuk Kim: "The major change in this version is mitigating cpu overheads on write paths by replacing redundant inode page updates with mark_inode_dirty calls. And we tried to reduce lock contentions as well to improve filesystem scalability. Other feature is setting F2FS automatically when detecting host-managed SMR. Enhancements: - ioctl to move a range of data between files - inject orphan inode errors - avoid flush commands congestion - support lazytime Bug fixes: - return proper results for some dentry operations - fix deadlock in add_link failure - disable extent_cache for fcollapse/finsert" * tag 'for-f2fs-4.8' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (68 commits) f2fs: clean up coding style and redundancy f2fs: get victim segment again after new cp f2fs: handle error case with f2fs_bug_on f2fs: avoid data race when deciding checkpoin in f2fs_sync_file f2fs: support an ioctl to move a range of data blocks f2fs: fix to report error number of f2fs_find_entry f2fs: avoid memory allocation failure due to a long length f2fs: reset default idle interval value f2fs: use blk_plug in all the possible paths f2fs: fix to avoid data update racing between GC and DIO f2fs: add maximum prefree segments f2fs: disable extent_cache for fcollapse/finsert inodes f2fs: refactor __exchange_data_block for speed up f2fs: fix ERR_PTR returned by bio f2fs: avoid mark_inode_dirty f2fs: move i_size_write in f2fs_write_end f2fs: fix to avoid redundant discard during fstrim f2fs: avoid mismatching block range for discard f2fs: fix incorrect f_bfree calculation in ->statfs f2fs: use percpu_rw_semaphore ...
2016-07-27 17:36:31 +00:00
__submit_bio(F2FS_I_SB(inode), bio, DATA);
return 0;
}
static int f2fs_read_data_page(struct file *file, struct page *page)
{
struct inode *inode = page->mapping->host;
int ret = -EAGAIN;
trace_f2fs_readpage(page, DATA);
/* If the file has inline data, try to read it directly */
if (f2fs_has_inline_data(inode))
ret = f2fs_read_inline_data(inode, page);
if (ret == -EAGAIN)
ret = f2fs_mpage_readpages(page->mapping, NULL, page, 1);
return ret;
}
static int f2fs_read_data_pages(struct file *file,
struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
struct inode *inode = mapping->host;
struct page *page = list_last_entry(pages, struct page, lru);
trace_f2fs_readpages(inode, page, nr_pages);
/* If the file has inline data, skip readpages */
if (f2fs_has_inline_data(inode))
return 0;
return f2fs_mpage_readpages(mapping, pages, NULL, nr_pages);
}
static int encrypt_one_page(struct f2fs_io_info *fio)
{
struct inode *inode = fio->page->mapping->host;
gfp_t gfp_flags = GFP_NOFS;
if (!f2fs_encrypted_file(inode))
return 0;
/* wait for GCed encrypted page writeback */
f2fs_wait_on_block_writeback(fio->sbi, fio->old_blkaddr);
retry_encrypt:
fio->encrypted_page = fscrypt_encrypt_page(inode, fio->page,
PAGE_SIZE, 0, fio->page->index, gfp_flags);
if (!IS_ERR(fio->encrypted_page))
return 0;
/* flush pending IOs and wait for a while in the ENOMEM case */
if (PTR_ERR(fio->encrypted_page) == -ENOMEM) {
f2fs_flush_merged_writes(fio->sbi);
congestion_wait(BLK_RW_ASYNC, HZ/50);
gfp_flags |= __GFP_NOFAIL;
goto retry_encrypt;
}
return PTR_ERR(fio->encrypted_page);
}
static inline bool need_inplace_update(struct f2fs_io_info *fio)
{
struct inode *inode = fio->page->mapping->host;
if (f2fs_is_pinned_file(inode))
return true;
if (S_ISDIR(inode->i_mode) || f2fs_is_atomic_file(inode))
return false;
if (is_cold_data(fio->page))
return false;
if (IS_ATOMIC_WRITTEN_PAGE(fio->page))
return false;
return need_inplace_update_policy(inode, fio);
}
static inline bool valid_ipu_blkaddr(struct f2fs_io_info *fio)
{
if (fio->old_blkaddr == NEW_ADDR)
return false;
if (fio->old_blkaddr == NULL_ADDR)
return false;
return true;
}
int do_write_data_page(struct f2fs_io_info *fio)
{
struct page *page = fio->page;
struct inode *inode = page->mapping->host;
struct dnode_of_data dn;
struct extent_info ei = {0,0,0};
bool ipu_force = false;
int err = 0;
set_new_dnode(&dn, inode, NULL, NULL, 0);
if (need_inplace_update(fio) &&
f2fs_lookup_extent_cache(inode, page->index, &ei)) {
fio->old_blkaddr = ei.blk + page->index - ei.fofs;
if (valid_ipu_blkaddr(fio)) {
ipu_force = true;
fio->need_lock = LOCK_DONE;
goto got_it;
}
}
/* Deadlock due to between page->lock and f2fs_lock_op */
if (fio->need_lock == LOCK_REQ && !f2fs_trylock_op(fio->sbi))
return -EAGAIN;
err = get_dnode_of_data(&dn, page->index, LOOKUP_NODE);
if (err)
goto out;
fio->old_blkaddr = dn.data_blkaddr;
/* This page is already truncated */
if (fio->old_blkaddr == NULL_ADDR) {
ClearPageUptodate(page);
goto out_writepage;
}
got_it:
/*
* If current allocation needs SSR,
* it had better in-place writes for updated data.
*/
if (ipu_force || (valid_ipu_blkaddr(fio) && need_inplace_update(fio))) {
err = encrypt_one_page(fio);
if (err)
goto out_writepage;
set_page_writeback(page);
f2fs_put_dnode(&dn);
if (fio->need_lock == LOCK_REQ)
f2fs_unlock_op(fio->sbi);
err = rewrite_data_page(fio);
trace_f2fs_do_write_data_page(fio->page, IPU);
set_inode_flag(inode, FI_UPDATE_WRITE);
return err;
}
if (fio->need_lock == LOCK_RETRY) {
if (!f2fs_trylock_op(fio->sbi)) {
err = -EAGAIN;
goto out_writepage;
}
fio->need_lock = LOCK_REQ;
}
err = encrypt_one_page(fio);
if (err)
goto out_writepage;
set_page_writeback(page);
/* LFS mode write path */
write_data_page(&dn, fio);
trace_f2fs_do_write_data_page(page, OPU);
set_inode_flag(inode, FI_APPEND_WRITE);
if (page->index == 0)
set_inode_flag(inode, FI_FIRST_BLOCK_WRITTEN);
out_writepage:
f2fs_put_dnode(&dn);
out:
if (fio->need_lock == LOCK_REQ)
f2fs_unlock_op(fio->sbi);
return err;
}
static int __write_data_page(struct page *page, bool *submitted,
struct writeback_control *wbc,
enum iostat_type io_type)
{
struct inode *inode = page->mapping->host;
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
loff_t i_size = i_size_read(inode);
const pgoff_t end_index = ((unsigned long long) i_size)
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
>> PAGE_SHIFT;
loff_t psize = (page->index + 1) << PAGE_SHIFT;
unsigned offset = 0;
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
bool need_balance_fs = false;
int err = 0;
struct f2fs_io_info fio = {
.sbi = sbi,
.ino = inode->i_ino,
.type = DATA,
.op = REQ_OP_WRITE,
.op_flags = wbc_to_write_flags(wbc),
.old_blkaddr = NULL_ADDR,
.page = page,
.encrypted_page = NULL,
.submitted = false,
.need_lock = LOCK_RETRY,
.io_type = io_type,
.io_wbc = wbc,
};
trace_f2fs_writepage(page, DATA);
if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
goto redirty_out;
if (page->index < end_index)
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
goto write;
/*
* If the offset is out-of-range of file size,
* this page does not have to be written to disk.
*/
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
offset = i_size & (PAGE_SIZE - 1);
if ((page->index >= end_index + 1) || !offset)
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
goto out;
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, offset, PAGE_SIZE);
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
write:
if (f2fs_is_drop_cache(inode))
goto out;
/* we should not write 0'th page having journal header */
if (f2fs_is_volatile_file(inode) && (!page->index ||
(!wbc->for_reclaim &&
available_free_memory(sbi, BASE_CHECK))))
goto redirty_out;
/* we should bypass data pages to proceed the kworkder jobs */
if (unlikely(f2fs_cp_error(sbi))) {
mapping_set_error(page->mapping, -EIO);
goto out;
}
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
/* Dentry blocks are controlled by checkpoint */
if (S_ISDIR(inode->i_mode)) {
fio.need_lock = LOCK_DONE;
err = do_write_data_page(&fio);
goto done;
}
if (!wbc->for_reclaim)
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
need_balance_fs = true;
else if (has_not_enough_free_secs(sbi, 0, 0))
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
goto redirty_out;
else
set_inode_flag(inode, FI_HOT_DATA);
err = -EAGAIN;
if (f2fs_has_inline_data(inode)) {
err = f2fs_write_inline_data(inode, page);
if (!err)
goto out;
}
if (err == -EAGAIN) {
err = do_write_data_page(&fio);
if (err == -EAGAIN) {
fio.need_lock = LOCK_REQ;
err = do_write_data_page(&fio);
}
}
down_write(&F2FS_I(inode)->i_sem);
if (F2FS_I(inode)->last_disk_size < psize)
F2FS_I(inode)->last_disk_size = psize;
up_write(&F2FS_I(inode)->i_sem);
done:
if (err && err != -ENOENT)
goto redirty_out;
f2fs: introduce a new global lock scheme In the previous version, f2fs uses global locks according to the usage types, such as directory operations, block allocation, block write, and so on. Reference the following lock types in f2fs.h. enum lock_type { RENAME, /* for renaming operations */ DENTRY_OPS, /* for directory operations */ DATA_WRITE, /* for data write */ DATA_NEW, /* for data allocation */ DATA_TRUNC, /* for data truncate */ NODE_NEW, /* for node allocation */ NODE_TRUNC, /* for node truncate */ NODE_WRITE, /* for node write */ NR_LOCK_TYPE, }; In that case, we lose the performance under the multi-threading environment, since every types of operations must be conducted one at a time. In order to address the problem, let's share the locks globally with a mutex array regardless of any types. So, let users grab a mutex and perform their jobs in parallel as much as possbile. For this, I propose a new global lock scheme as follows. 0. Data structure - f2fs_sb_info -> mutex_lock[NR_GLOBAL_LOCKS] - f2fs_sb_info -> node_write 1. mutex_lock_op(sbi) - try to get an avaiable lock from the array. - returns the index of the gottern lock variable. 2. mutex_unlock_op(sbi, index of the lock) - unlock the given index of the lock. 3. mutex_lock_all(sbi) - grab all the locks in the array before the checkpoint. 4. mutex_unlock_all(sbi) - release all the locks in the array after checkpoint. 5. block_operations() - call mutex_lock_all() - sync_dirty_dir_inodes() - grab node_write - sync_node_pages() Note that, the pairs of mutex_lock_op()/mutex_unlock_op() and mutex_lock_all()/mutex_unlock_all() should be used together. Signed-off-by: Jaegeuk Kim <jaegeuk.kim@samsung.com>
2012-11-22 07:21:29 +00:00
out:
inode_dec_dirty_pages(inode);
if (err)
ClearPageUptodate(page);
if (wbc->for_reclaim) {
f2fs_submit_merged_write_cond(sbi, inode, 0, page->index, DATA);
clear_inode_flag(inode, FI_HOT_DATA);
remove_dirty_inode(inode);
submitted = NULL;
}
unlock_page(page);
if (!S_ISDIR(inode->i_mode))
f2fs_balance_fs(sbi, need_balance_fs);
if (unlikely(f2fs_cp_error(sbi))) {
f2fs_submit_merged_write(sbi, DATA);
submitted = NULL;
}
if (submitted)
*submitted = fio.submitted;
return 0;
redirty_out:
redirty_page_for_writepage(wbc, page);
if (!err)
return AOP_WRITEPAGE_ACTIVATE;
unlock_page(page);
return err;
}
static int f2fs_write_data_page(struct page *page,
struct writeback_control *wbc)
{
return __write_data_page(page, NULL, wbc, FS_DATA_IO);
}
/*
* This function was copied from write_cche_pages from mm/page-writeback.c.
* The major change is making write step of cold data page separately from
* warm/hot data page.
*/
static int f2fs_write_cache_pages(struct address_space *mapping,
struct writeback_control *wbc,
enum iostat_type io_type)
{
int ret = 0;
int done = 0;
struct pagevec pvec;
int nr_pages;
pgoff_t uninitialized_var(writeback_index);
pgoff_t index;
pgoff_t end; /* Inclusive */
pgoff_t done_index;
pgoff_t last_idx = ULONG_MAX;
int cycled;
int range_whole = 0;
int tag;
pagevec_init(&pvec);
if (get_dirty_pages(mapping->host) <=
SM_I(F2FS_M_SB(mapping))->min_hot_blocks)
set_inode_flag(mapping->host, FI_HOT_DATA);
else
clear_inode_flag(mapping->host, FI_HOT_DATA);
if (wbc->range_cyclic) {
writeback_index = mapping->writeback_index; /* prev offset */
index = writeback_index;
if (index == 0)
cycled = 1;
else
cycled = 0;
end = -1;
} else {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
index = wbc->range_start >> PAGE_SHIFT;
end = wbc->range_end >> PAGE_SHIFT;
if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX)
range_whole = 1;
cycled = 1; /* ignore range_cyclic tests */
}
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
tag = PAGECACHE_TAG_TOWRITE;
else
tag = PAGECACHE_TAG_DIRTY;
retry:
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
tag_pages_for_writeback(mapping, index, end);
done_index = index;
while (!done && (index <= end)) {
int i;
nr_pages = pagevec_lookup_range_tag(&pvec, mapping, &index, end,
tag);
if (nr_pages == 0)
break;
for (i = 0; i < nr_pages; i++) {
struct page *page = pvec.pages[i];
bool submitted = false;
done_index = page->index;
retry_write:
lock_page(page);
if (unlikely(page->mapping != mapping)) {
continue_unlock:
unlock_page(page);
continue;
}
if (!PageDirty(page)) {
/* someone wrote it for us */
goto continue_unlock;
}
if (PageWriteback(page)) {
if (wbc->sync_mode != WB_SYNC_NONE)
f2fs_wait_on_page_writeback(page,
DATA, true);
else
goto continue_unlock;
}
BUG_ON(PageWriteback(page));
if (!clear_page_dirty_for_io(page))
goto continue_unlock;
ret = __write_data_page(page, &submitted, wbc, io_type);
if (unlikely(ret)) {
/*
* keep nr_to_write, since vfs uses this to
* get # of written pages.
*/
if (ret == AOP_WRITEPAGE_ACTIVATE) {
unlock_page(page);
ret = 0;
continue;
} else if (ret == -EAGAIN) {
ret = 0;
if (wbc->sync_mode == WB_SYNC_ALL) {
cond_resched();
congestion_wait(BLK_RW_ASYNC,
HZ/50);
goto retry_write;
}
continue;
}
done_index = page->index + 1;
done = 1;
break;
} else if (submitted) {
last_idx = page->index;
}
/* give a priority to WB_SYNC threads */
if ((atomic_read(&F2FS_M_SB(mapping)->wb_sync_req) ||
--wbc->nr_to_write <= 0) &&
wbc->sync_mode == WB_SYNC_NONE) {
done = 1;
break;
}
}
pagevec_release(&pvec);
cond_resched();
}
if (!cycled && !done) {
cycled = 1;
index = 0;
end = writeback_index - 1;
goto retry;
}
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
mapping->writeback_index = done_index;
if (last_idx != ULONG_MAX)
f2fs_submit_merged_write_cond(F2FS_M_SB(mapping), mapping->host,
0, last_idx, DATA);
return ret;
}
int __f2fs_write_data_pages(struct address_space *mapping,
struct writeback_control *wbc,
enum iostat_type io_type)
{
struct inode *inode = mapping->host;
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
struct blk_plug plug;
int ret;
/* deal with chardevs and other special file */
if (!mapping->a_ops->writepage)
return 0;
/* skip writing if there is no dirty page in this inode */
if (!get_dirty_pages(inode) && wbc->sync_mode == WB_SYNC_NONE)
return 0;
/* during POR, we don't need to trigger writepage at all. */
if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING)))
goto skip_write;
if (S_ISDIR(inode->i_mode) && wbc->sync_mode == WB_SYNC_NONE &&
get_dirty_pages(inode) < nr_pages_to_skip(sbi, DATA) &&
available_free_memory(sbi, DIRTY_DENTS))
goto skip_write;
/* skip writing during file defragment */
if (is_inode_flag_set(inode, FI_DO_DEFRAG))
goto skip_write;
trace_f2fs_writepages(mapping->host, wbc, DATA);
/* to avoid spliting IOs due to mixed WB_SYNC_ALL and WB_SYNC_NONE */
if (wbc->sync_mode == WB_SYNC_ALL)
atomic_inc(&sbi->wb_sync_req);
else if (atomic_read(&sbi->wb_sync_req))
goto skip_write;
blk_start_plug(&plug);
ret = f2fs_write_cache_pages(mapping, wbc, io_type);
blk_finish_plug(&plug);
if (wbc->sync_mode == WB_SYNC_ALL)
atomic_dec(&sbi->wb_sync_req);
/*
* if some pages were truncated, we cannot guarantee its mapping->host
* to detect pending bios.
*/
remove_dirty_inode(inode);
return ret;
skip_write:
wbc->pages_skipped += get_dirty_pages(inode);
trace_f2fs_writepages(mapping->host, wbc, DATA);
return 0;
}
static int f2fs_write_data_pages(struct address_space *mapping,
struct writeback_control *wbc)
{
struct inode *inode = mapping->host;
return __f2fs_write_data_pages(mapping, wbc,
F2FS_I(inode)->cp_task == current ?
FS_CP_DATA_IO : FS_DATA_IO);
}
static void f2fs_write_failed(struct address_space *mapping, loff_t to)
{
struct inode *inode = mapping->host;
loff_t i_size = i_size_read(inode);
if (to > i_size) {
down_write(&F2FS_I(inode)->i_mmap_sem);
truncate_pagecache(inode, i_size);
truncate_blocks(inode, i_size, true);
up_write(&F2FS_I(inode)->i_mmap_sem);
}
}
static int prepare_write_begin(struct f2fs_sb_info *sbi,
struct page *page, loff_t pos, unsigned len,
block_t *blk_addr, bool *node_changed)
{
struct inode *inode = page->mapping->host;
pgoff_t index = page->index;
struct dnode_of_data dn;
struct page *ipage;
bool locked = false;
struct extent_info ei = {0,0,0};
int err = 0;
/*
* we already allocated all the blocks, so we don't need to get
* the block addresses when there is no need to fill the page.
*/
if (!f2fs_has_inline_data(inode) && len == PAGE_SIZE &&
!is_inode_flag_set(inode, FI_NO_PREALLOC))
return 0;
if (f2fs_has_inline_data(inode) ||
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
(pos & PAGE_MASK) >= i_size_read(inode)) {
__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, true);
locked = true;
}
restart:
/* check inline_data */
ipage = get_node_page(sbi, inode->i_ino);
if (IS_ERR(ipage)) {
err = PTR_ERR(ipage);
goto unlock_out;
}
set_new_dnode(&dn, inode, ipage, ipage, 0);
if (f2fs_has_inline_data(inode)) {
if (pos + len <= MAX_INLINE_DATA(inode)) {
read_inline_data(page, ipage);
set_inode_flag(inode, FI_DATA_EXIST);
f2fs: fix deadlock when flush inline data Below backtrace info was reported by Yunlei He: Call Trace: [<ffffffff817a9395>] schedule+0x35/0x80 [<ffffffff817abb7d>] rwsem_down_read_failed+0xed/0x130 [<ffffffff813c12a8>] call_rwsem_down_read_failed+0x18/0x [<ffffffff817ab1d0>] down_read+0x20/0x30 [<ffffffffa02a1a12>] f2fs_evict_inode+0x242/0x3a0 [f2fs] [<ffffffff81217057>] evict+0xc7/0x1a0 [<ffffffff81217cd6>] iput+0x196/0x200 [<ffffffff812134f9>] __dentry_kill+0x179/0x1e0 [<ffffffff812136f9>] dput+0x199/0x1f0 [<ffffffff811fe77b>] __fput+0x18b/0x220 [<ffffffff811fe84e>] ____fput+0xe/0x10 [<ffffffff81097427>] task_work_run+0x77/0x90 [<ffffffff81074d62>] exit_to_usermode_loop+0x73/0xa2 [<ffffffff81003b7a>] do_syscall_64+0xfa/0x110 [<ffffffff817acf65>] entry_SYSCALL64_slow_path+0x25/0x25 Call Trace: [<ffffffff817a9395>] schedule+0x35/0x80 [<ffffffff81216dc3>] __wait_on_freeing_inode+0xa3/0xd0 [<ffffffff810bc300>] ? autoremove_wake_function+0x40/0x4 [<ffffffff8121771d>] find_inode_fast+0x7d/0xb0 [<ffffffff8121794a>] ilookup+0x6a/0xd0 [<ffffffffa02bc740>] sync_node_pages+0x210/0x650 [f2fs] [<ffffffff8122e690>] ? do_fsync+0x70/0x70 [<ffffffffa02b085e>] block_operations+0x9e/0xf0 [f2fs] [<ffffffff8137b795>] ? bio_endio+0x55/0x60 [<ffffffffa02b0942>] write_checkpoint+0x92/0xba0 [f2fs] [<ffffffff8117da57>] ? mempool_free_slab+0x17/0x20 [<ffffffff8117de8b>] ? mempool_free+0x2b/0x80 [<ffffffff8122e690>] ? do_fsync+0x70/0x70 [<ffffffffa02a53e3>] f2fs_sync_fs+0x63/0xd0 [f2fs] [<ffffffff8129630f>] ? ext4_sync_fs+0xbf/0x190 [<ffffffff8122e6b0>] sync_fs_one_sb+0x20/0x30 [<ffffffff812002e9>] iterate_supers+0xb9/0x110 [<ffffffff8122e7b5>] sys_sync+0x55/0x90 [<ffffffff81003ae9>] do_syscall_64+0x69/0x110 [<ffffffff817acf65>] entry_SYSCALL64_slow_path+0x25/0x25 With following excuting serials, we will set inline_node in inode page after inode was unlinked, result in a deadloop described as below: 1. open file 2. write file 3. unlink file 4. write file 5. close file Thread A Thread B - dput - iput_final - inode->i_state |= I_FREEING - evict - f2fs_evict_inode - f2fs_sync_fs - write_checkpoint - block_operations - f2fs_lock_all (down_write(cp_rwsem)) - f2fs_lock_op (down_read(cp_rwsem)) - sync_node_pages - ilookup - find_inode_fast - __wait_on_freeing_inode (wait on I_FREEING clear) Here, we change to set inline_node flag only for linked inode for fixing. Reported-by: Yunlei He <heyunlei@huawei.com> Signed-off-by: Chao Yu <yuchao0@huawei.com> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Cc: stable@vger.kernel.org # v4.6 Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2016-05-11 11:48:44 +00:00
if (inode->i_nlink)
set_inline_node(ipage);
} else {
err = f2fs_convert_inline_page(&dn, page);
if (err)
goto out;
if (dn.data_blkaddr == NULL_ADDR)
err = f2fs_get_block(&dn, index);
}
} else if (locked) {
err = f2fs_get_block(&dn, index);
} else {
if (f2fs_lookup_extent_cache(inode, index, &ei)) {
dn.data_blkaddr = ei.blk + index - ei.fofs;
} else {
/* hole case */
err = get_dnode_of_data(&dn, index, LOOKUP_NODE);
if (err || dn.data_blkaddr == NULL_ADDR) {
f2fs_put_dnode(&dn);
__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO,
true);
locked = true;
goto restart;
}
}
}
/* convert_inline_page can make node_changed */
*blk_addr = dn.data_blkaddr;
*node_changed = dn.node_changed;
out:
f2fs_put_dnode(&dn);
unlock_out:
if (locked)
__do_map_lock(sbi, F2FS_GET_BLOCK_PRE_AIO, false);
return err;
}
static int f2fs_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;
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
struct page *page = NULL;
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
pgoff_t index = ((unsigned long long) pos) >> PAGE_SHIFT;
bool need_balance = false;
block_t blkaddr = NULL_ADDR;
int err = 0;
trace_f2fs_write_begin(inode, pos, len, flags);
if (f2fs_is_atomic_file(inode) &&
!available_free_memory(sbi, INMEM_PAGES)) {
err = -ENOMEM;
goto fail;
}
/*
* We should check this at this moment to avoid deadlock on inode page
* and #0 page. The locking rule for inline_data conversion should be:
* lock_page(page #0) -> lock_page(inode_page)
*/
if (index != 0) {
err = f2fs_convert_inline_inode(inode);
if (err)
goto fail;
}
repeat:
/*
* Do not use grab_cache_page_write_begin() to avoid deadlock due to
* wait_for_stable_page. Will wait that below with our IO control.
*/
page = f2fs_pagecache_get_page(mapping, index,
FGP_LOCK | FGP_WRITE | FGP_CREAT, GFP_NOFS);
if (!page) {
err = -ENOMEM;
goto fail;
}
*pagep = page;
err = prepare_write_begin(sbi, page, pos, len,
&blkaddr, &need_balance);
if (err)
goto fail;
if (need_balance && has_not_enough_free_secs(sbi, 0, 0)) {
unlock_page(page);
f2fs_balance_fs(sbi, true);
lock_page(page);
if (page->mapping != mapping) {
/* The page got truncated from under us */
f2fs_put_page(page, 1);
goto repeat;
}
}
f2fs_wait_on_page_writeback(page, DATA, false);
f2fs crypto: fix racing of accessing encrypted page among different competitors Since we use different page cache (normally inode's page cache for R/W and meta inode's page cache for GC) to cache the same physical block which is belong to an encrypted inode. Writeback of these two page cache should be exclusive, but now we didn't handle writeback state well, so there may be potential racing problem: a) kworker: f2fs_gc: - f2fs_write_data_pages - f2fs_write_data_page - do_write_data_page - write_data_page - f2fs_submit_page_mbio (page#1 in inode's page cache was queued in f2fs bio cache, and be ready to write to new blkaddr) - gc_data_segment - move_encrypted_block - pagecache_get_page (page#2 in meta inode's page cache was cached with the invalid datas of physical block located in new blkaddr) - f2fs_submit_page_mbio (page#1 was submitted, later, page#2 with invalid data will be submitted) b) f2fs_gc: - gc_data_segment - move_encrypted_block - f2fs_submit_page_mbio (page#1 in meta inode's page cache was queued in f2fs bio cache, and be ready to write to new blkaddr) user thread: - f2fs_write_begin - f2fs_submit_page_bio (we submit the request to block layer to update page#2 in inode's page cache with physical block located in new blkaddr, so here we may read gabbage data from new blkaddr since GC hasn't writebacked the page#1 yet) This patch fixes above potential racing problem for encrypted inode. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2015-10-08 05:27:34 +00:00
/* wait for GCed encrypted page writeback */
if (f2fs_encrypted_file(inode))
f2fs_wait_on_block_writeback(sbi, blkaddr);
f2fs crypto: fix racing of accessing encrypted page among different competitors Since we use different page cache (normally inode's page cache for R/W and meta inode's page cache for GC) to cache the same physical block which is belong to an encrypted inode. Writeback of these two page cache should be exclusive, but now we didn't handle writeback state well, so there may be potential racing problem: a) kworker: f2fs_gc: - f2fs_write_data_pages - f2fs_write_data_page - do_write_data_page - write_data_page - f2fs_submit_page_mbio (page#1 in inode's page cache was queued in f2fs bio cache, and be ready to write to new blkaddr) - gc_data_segment - move_encrypted_block - pagecache_get_page (page#2 in meta inode's page cache was cached with the invalid datas of physical block located in new blkaddr) - f2fs_submit_page_mbio (page#1 was submitted, later, page#2 with invalid data will be submitted) b) f2fs_gc: - gc_data_segment - move_encrypted_block - f2fs_submit_page_mbio (page#1 in meta inode's page cache was queued in f2fs bio cache, and be ready to write to new blkaddr) user thread: - f2fs_write_begin - f2fs_submit_page_bio (we submit the request to block layer to update page#2 in inode's page cache with physical block located in new blkaddr, so here we may read gabbage data from new blkaddr since GC hasn't writebacked the page#1 yet) This patch fixes above potential racing problem for encrypted inode. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2015-10-08 05:27:34 +00:00
if (len == PAGE_SIZE || PageUptodate(page))
return 0;
if (!(pos & (PAGE_SIZE - 1)) && (pos + len) >= i_size_read(inode)) {
zero_user_segment(page, len, PAGE_SIZE);
return 0;
}
if (blkaddr == NEW_ADDR) {
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
zero_user_segment(page, 0, PAGE_SIZE);
SetPageUptodate(page);
} else {
err = f2fs_submit_page_read(inode, page, blkaddr);
if (err)
goto fail;
lock_page(page);
if (unlikely(page->mapping != mapping)) {
f2fs_put_page(page, 1);
goto repeat;
}
if (unlikely(!PageUptodate(page))) {
err = -EIO;
goto fail;
}
}
return 0;
fail:
f2fs_put_page(page, 1);
f2fs_write_failed(mapping, pos + len);
if (f2fs_is_atomic_file(inode))
drop_inmem_pages_all(sbi);
return err;
}
static int f2fs_write_end(struct file *file,
struct address_space *mapping,
loff_t pos, unsigned len, unsigned copied,
struct page *page, void *fsdata)
{
struct inode *inode = page->mapping->host;
trace_f2fs_write_end(inode, pos, len, copied);
/*
* This should be come from len == PAGE_SIZE, and we expect copied
* should be PAGE_SIZE. Otherwise, we treat it with zero copied and
* let generic_perform_write() try to copy data again through copied=0.
*/
if (!PageUptodate(page)) {
if (unlikely(copied != len))
copied = 0;
else
SetPageUptodate(page);
}
if (!copied)
goto unlock_out;
set_page_dirty(page);
if (pos + copied > i_size_read(inode))
f2fs_i_size_write(inode, pos + copied);
unlock_out:
f2fs_put_page(page, 1);
f2fs_update_time(F2FS_I_SB(inode), REQ_TIME);
return copied;
}
static int check_direct_IO(struct inode *inode, struct iov_iter *iter,
loff_t offset)
{
unsigned blocksize_mask = inode->i_sb->s_blocksize - 1;
if (offset & blocksize_mask)
return -EINVAL;
if (iov_iter_alignment(iter) & blocksize_mask)
return -EINVAL;
return 0;
}
static ssize_t f2fs_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
{
struct address_space *mapping = iocb->ki_filp->f_mapping;
struct inode *inode = mapping->host;
size_t count = iov_iter_count(iter);
loff_t offset = iocb->ki_pos;
int rw = iov_iter_rw(iter);
int err;
err = check_direct_IO(inode, iter, offset);
if (err)
return err;
if (__force_buffered_io(inode, rw))
return 0;
trace_f2fs_direct_IO_enter(inode, offset, count, rw);
down_read(&F2FS_I(inode)->dio_rwsem[rw]);
err = blockdev_direct_IO(iocb, inode, iter, get_data_block_dio);
up_read(&F2FS_I(inode)->dio_rwsem[rw]);
if (rw == WRITE) {
if (err > 0) {
f2fs_update_iostat(F2FS_I_SB(inode), APP_DIRECT_IO,
err);
set_inode_flag(inode, FI_UPDATE_WRITE);
} else if (err < 0) {
f2fs_write_failed(mapping, offset + count);
}
}
trace_f2fs_direct_IO_exit(inode, offset, count, rw, err);
return err;
}
void f2fs_invalidate_page(struct page *page, unsigned int offset,
unsigned int length)
{
struct inode *inode = page->mapping->host;
struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
if (inode->i_ino >= F2FS_ROOT_INO(sbi) &&
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time ago with promise that one day it will be possible to implement page cache with bigger chunks than PAGE_SIZE. This promise never materialized. And unlikely will. We have many places where PAGE_CACHE_SIZE assumed to be equal to PAGE_SIZE. And it's constant source of confusion on whether PAGE_CACHE_* or PAGE_* constant should be used in a particular case, especially on the border between fs and mm. Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much breakage to be doable. Let's stop pretending that pages in page cache are special. They are not. The changes are pretty straight-forward: - <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>; - PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN}; - page_cache_get() -> get_page(); - page_cache_release() -> put_page(); This patch contains automated changes generated with coccinelle using script below. For some reason, coccinelle doesn't patch header files. I've called spatch for them manually. The only adjustment after coccinelle is revert of changes to PAGE_CAHCE_ALIGN definition: we are going to drop it later. There are few places in the code where coccinelle didn't reach. I'll fix them manually in a separate patch. Comments and documentation also will be addressed with the separate patch. virtual patch @@ expression E; @@ - E << (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ expression E; @@ - E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) + E @@ @@ - PAGE_CACHE_SHIFT + PAGE_SHIFT @@ @@ - PAGE_CACHE_SIZE + PAGE_SIZE @@ @@ - PAGE_CACHE_MASK + PAGE_MASK @@ expression E; @@ - PAGE_CACHE_ALIGN(E) + PAGE_ALIGN(E) @@ expression E; @@ - page_cache_get(E) + get_page(E) @@ expression E; @@ - page_cache_release(E) + put_page(E) Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-04-01 12:29:47 +00:00
(offset % PAGE_SIZE || length != PAGE_SIZE))
return;
if (PageDirty(page)) {
if (inode->i_ino == F2FS_META_INO(sbi)) {
dec_page_count(sbi, F2FS_DIRTY_META);
} else if (inode->i_ino == F2FS_NODE_INO(sbi)) {
dec_page_count(sbi, F2FS_DIRTY_NODES);
} else {
inode_dec_dirty_pages(inode);
remove_dirty_inode(inode);
}
}
/* This is atomic written page, keep Private */
if (IS_ATOMIC_WRITTEN_PAGE(page))
f2fs: fix stale ATOMIC_WRITTEN_PAGE private pointer When I forced to enable atomic operations intentionally, I could hit the below panic, since we didn't clear page->private in f2fs_invalidate_page called by file truncation. The panic occurs due to NULL mapping having page->private. BUG: unable to handle kernel paging request at ffffffffffffffff IP: drop_buffers+0x38/0xe0 PGD 5d00c067 PUD 5d00e067 PMD 0 CPU: 3 PID: 1648 Comm: fsstress Tainted: G D OE 4.10.0+ #5 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 task: ffff9151952863c0 task.stack: ffffaaec40db4000 RIP: 0010:drop_buffers+0x38/0xe0 RSP: 0018:ffffaaec40db74c8 EFLAGS: 00010292 Call Trace: ? page_referenced+0x8b/0x170 try_to_free_buffers+0xc5/0xe0 try_to_release_page+0x49/0x50 shrink_page_list+0x8bc/0x9f0 shrink_inactive_list+0x1dd/0x500 ? shrink_active_list+0x2c0/0x430 shrink_node_memcg+0x5eb/0x7c0 shrink_node+0xe1/0x320 do_try_to_free_pages+0xef/0x2e0 try_to_free_pages+0xe9/0x190 __alloc_pages_slowpath+0x390/0xe70 __alloc_pages_nodemask+0x291/0x2b0 alloc_pages_current+0x95/0x140 __page_cache_alloc+0xc4/0xe0 pagecache_get_page+0xab/0x2a0 grab_cache_page_write_begin+0x20/0x40 get_read_data_page+0x2e6/0x4c0 [f2fs] ? f2fs_mark_inode_dirty_sync+0x16/0x30 [f2fs] ? truncate_data_blocks_range+0x238/0x2b0 [f2fs] get_lock_data_page+0x30/0x190 [f2fs] __exchange_data_block+0xaaf/0xf40 [f2fs] f2fs_fallocate+0x418/0xd00 [f2fs] vfs_fallocate+0x157/0x220 SyS_fallocate+0x48/0x80 Signed-off-by: Yunlei He <heyunlei@huawei.com> Signed-off-by: Chao Yu <yuchao0@huawei.com> [Chao Yu: use INMEM_INVALIDATE for better tracing] Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
2017-03-17 01:55:52 +00:00
return drop_inmem_page(inode, page);
set_page_private(page, 0);
ClearPagePrivate(page);
}
int f2fs_release_page(struct page *page, gfp_t wait)
{
/* If this is dirty page, keep PagePrivate */
if (PageDirty(page))
return 0;
/* This is atomic written page, keep Private */
if (IS_ATOMIC_WRITTEN_PAGE(page))
return 0;
set_page_private(page, 0);
ClearPagePrivate(page);
return 1;
}
/*
* This was copied from __set_page_dirty_buffers which gives higher performance
* in very high speed storages. (e.g., pmem)
*/
void f2fs_set_page_dirty_nobuffers(struct page *page)
{
struct address_space *mapping = page->mapping;
unsigned long flags;
if (unlikely(!mapping))
return;
spin_lock(&mapping->private_lock);
lock_page_memcg(page);
SetPageDirty(page);
spin_unlock(&mapping->private_lock);
spin_lock_irqsave(&mapping->tree_lock, flags);
WARN_ON_ONCE(!PageUptodate(page));
account_page_dirtied(page, mapping);
radix_tree_tag_set(&mapping->page_tree,
page_index(page), PAGECACHE_TAG_DIRTY);
spin_unlock_irqrestore(&mapping->tree_lock, flags);
unlock_page_memcg(page);
__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
return;
}
static int f2fs_set_data_page_dirty(struct page *page)
{
struct address_space *mapping = page->mapping;
struct inode *inode = mapping->host;
trace_f2fs_set_page_dirty(page, DATA);
if (!PageUptodate(page))
SetPageUptodate(page);
if (f2fs_is_atomic_file(inode) && !f2fs_is_commit_atomic_write(inode)) {
if (!IS_ATOMIC_WRITTEN_PAGE(page)) {
register_inmem_page(inode, page);
return 1;
}
/*
* Previously, this page has been registered, we just
* return here.
*/
return 0;
}
if (!PageDirty(page)) {
f2fs_set_page_dirty_nobuffers(page);
update_dirty_page(inode, page);
return 1;
}
return 0;
}
static sector_t f2fs_bmap(struct address_space *mapping, sector_t block)
{
struct inode *inode = mapping->host;
if (f2fs_has_inline_data(inode))
return 0;
/* make sure allocating whole blocks */
if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
filemap_write_and_wait(mapping);
return generic_block_bmap(mapping, block, get_data_block_bmap);
}
#ifdef CONFIG_MIGRATION
#include <linux/migrate.h>
int f2fs_migrate_page(struct address_space *mapping,
struct page *newpage, struct page *page, enum migrate_mode mode)
{
int rc, extra_count;
struct f2fs_inode_info *fi = F2FS_I(mapping->host);
bool atomic_written = IS_ATOMIC_WRITTEN_PAGE(page);
BUG_ON(PageWriteback(page));
/* migrating an atomic written page is safe with the inmem_lock hold */
if (atomic_written) {
if (mode != MIGRATE_SYNC)
return -EBUSY;
if (!mutex_trylock(&fi->inmem_lock))
return -EAGAIN;
}
/*
* A reference is expected if PagePrivate set when move mapping,
* however F2FS breaks this for maintaining dirty page counts when
* truncating pages. So here adjusting the 'extra_count' make it work.
*/
extra_count = (atomic_written ? 1 : 0) - page_has_private(page);
rc = migrate_page_move_mapping(mapping, newpage,
page, NULL, mode, extra_count);
if (rc != MIGRATEPAGE_SUCCESS) {
if (atomic_written)
mutex_unlock(&fi->inmem_lock);
return rc;
}
if (atomic_written) {
struct inmem_pages *cur;
list_for_each_entry(cur, &fi->inmem_pages, list)
if (cur->page == page) {
cur->page = newpage;
break;
}
mutex_unlock(&fi->inmem_lock);
put_page(page);
get_page(newpage);
}
if (PagePrivate(page))
SetPagePrivate(newpage);
set_page_private(newpage, page_private(page));
mm/migrate: new migrate mode MIGRATE_SYNC_NO_COPY Introduce a new migration mode that allow to offload the copy to a device DMA engine. This changes the workflow of migration and not all address_space migratepage callback can support this. This is intended to be use by migrate_vma() which itself is use for thing like HMM (see include/linux/hmm.h). No additional per-filesystem migratepage testing is needed. I disables MIGRATE_SYNC_NO_COPY in all problematic migratepage() callback and i added comment in those to explain why (part of this patch). The commit message is unclear it should say that any callback that wish to support this new mode need to be aware of the difference in the migration flow from other mode. Some of these callbacks do extra locking while copying (aio, zsmalloc, balloon, ...) and for DMA to be effective you want to copy multiple pages in one DMA operations. But in the problematic case you can not easily hold the extra lock accross multiple call to this callback. Usual flow is: For each page { 1 - lock page 2 - call migratepage() callback 3 - (extra locking in some migratepage() callback) 4 - migrate page state (freeze refcount, update page cache, buffer head, ...) 5 - copy page 6 - (unlock any extra lock of migratepage() callback) 7 - return from migratepage() callback 8 - unlock page } The new mode MIGRATE_SYNC_NO_COPY: 1 - lock multiple pages For each page { 2 - call migratepage() callback 3 - abort in all problematic migratepage() callback 4 - migrate page state (freeze refcount, update page cache, buffer head, ...) } // finished all calls to migratepage() callback 5 - DMA copy multiple pages 6 - unlock all the pages To support MIGRATE_SYNC_NO_COPY in the problematic case we would need a new callback migratepages() (for instance) that deals with multiple pages in one transaction. Because the problematic cases are not important for current usage I did not wanted to complexify this patchset even more for no good reason. Link: http://lkml.kernel.org/r/20170817000548.32038-14-jglisse@redhat.com Signed-off-by: Jérôme Glisse <jglisse@redhat.com> Cc: Aneesh Kumar <aneesh.kumar@linux.vnet.ibm.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Dan Williams <dan.j.williams@intel.com> Cc: David Nellans <dnellans@nvidia.com> Cc: Evgeny Baskakov <ebaskakov@nvidia.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: John Hubbard <jhubbard@nvidia.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Mark Hairgrove <mhairgrove@nvidia.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Cc: Ross Zwisler <ross.zwisler@linux.intel.com> Cc: Sherry Cheung <SCheung@nvidia.com> Cc: Subhash Gutti <sgutti@nvidia.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Bob Liu <liubo95@huawei.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-09-08 23:12:06 +00:00
if (mode != MIGRATE_SYNC_NO_COPY)
migrate_page_copy(newpage, page);
else
migrate_page_states(newpage, page);
return MIGRATEPAGE_SUCCESS;
}
#endif
const struct address_space_operations f2fs_dblock_aops = {
.readpage = f2fs_read_data_page,
.readpages = f2fs_read_data_pages,
.writepage = f2fs_write_data_page,
.writepages = f2fs_write_data_pages,
.write_begin = f2fs_write_begin,
.write_end = f2fs_write_end,
.set_page_dirty = f2fs_set_data_page_dirty,
.invalidatepage = f2fs_invalidate_page,
.releasepage = f2fs_release_page,
.direct_IO = f2fs_direct_IO,
.bmap = f2fs_bmap,
#ifdef CONFIG_MIGRATION
.migratepage = f2fs_migrate_page,
#endif
};