- 24 Feb, 2017 2 commits
-
-
Hou Pengyang authored
As data segment gc may lead dnode dirty, so the greedy cost for data segment should be valid blocks * 2, that is data segment is prior to node segment. Signed-off-by:
Hou Pengyang <houpengyang@huawei.com> Reviewed-by:
Chao Yu <yuchao0@huawei.com> Signed-off-by:
Jaegeuk Kim <jaegeuk@kernel.org>
-
Yunlong Song authored
SIT information should be updated before segment allocation, since SSR needs latest valid block information. Current code does not update the old_blkaddr info in sit_entry, so adjust the allocate_segment to its proper location. Commit 5e443818 ("f2fs: handle dirty segments inside refresh_sit_entry") puts it into wrong location. Signed-off-by:
Yunlong Song <yunlong.song@huawei.com> Reviewed-by:
-
- 23 Feb, 2017 38 commits
-
-
Hou Pengyang authored
For foreground gc, greedy algorithm should be adapted, which makes this formula work well: (2 * (100 / config.overprovision + 1) + 6) But currently, we fg_gc have a prior to select bg_gc victim segments to gc first, these victims are selected by cost-benefit algorithm, we can't guarantee such segments have the small valid blocks, which may destroy the f2fs rule, on the worstest case, would consume all the free segments. This patch fix this by add a filter in check_bg_victims, if segment's has # of valid blocks over overprovision ratio, skip such segments. Cc: <stable@vger.kernel.org> Signed-off-by:
-
Jaegeuk Kim authored
Otherwise we can get livelock like below. [79880.428136] dbench D 0 18405 18404 0x00000000 [79880.428139] Call Trace: [79880.428142] __schedule+0x219/0x6b0 [79880.428144] schedule+0x36/0x80 [79880.428147] schedule_timeout+0x243/0x2e0 [79880.428152] ? update_sd_lb_stats+0x16b/0x5f0 [79880.428155] ? ktime_get+0x3c/0xb0 [79880.428157] io_schedule_timeout+0xa6/0x110 [79880.428161] __lock_page+0xf7/0x130 [79880.428164] ? unlock_page+0x30/0x30 [79880.428167] pagecache_get_page+0x16b/0x250 [79880.428171] grab_cache_page_write_begin+0x20/0x40 [79880.428182] f2fs_write_begin+0xa2/0xdb0 [f2fs] [79880.428192] ? f2fs_mark_inode_dirty_sync+0x16/0x30 [f2fs] [79880.428197] ? kmem_cache_free+0x79/0x200 [79880.428203] ? __mark_inode_dirty+0x17f/0x360 [79880.428206] generic_perform_write+0xbb/0x190 [79880.428213] ? file_update_time+0xa4/0xf0 [79880.428217] __generic_file_write_iter+0x19b/0x1e0 [79880.428226] f2fs_file_write_iter+0x9c/0x180 [f2fs] [79880.428231] __vfs_write+0xc5/0x140 [79880.428235] vfs_write+0xb2/0x1b0 [79880.428238] SyS_write+0x46/0xa0 [79880.428242] entry_SYSCALL_64_fastpath+0x1e/0xad Fixes: cae96a5c8ab6 ("f2fs: check io submission more precisely") Reviewed-by: -
Yunlei He authored
In FG_GC process, it will search victim section twice. This will cause some dirty section with less valid blocks skip garbage collection. section # 26425 : valid blocks # 3 142.037567: get_victim_by_default: victim 26425 : valid blocks # 3 142.037585: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26425 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 244 142.039494: f2fs_get_victim: dev = (259,30), type = Hot DATA, policy = (Background GC, SSR-mode, Greedy), victim = 19022 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 24 142.070247: new_curseg: Debug: alloc new segment 26746 142.244341: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26054 ofs_unit = 1, pre_victim_secno = 26054, prefree = 0, free = 243 142.254475: do_garbage_collect: Debug: FG_GC, seg_freed = 1 142.293131: f2fs_get_victim: dev = (259,30), type = Warm DATA, policy = (Background GC, SSR-mode, Greedy), victim = 23466 ofs_unit = 1, pre_victim_secno = -1, prefree = 0, free = 244 142.319001: f2fs_get_victim: dev = (259,30), type = Warm DATA, policy = (Background GC, SSR-mode, Greedy), victim = 23467 ofs_unit = 1, pre_victim_secno = -1, prefree = 0, free = 244 142.368879: get_victim_by_default: victim 26425 : valid blocks # 3 142.368894: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26425 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 244 142.378127: f2fs_get_victim: dev = (259,30), type = Hot DATA, policy = (Background GC, SSR-mode, Greedy), victim = 19612 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 24 142.416917: new_curseg: Debug: alloc new segment 26054 142.656794: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 25404 ofs_unit = 1, pre_victim_secno = 25404, prefree = 0, free = 243 142.662139: do_garbage_collect: Debug: FG_GC, seg_freed = 1 142.684159: new_curseg: Debug: alloc new segment 25197 142.685059: get_victim_by_default: victim 26425 : valid blocks # 3 142.685079: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26425 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 243 142.701427: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26238 ofs_unit = 1, pre_victim_secno = 26238, prefree = 0, free = 243 142.707105: do_garbage_collect: Debug: FG_GC, seg_freed = 1 142.802444: f2fs_get_victim: dev = (259,30), type = Warm DATA, policy = (Background GC, SSR-mode, Greedy), victim = 23473 ofs_unit = 1, pre_victim_secno = -1, prefree = 0, free = 244 142.804422: get_victim_by_default: victim 26425 : valid blocks # 3 142.804443: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26425 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 244 142.851567: f2fs_get_victim: dev = (259,30), type = Hot DATA, policy = (Background GC, SSR-mode, Greedy), victim = 19092 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 24 142.865014: new_curseg: Debug: alloc new segment 26238 143.082245: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26307 ofs_unit = 1, pre_victim_secno = 26307, prefree = 0, free = 244 143.088252: do_garbage_collect: Debug: FG_GC, seg_freed = 1 143.128307: new_curseg: Debug: alloc new segment 25404 143.181846: get_victim_by_default: victim 26425 : valid blocks # 3 143.181872: f2fs_get_victim: dev = (259,30), type = No TYPE, policy = (Foreground GC, LFS-mode, Greedy), victim = 26425 ofs_unit = 1, pre_victim_secno = 26425, prefree = 0, free = 244 Signed-off-by:
Yunlei He <heyunlei@huawei.com> Reviewed-by:
-
Jaegeuk Kim authored
This patch adds min_cost of each victims. Signed-off-by: -
Jaegeuk Kim authored
It turns out a stakable filesystem like sdcardfs in AOSP can trigger multiple vfs_create() to lower filesystem. In that case, f2fs will add multiple dentries having same name which breaks filesystem consistency. Until upper layer fixes, let's work around by f2fs, which shows actually not much performance regression. Cc: <stable@vger.kernel.org> Signed-off-by: -
Jaegeuk Kim authored
This patch shows actual device information in the tracepoints. Signed-off-by: -
Jaegeuk Kim authored
We have had node chains, but haven't used it so far due to stale node blocks. Now, we have crc|cp_ver in node footer and give random cp_ver at format time, we can start to use it again. Signed-off-by: -
Chao Yu authored
In android, since SElinux is enable, security policy will be appliedd for each file, it stores in inode as an xattr entry, so it will take one 4k size node block additionally for each file. Let's enable inline_xattr by default in order to save storage space. Signed-off-by:
-
Chao Yu authored
This patch introduces new mount option 'noinline_xattr', so we can disable inline xattr functionality which is already set as a default mount option. Signed-off-by:
-
Jaegeuk Kim authored
We've not seen this buggy case for a long time, so it's time to avoid this unnecessary get_node_info() call which reading NAT page to cache nat entry. Signed-off-by: -
Jaegeuk Kim authored
Let's avoid build_free_nids() in checkpoint path. Signed-off-by: -
Chao Yu authored
Currently, if we call fsync after updating the xattr date belongs to the file, f2fs needs to trigger checkpoint to keep xattr data consistent. But, this policy cause low performance as checkpoint will block most foreground operations and cause unneeded and unrelated IOs around checkpoint. This patch will reuse regular file recovery policy for xattr node block, so, we change to write xattr node block tagged with fsync flag to warm area instead of cold area, and during recovery, we search warm node chain for fsynced xattr block, and do the recovery. So, for below application IO pattern, performance can be improved obviously: - touch file - create/update/delete xattr entry in file - fsync file Signed-off-by:
-
Bhumika Goyal authored
Declare fscrypt_operations structure as const as it is only stored in the s_cop field of a super_block structure. This field is of type const, so fscrypt_operations structure having this property can be made const too. File size before: fs/f2fs/super.o text data bss dec hex filename 54131 31355 184 85670 14ea6 fs/f2fs/super.o File size after: fs/f2fs/super.o text data bss dec hex filename 54227 31259 184 85670 14ea6 fs/f2fs/super.o Signed-off-by:
Bhumika Goyal <bhumirks@gmail.com> Signed-off-by:
-
Jaegeuk Kim authored
If we mounted f2fs successfully, let's show current checkpoint version. Signed-off-by: -
Tiezhu Yang authored
There is a typo "f2f2" in f2fs.txt, this patch fixes it. Signed-off-by:
Tiezhu Yang <kernelpatch@126.com> Signed-off-by:
-
Jaegeuk Kim authored
This patch removes REQ_PREFLUSH in the nobarrier case. Signed-off-by: -
Jaegeuk Kim authored
If the cached bio has the last page's index, then we need to submit it. Otherwise, we don't need to submit it and can wait for further IO merges. Reviewed-by:
-
Jaegeuk Kim authored
This patch check IO submission more precisely than previous rough check. Signed-off-by: -
Jaegeuk Kim authored
This patch introduces __write_data_page to call it by f2fs_write_cache_pages directly.. Signed-off-by: -
Jaegeuk Kim authored
We need to flush data writes before flushing last node block writes by using FUA with PREFLUSH. We don't need to guarantee precedent node writes since if those are not written, we can't reach to the last node block when scanning node block chain during roll-forward recovery. Afterwards f2fs_wait_on_page_writeback guarantees all the IO submission to disk, which builds a valid node block chain. Reviewed-by:
-
Jaegeuk Kim authored
This patch just moves write_node_page and introduces an inner function. Reviewed-by:
-
Jaegeuk Kim authored
This patch moves the tracepoint location for flush command. Reviewed-by:
-
Jaegeuk Kim authored
This patch shows cached # of APPEND and UPDATE inode entries. Reviewed-by:
-
DongOh Shin authored
1) Nine coding style warnings below have been resolved: "Missing a blank line after declarations" 2) 435 coding style warnings below have been resolved: "function definition argument 'x' should also have an identifier name" 3) Two coding style warnings below have been resolved: "macros should not use a trailing semicolon" Signed-off-by:
DongOh Shin <doscode.kr@gmail.com> Signed-off-by:
-
DongOh Shin authored
Two coding style errors below have been resolved: "Macros with complex values should be enclosed in parentheses" And a coding style error below has been resolved: "space prohibited before that ',' (ctx:WxW)" Signed-off-by:
-
Jaegeuk Kim authored
We missed two functions declared as static functions. Signed-off-by: -
Kaixu Xia authored
This patch shows the fault injection mount option in f2fs_show_options(). Signed-off-by:
Kaixu Xia <xiakaixu@huawei.com> Signed-off-by:
-
Chao Yu authored
We only allocate flush merge control structure sbi::sm_info::fcc_info when flush_merge option is on, but in f2fs_issue_flush we still try to access member of the control structure without that option, it incurs panic as show below, fix it. Call Trace: __remove_ino_entry+0xa9/0xc0 [f2fs] f2fs_do_sync_file.isra.27+0x214/0x6d0 [f2fs] f2fs_sync_file+0x18/0x20 [f2fs] vfs_fsync_range+0x3d/0xb0 __do_page_fault+0x261/0x4d0 do_fsync+0x3d/0x70 SyS_fsync+0x10/0x20 do_syscall_64+0x6e/0x180 entry_SYSCALL64_slow_path+0x25/0x25 RIP: 0033:0x7f18ce260de0 RSP: 002b:00007ffdd4589258 EFLAGS: 00000246 ORIG_RAX: 000000000000004a RAX: ffffffffffffffda RBX: 0000000000000001 RCX: 00007f18ce260de0 RDX: 0000000000000006 RSI: 00000000016c0360 RDI: 0000000000000003 RBP: 00000000016c0360 R08: 000000000000ffff R09: 000000000000001f R10: 00007ffdd4589020 R11: 0000000000000246 R12: 00000000016c0100 R13: 0000000000000000 R14: 00000000016c1f00 R15: 00000000016c0100 Code: fb 81 e3 00 08 00 00 48 89 45 a0 0f 1f 44 00 00 31 c0 85 db 75 27 41 81 e7 00 04 00 00 74 0c 41 8b 45 20 85 c0 0f 85 81 00 00 00 <f0> 41 ff 45 20 4c 89 e7 e8 f8 e9 ff ff f0 41 ff 4d 20 48 83 c4 RIP: f2fs_issue_flush+0x5b/0x170 [f2fs] RSP: ffffc90003b5fd78 CR2: 0000000000000020 ---[ end trace a09314c24f037648 ]--- Reported-by:
Shuoran Liu <liushuoran@huawei.com> Signed-off-by:
-
Chao Yu authored
We use following method to calculate size with current page index: size = index << PAGE_SHIFT If type of index has only 32-bits size, left shifting will incur overflow, which makes result incorrect. So let's cast index with 64-bits type to avoid such issue. Signed-off-by:
-
Chao Yu authored
Previously, in getxattr we will load all entries both in inline xattr and xattr node block, and then do the lookup in all entries, but our lookup flow shows low efficiency, since if we can lookup and hit in inline xattr of inode page cache first, we don't need to load and lookup xattr node block, which can obviously save cpu time and IO latency. Signed-off-by:
-
Masanari Iida authored
This patch fix a typo in f2fs.txt Signed-off-by:
Masanari Iida <standby24x7@gmail.com> Signed-off-by:
-
Wei Fang authored
A dead loop can be triggered in f2fs_fiemap() using the test case as below: ... fd = open(); fallocate(fd, 0, 0, 4294967296); ioctl(fd, FS_IOC_FIEMAP, fiemap_buf); ... It's caused by an overflow in __get_data_block(): ... bh->b_size = map.m_len << inode->i_blkbits; ... map.m_len is an unsigned int, and bh->b_size is a size_t which is 64 bits on 64 bits archtecture, type conversion from an unsigned int to a size_t will result in an overflow. In the above-mentioned case, bh->b_size will be zero, and f2fs_fiemap() will call get_data_block() at block 0 again an again. Fix this by adding a force conversion before left shift. Signed-off-by:
Wei Fang <fangwei1@huawei.com> Acked-by:
-
Jaegeuk Kim authored
Sheng Yong reports needless preallocation if write(small_buffer, large_size) is called. In that case, f2fs preallocates large_size, but vfs returns early due to small_buffer size. Let's detect it before preallocation phase in f2fs. Reported-by:
Sheng Yong <shengyong1@huawei.com> Signed-off-by:
-
Jaegeuk Kim authored
This patch adds stat information for flush and discard commands. Reviewed-by:
-
Jaegeuk Kim authored
This patch adds a kernel thread to issue discard commands. It proposes three states, D_PREP, D_SUBMIT, and D_DONE to identify current bio status. Reviewed-by:
-
Jaegeuk Kim authored
This patch adds discard_cmd_control with the existing discarding controls. Reviewed-by:
-
Jaegeuk Kim authored
This patch modifies stat information more clearly. Reviewed-by:
-
Jaegeuk Kim authored
This patch simply cleans up the names for flush/discard commands. Reviewed-by:
-
