- 04 Jun, 2018 3 commits
-
-
Chao Yu authored
Although mixed sync/async IOs can have continuous LBA, as they have different IO priority, block IO scheduler will add them into different queues and commit them separately, result in splited IOs which causes wrose performance. This patch gives high priority to synchronous IO of nodes, means that once synchronous flow starts, it can interrupt asynchronous writeback flow of system flusher, so more big IOs can be expected. Signed-off-by:
Chao Yu <yuchao0@huawei.com> Signed-off-by:
Jaegeuk Kim <jaegeuk@kernel.org>
-
Chao Yu authored
We should never falsify wbc->sync_mode passed from mm, otherwise mm can trigger writeback with wrong IO priority. Signed-off-by:
-
Chao Yu authored
If we change system time to the past, get_mtime() will return a overflowed time, and SIT_I(sbi)->max_mtime will be udpated incorrectly, this patch fixes the two issues. Signed-off-by:
-
- 31 May, 2018 37 commits
-
-
youngjun yoo authored
clean up checkpatch error: ERROR: space required after that ':' ERROR: space required after that ',' Signed-off-by:
youngjun yoo <youngjun.willow@gmail.com> Reviewed-by:
-
youngjun yoo authored
clean up checkpatch warning: WARNING: Missing a blank line after declarations Signed-off-by:
-
youngjun yoo authored
clean up checkpatch warning: WARNING: Prefer 'unsigned int' to bare use of 'unsigned' Signed-off-by:
-
Chao Yu authored
As Ted reported: "Hi, I was looking at f2fs's sources recently, and I noticed that there is a very large number of non-static symbols which don't have a f2fs prefix. There's well over a hundred (see attached below). As one example, in fs/f2fs/dir.c there is: unsigned char get_de_type(struct f2fs_dir_entry *de) This function is clearly only useful for f2fs, but it has a generic name. This means that if any other file system tries to have the same symbol name, there will be a symbol conflict and the kernel would not successfully build. It also means that when someone is looking f2fs sources, it's not at all obvious whether a function such as read_data_page(), invalidate_blocks(), is a generic kernel function found in the fs, mm, or block layers, or a f2fs specific function. You might want to fix this at some point. Hopefully Kent's bcachefs isn't similarly using genericly named functions, since that might cause conflicts with f2fs's functions --- but just as this would be a problem that we would rightly insist that Kent fix, this is something that we should have rightly insisted that f2fs should have fixed before it was integrated into the mainline kernel. acquire_orphan_inode add_ino_entry add_orphan_inode allocate_data_block allocate_new_segments alloc_nid alloc_nid_done alloc_nid_failed available_free_memory ...." This patch adds "f2fs_" prefix for all non-static symbols in order to: a) avoid conflict with other kernel generic symbols; b) to indicate the function is f2fs specific one instead of generic one; Reported-by:
Theodore Ts'o <tytso@mit.edu> Signed-off-by:
-
Chao Yu authored
Signed-off-by:
-
Chao Yu authored
Signed-off-by:
-
Chao Yu authored
Configure io_bits with 2 and enable LFS mode, generic/017 reports below dmesg: BUG: unable to handle kernel NULL pointer dereference at 00000039 *pdpt = 000000002fcb2001 *pde = 0000000000000000 Oops: 0000 [#1] PREEMPT SMP Modules linked in: crc32_generic zram f2fs(O) bnep rfcomm bluetooth ecdh_generic snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_seq_midi snd_seq_midi_event snd_rawmidi pcbc snd_seq joydev aesni_intel aes_i586 snd_seq_device snd_timer crypto_simd cryptd snd soundcore i2c_piix4 serio_raw mac_hid video parport_pc ppdev lp parport hid_generic usbhid psmouse hid e1000 CPU: 2 PID: 20779 Comm: xfs_io Tainted: G O 4.17.0-rc2 #38 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 EIP: is_checkpointed_data+0x84/0xd0 [f2fs] EFLAGS: 00010207 CPU: 2 EAX: 00000000 EBX: f5cd7000 ECX: fffffe32 EDX: 00000039 ESI: 000001cd EDI: ec95fb6c EBP: e264bd80 ESP: e264bd6c DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 CR0: 80050033 CR2: 00000039 CR3: 2fe55660 CR4: 000406f0 Call Trace: __exchange_data_block+0xb3f/0x1000 [f2fs] f2fs_fallocate+0xab9/0x16b0 [f2fs] vfs_fallocate+0x17c/0x2d0 ksys_fallocate+0x42/0x70 sys_fallocate+0x31/0x40 do_fast_syscall_32+0xaa/0x22c entry_SYSENTER_32+0x4c/0x7b EIP: 0xb7f98c51 EFLAGS: 00000293 CPU: 2 EAX: ffffffda EBX: 00000003 ECX: 00000008 EDX: 01001000 ESI: 00000000 EDI: 00001000 EBP: 00000000 ESP: bfc0357c DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b Code: 00 00 d3 e8 8b 4d ec 2b 02 8b 55 f0 6b c0 1c 03 41 70 29 d6 8b 93 d0 06 00 00 8b 40 0c 83 ea 01 21 d6 89 f2 89 f1 c1 ea 03 f7 d1 <0f> be 14 10 83 e1 07 b8 01 00 00 00 d3 e0 85 c2 89 f8 0f 95 c3 EIP: is_checkpointed_data+0x84/0xd0 [f2fs] SS:ESP: 0068:e264bd6c CR2: 0000000000000039 ---[ end trace 9a4d4087cce6080a ]--- This is because in recovery flow of __exchange_data_block, we didn't pass olen to __roll_back_blkaddrs, instead we passed len, which indicates wrong array size, result in copying random block address into dnode page. Later, once that random block address was accessed by is_checkpointed_data, it can cause NULL pointer dereference. Signed-off-by:
-
Chao Yu authored
Configure io_bits with 2 and enable LFS mode, generic/013 reports below dmesg: BUG: unable to handle kernel NULL pointer dereference at 00000104 *pdpt = 0000000029b7b001 *pde = 0000000000000000 Oops: 0002 [#1] PREEMPT SMP Modules linked in: crc32_generic zram f2fs(O) rfcomm bnep bluetooth ecdh_generic snd_intel8x0 snd_ac97_codec ac97_bus snd_pcm snd_seq_midi snd_seq_midi_event snd_rawmidi snd_seq pcbc joydev snd_seq_device aesni_intel snd_timer aes_i586 snd crypto_simd cryptd soundcore i2c_piix4 serio_raw mac_hid video parport_pc ppdev lp parport hid_generic psmouse usbhid hid e1000 CPU: 0 PID: 11161 Comm: fsstress Tainted: G O 4.17.0-rc2 #38 Hardware name: innotek GmbH VirtualBox/VirtualBox, BIOS VirtualBox 12/01/2006 EIP: f2fs_submit_page_write+0x28d/0x550 [f2fs] EFLAGS: 00010206 CPU: 0 EAX: e863dcd8 EBX: 00000000 ECX: 00000100 EDX: 00000200 ESI: e863dcf4 EDI: f6f82768 EBP: e863dbb0 ESP: e863db74 DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068 CR0: 80050033 CR2: 00000104 CR3: 29a62020 CR4: 000406f0 Call Trace: do_write_page+0x6f/0xc0 [f2fs] write_data_page+0x4a/0xd0 [f2fs] do_write_data_page+0x327/0x630 [f2fs] __write_data_page+0x34b/0x820 [f2fs] __f2fs_write_data_pages+0x42d/0x8c0 [f2fs] f2fs_write_data_pages+0x27/0x30 [f2fs] do_writepages+0x1a/0x70 __filemap_fdatawrite_range+0x94/0xd0 filemap_write_and_wait_range+0x3d/0xa0 __generic_file_write_iter+0x11a/0x1f0 f2fs_file_write_iter+0xdd/0x3b0 [f2fs] __vfs_write+0xd2/0x150 vfs_write+0x9b/0x190 ksys_write+0x45/0x90 sys_write+0x16/0x20 do_fast_syscall_32+0xaa/0x22c entry_SYSENTER_32+0x4c/0x7b EIP: 0xb7fc8c51 EFLAGS: 00000246 CPU: 0 EAX: ffffffda EBX: 00000003 ECX: 09cde000 EDX: 00001000 ESI: 00000003 EDI: 00001000 EBP: 00000000 ESP: bfbded38 DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b Code: e8 f9 77 34 c9 8b 45 e0 8b 80 b8 00 00 00 39 45 d8 0f 84 bb 02 00 00 8b 45 e0 8b 80 b8 00 00 00 8d 50 d8 8b 08 89 55 f0 8b 50 04 <89> 51 04 89 0a c7 00 00 01 00 00 c7 40 04 00 02 00 00 8b 45 dc EIP: f2fs_submit_page_write+0x28d/0x550 [f2fs] SS:ESP: 0068:e863db74 CR2: 0000000000000104 ---[ end trace 4cac79c0d1305ee6 ]--- allocate_data_block will submit all sequential pending IOs sorted by a FIFO list, If we failed to submit other user's IO due to unaligned write, we will retry to allocate new block address for current IO, then it will initialize fio.list again, if fio was in the list before, it can break FIFO list, result in above panic. Thread A Thread B - do_write_page - allocate_data_block - list_add_tail : fioA cached in FIFO list. - do_write_page - allocate_data_block - list_add_tail : fioB cached in FIFO list. - f2fs_submit_page_write : fail to submit IO - allocate_data_block - INIT_LIST_HEAD - f2fs_submit_page_write - list_del <-- NULL pointer dereference This patch adds fio.retry parameter to indicate failure status for each IO, and avoid bailing out if there is still pending IO in FIFO list for fixing. Signed-off-by:
-
Anatoly Pugachev authored
The following patch disables loading of f2fs module on architectures which have PAGE_SIZE > 4096 , since it is impossible to mount f2fs on such architectures , log messages are: mount: /mnt: wrong fs type, bad option, bad superblock on /dev/vdiskb1, missing codepage or helper program, or other error. /dev/vdiskb1: F2FS filesystem, UUID=1d8b9ca4-2389-4910-af3b-10998969f09c, volume name "" May 15 18:03:13 ttip kernel: F2FS-fs (vdiskb1): Invalid page_cache_size (8192), supports only 4KB May 15 18:03:13 ttip kernel: F2FS-fs (vdiskb1): Can't find valid F2FS filesystem in 1th superblock May 15 18:03:13 ttip kernel: F2FS-fs (vdiskb1): Invalid page_cache_size (8192), supports only 4KB May 15 18:03:13 ttip kernel: F2FS-fs (vdiskb1): Can't find valid F2FS filesystem in 2th superblock May 15 18:03:13 ttip kernel: F2FS-fs (vdiskb1): Invalid page_cache_size (8192), supports only 4KB which was introduced by git commit 5c9b4692 tested on git kernel 4.17.0-rc6-00309-gec30dcf7 with patch applied: modprobe: ERROR: could not insert 'f2fs': Invalid argument May 28 01:40:28 v215 kernel: F2FS not supported on PAGE_SIZE(8192) != 4096 Signed-off-by:
Anatoly Pugachev <matorola@gmail.com> Reviewed-by:
-
Chao Yu authored
This patch fixes error path of move_data_page: - clear cold data flag if it fails to write page. - redirty page for non-ENOMEM case. Signed-off-by:
-
Chao Yu authored
As description in commit "f2fs: don't drop any page on f2fs_cp_error() case": "We still provide readdir() after shtudown, so we should keep pages to avoid additional IOs." In order to provider lastest directory structure, let's keep dentry pages in cache after fs shutdown. Signed-off-by:
-
Chao Yu authored
Thread A Thread B - f2fs_remount - stop_gc_thread - f2fs_sbi_store sbi->gc_thread = NULL; access sbi->gc_thread->gc_* Previously, we allocate memory for sbi->gc_thread based on background gc thread mount option, the memory can be released if we turn off that mount option, but still there are several places access gc_thread pointer without considering race condition, result in NULL point dereference. In order to fix this issue, use sb->s_umount to exclude those operations. Signed-off-by:
-
Chao Yu authored
Introduce clear_radix_tree_dirty_tag to include common codes for cleanup. Signed-off-by:
-
Chao Yu authored
- f2fs_fill_super - recover_fsync_data - recover_data - del_fsync_inode - iput - iput_final - write_inode_now - f2fs_write_inode - f2fs_balance_fs - f2fs_balance_fs_bg - sync_dirty_inodes With data_flush mount option, during recovery, in order to avoid entering above writeback flow, let's detect recovery status and do skip in f2fs_balance_fs_bg. Signed-off-by:Yunlei He <heyunlei@huawei.com> Signed-off-by:
-
Sheng Yong authored
If SBI_NEED_FSCK is set, discard_wake will never be cleared. As a result, the condition of wait_event_interruptible_timeout() is always true, which gets discard thread run too frequently. Signed-off-by:
Sheng Yong <shengyong1@huawei.com> Reviewed-by:
-
Yunlei He authored
This patch modify discard thread wait policy as below: issued io_interrupted wait time(ms) 1. 8 0 50 2. (0,8) 1 50 3. 0 1 500 (dev is busy) 4. 0 0 60000 (no candidates) Signed-off-by:
-
Chao Yu authored
f2fs doesn't allow abuse on atomic write class interface, so except limiting in-mem pages' total memory usage capacity, we need to limit atomic-write usage as well when filesystem is seriously fragmented, otherwise we may run into infinite loop during foreground GC because target blocks in victim segment are belong to atomic opened file for long time. Now, we will detect failure due to atomic write in foreground GC, if the count exceeds threshold, we will drop all atomic written data in cache, by this, I expect it can keep our system running safely to prevent Dos attack. In addition, his patch adds to show GC skip information in debugfs, now it just shows count of skipped caused by atomic write. Signed-off-by:
-
Jaegeuk Kim authored
This is to avoid sbi->gc_thread pointer access. Signed-off-by: -
Chao Yu authored
For non-migration IO, we will keep order of data/node blocks' submitting as allocation sequence by sorting IOs in per log io_list list, but for migration IO, it could be out-of-order. In LFS mode, we should keep all IOs including migration IO be ordered, so that this patch fixes to add an additional lock to keep submitting order. Signed-off-by:
Yunlong Song <yunlong.song@huawei.com> Signed-off-by:
-
Chao Yu authored
After revoking atomic write, related LBA can be reused by others, so we need to wait page writeback before reusing the LBA, in order to avoid interference between old atomic written in-flight IO and new IO. Signed-off-by:
-
Sahitya Tummala authored
f2fs_ioc_shutdown() ioctl gets stuck in the below path when issued with F2FS_GOING_DOWN_FULLSYNC option. __switch_to+0x90/0xc4 percpu_down_write+0x8c/0xc0 freeze_super+0xec/0x1e4 freeze_bdev+0xc4/0xcc f2fs_ioctl+0xc0c/0x1ce0 f2fs_compat_ioctl+0x98/0x1f0 Signed-off-by:
Sahitya Tummala <stummala@codeaurora.org> Reviewed-by:
-
Chao Yu authored
This patch changes to detect synchronous writeback more earlier before, in order to avoid unnecessary page writeback before exiting asynchronous writeback. Signed-off-by:
-
Chao Yu authored
- rename is_valid_blkaddr() to is_valid_meta_blkaddr() for readability. - introduce is_valid_blkaddr() for cleanup. No logic change in this patch. Signed-off-by:
-
Chao Yu authored
Since sit_i.min_mtime's type is unsigned long long, so we should initialize it with max value of the type ULLONG_MAX instead of LLONG_MAX. Signed-off-by:
-
Chao Yu authored
1. thread A: commit_inmem_pages submit data into block layer, but haven't waited it writeback. 2. thread A: commit_inmem_pages update related node. 3. thread B: do checkpoint, flush all nodes to disk. 4. SPOR Then, atomic file becomes corrupted since nodes is flushed before data. This patch fixes to treat atomic page as checkpoint guaranteed one, then in checkpoint, we can make sure all atomic page can be writebacked with metadata of atomic file. Signed-off-by:
-
Chao Yu authored
i_current_depth is used only for directory inode, but its space is shared with i_gc_failures field used for regular inode, in order to avoid affecting i_gc_failures' value, this patch fixes to initialize the union's fields according to inode type. Signed-off-by:
-
Chao Yu authored
For extreme case: 10 section, op = 10%, no_fggc_threshold = 90% All section usage: 85% 85% 85% 85% 90% 90% 95% 95% 95% 95% During foreground GC, if we skip select dirty section whose usage is larger than no_fggc_threshold, we can only recycle 80% invalid space from four 85% usage sections and two 90% usage sections, result in encountering out-of-space issue. This reverts commit e93b9865 to fix this issue, besides, we keep the logic that we scan all dirty section when searching a victim, so that GC can select victim with least valid blocks. Signed-off-by:
-
Jaegeuk Kim authored
We still provide readdir() after shtudown, so we should keep pages to avoid additional IOs. Signed-off-by: -
Colin Ian King authored
Trivial fix to spelling mistake in extension list text Signed-off-by:
Colin Ian King <colin.king@canonical.com> Reviewed-by:
-
Jaegeuk Kim authored
In order to avoid the below overflow issue, we should have checked the boundaries in superblock before reaching out to allocation. As Linus suggested, the right place should be sanity_check_raw_super(). Dr Silvio Cesare of InfoSect reported: There are integer overflows with using the cp_payload superblock field in the f2fs filesystem potentially leading to memory corruption. include/linux/f2fs_fs.h struct f2fs_super_block { ... __le32 cp_payload; fs/f2fs/f2fs.h typedef u32 block_t; /* * should not change u32, since it is the on-disk block * address format, __le32. */ ... static inline block_t __cp_payload(struct f2fs_sb_info *sbi) { return le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_payload); } fs/f2fs/checkpoint.c block_t start_blk, orphan_blocks, i, j; ... start_blk = __start_cp_addr(sbi) + 1 + __cp_payload(sbi); orphan_blocks = __start_sum_addr(sbi) - 1 - __cp_payload(sbi); +++ integer overflows ... unsigned int cp_blks = 1 + __cp_payload(sbi); ... sbi->ckpt = kzalloc(cp_blks * blk_size, GFP_KERNEL); +++ integer overflow leading to incorrect heap allocation. int cp_payload_blks = __cp_payload(sbi); ... ckpt->cp_pack_start_sum = cpu_to_le32(1 + cp_payload_blks + orphan_blocks); +++ sign bug and integer overflow ... for (i = 1; i < 1 + cp_payload_blks; i++) +++ integer overflow ... sbi->max_orphans = (sbi->blocks_per_seg - F2FS_CP_PACKS - NR_CURSEG_TYPE - __cp_payload(sbi)) * F2FS_ORPHANS_PER_BLOCK; +++ integer overflow Reported-by:Greg KH <greg@kroah.com> Reported-by:
Silvio Cesare <silvio.cesare@gmail.com> Suggested-by:
Linus Torvalds <torvalds@linux-foundation.org> Reviewed-by:
-
Chao Yu authored
Volatile file's data will be updated oftenly, so it'd better to place its data into hot data segment. In addition, for atomic file, we change to check FI_ATOMIC_FILE instead of FI_HOT_DATA to make code readability better. Signed-off-by:
-
Chao Yu authored
Introduce release_discard_addr() to include common codes for cleanup. Signed-off-by:
Fengguang Wu <fengguang.wu@intel.com> Signed-off-by:
-
Chao Yu authored
In build_sit_entries(), if valid_blocks in SIT block is smaller than valid_blocks in journal, for below calculation: sbi->discard_blks += old_valid_blocks - se->valid_blocks; There will be two times potential overflow: - old_valid_blocks - se->valid_blocks will overflow, and be a very large number. - sbi->discard_blks += result will overflow again, comes out a correct result accidently. Anyway, it should be fixed. Fixes: d600af23 ("f2fs: avoid unneeded loop in build_sit_entries") Fixes: 1f43e2ad ("f2fs: introduce CP_TRIMMED_FLAG to avoid unneeded discard") Signed-off-by:
-
Chao Yu authored
RW semphore dio_rwsem in struct f2fs_inode_info is introduced to avoid race between dio and data gc, but now, it is more wildly used to avoid foreground operation vs data gc. So rename it to i_gc_rwsem to improve its readability. Signed-off-by:
-
Yunlei He authored
This patch move mnt_want_write_file after range check, it's needless to check arguments with it. Signed-off-by:
-
Yunlei He authored
This patch fix missing clear FI_NO_PREALLOC in some error case Signed-off-by:
-
Jaegeuk Kim authored
This is to give a option for user to be able to recover B/foo in the below case. mkdir A sync() rename(A, B) creat (B/foo) fsync (B/foo) ---crash--- Sugessted-by:
Velayudhan Pillai <vijay@cs.utexas.edu> Signed-off-by:
-
