summaryrefslogtreecommitdiffstats
path: root/fs/f2fs/segment.c
Commit message (Collapse)AuthorAgeFilesLines
* f2fs: monitor the number of background checkpointJaegeuk Kim2016-01-111-0/+1
| | | | | | This patch adds to show the number of background checkpoint. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: detect idle time depending on user behaviorJaegeuk Kim2016-01-111-1/+1
| | | | | | | This patch adds last time that user requested filesystem operations. This information is used to detect whether system is idle or not later. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: introduce time and interval facilityJaegeuk Kim2016-01-111-1/+1
| | | | | | This patch adds time and interval arrays to store some timing variables. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: clean up f2fs_balance_fsJaegeuk Kim2016-01-081-2/+4
| | | | | | This patch adds one parameter to clean up all the callers of f2fs_balance_fs. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: report error of do_checkpointChao Yu2015-12-301-2/+3
| | | | | | | | | | | do_checkpoint and write_checkpoint can fail due to reasons like triggering in a readonly fs or encountering IO error of storage device. So it's better to report such error info to user, let user be aware of failure of doing checkpoint. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: support data flush in backgroundChao Yu2015-12-171-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | Previously, when finishing a checkpoint, we have persisted all fs meta info including meta inode, node inode, dentry page of directory inode, so, after a sudden power cut, f2fs can recover from last checkpoint with full directory structure. But during checkpoint, we didn't flush dirty pages of regular and symlink inode, so such dirty datas still in memory will be lost in that moment of power off. In order to reduce the chance of lost data, this patch enables f2fs_balance_fs_bg with the ability of data flushing. It will try to flush user data before starting a checkpoint. So user's data written after last checkpoint which may not be fsynced could be saved. When we mount with data_flush option, after every period of cp_interval (could be configured in sysfs: /sys/fs/f2fs/device/cp_interval) seconds user data could be flushed into device once f2fs_balance_fs_bg was called in kworker thread or gc thread. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: enhance the bit operation for SSRJaegeuk Kim2015-12-091-30/+20
| | | | | | | | This patch enhances the existing bit operation when f2fs allocates SSR blocks. Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: clean up code with __has_cursum_spaceChao Yu2015-12-041-2/+2
| | | | | | | Clean up codes in lookup_journal_in_cursum() with __has_cursum_space(). Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: clear page uptodate when dropping cache for atomic writeChao Yu2015-12-041-0/+1
| | | | | | | | | We should clear uptodate flag for all pages atomic written when we drop them, otherwise before these cached pages were reclaimed or invalidated eventually, we will see invalid data when hitting them again. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: optimize __find_rev_next_bitFan Li2015-12-041-28/+18
| | | | | | | | | | 1. Skip __reverse_ulong if the bitmap is empty. 2. Reduce branches and codes. According to my test, the performance of this new version is 5% higher on an empty bitmap of 64bytes, and remains about the same in the worst scenario. Signed-off-by: Fan li <fanofcode.li@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: fix to clear GCed flag for atomic written pageChao Yu2015-10-221-0/+1
| | | | | | | | Atomic write page can be GCed, after committing this kind of page, we should clear the GCed flag for it. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: don't need to submit bio on error caseJaegeuk Kim2015-10-211-1/+1
| | | | | | If commit_atomic_write is failed, we don't need to submit any bio. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: refactor __find_rev_next_{zero}_bitJaegeuk Kim2015-10-211-57/+49
| | | | | | | This patch refactors __find_rev_next_{zero}_bit which was disabled previously due to bugs. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs crypto: fix racing of accessing encrypted page amongChao Yu2015-10-131-0/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | 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>
* f2fs: support lower priority asynchronous readahead in ra_meta_pagesChao Yu2015-10-121-3/+3
| | | | | | | | | | | | | | Now, we use ra_meta_pages to reads continuous physical blocks as much as possible to improve performance of following reads. However, ra_meta_pages uses a synchronous readahead approach by submitting bio with READ, as READ is with high priority, it can not be used in the case of preloading blocks, and it's not sure when these RAed pages will be used. This patch supports asynchronous readahead in ra_meta_pages by tagging bio with READA flag in order to allow preloading. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: don't tag REQ_META for temporary non-meta pagesChao Yu2015-10-121-0/+3
| | | | | | | | | | | | In recovery or checkpoint flow, we grab pages temperarily in meta inode's mapping for caching temperary data, actually, datas in these pages were not meta data of f2fs, but still we tag them with REQ_META flag. However, lower device like eMMC may do some optimization for data of such type. So in order to avoid wrong optimization, we'd better remove such flag for temperary non-meta pages. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: fix SSA updates resulting in corruptionJaegeuk Kim2015-10-121-1/+32
| | | | | | | | | The f2fs_collapse_range and f2fs_insert_range changes the block addresses directly. But that can cause uncovered SSA updates. In that case, we need to give up to change the block addresses and do buffered writes to keep filesystem consistency. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: introduce a periodic checkpoint flowJaegeuk Kim2015-10-091-1/+3
| | | | | | | | | | This patch introduces a periodic checkpoint feature. Note that, this is not enforcing to conduct checkpoints very strictly in terms of trigger timing, instead just hope to help user experiences. The default value is 60 seconds. Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: support synchronous gc in ioctlChao Yu2015-10-091-1/+1
| | | | | | | | | | | | This patch drops in batches gc triggered through ioctl, since user can easily control the gc by designing the loop around the ->ioctl. We support synchronous gc by forcing using FG_GC in f2fs_gc, so with it, user can make sure that in this round all blocks gced were persistent in the device until ioctl returned. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: use vmalloc to handle -ENOMEM errorJaegeuk Kim2015-10-091-16/+16
| | | | | | This patch introduces f2fs_kvmalloc to avoid -ENOMEM during mount. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: use __GFP_NOFAIL to avoid infinite loopJaegeuk Kim2015-08-241-1/+1
| | | | | | | | | | __GFP_NOFAIL can avoid retrying the whole path of kmem_cache_alloc and bio_alloc. And, it also fixes the use cases of GFP_ATOMIC correctly. Suggested-by: Chao Yu <chao2.yu@samsung.com> Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: handle failed bio allocationJaegeuk Kim2015-08-201-3/+12
| | | | | | | | | | | | | | | | As the below comment of bio_alloc_bioset, f2fs can allocate multiple bios at the same time. So, we can't guarantee that bio is allocated all the time. " * When @bs is not NULL, if %__GFP_WAIT is set then bio_alloc will always be * able to allocate a bio. This is due to the mempool guarantees. To make this * work, callers must never allocate more than 1 bio at a time from this pool. * Callers that need to allocate more than 1 bio must always submit the * previously allocated bio for IO before attempting to allocate a new one. * Failure to do so can cause deadlocks under memory pressure. " Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: shrink free_nids entriesChao Yu2015-08-201-0/+3
| | | | | | | | This patch introduces __count_free_nids/try_to_free_nids and registers them in slab shrinker for shrinking under memory pressure. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: do not assign a new segment for dio under space shortageJaegeuk Kim2015-08-141-1/+2
| | | | | | | | If there is not enough free segment, we should not assign a new segment explicitly. Otherwise, we can run out of free segment. Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: remove inmem radix treeChao Yu2015-08-111-16/+9
| | | | | | | | | | | | | | | Previously, we use radix tree to index all registered page entries for atomic file, but now we only use radix tree to see whether current page is indexed or not, since the other user of radix tree is gone in commit 042b7816aaeb ("f2fs: remove unnecessary call to invalidate inmemory pages"). So in this patch, we try to use one more efficient way: Introducing a macro ATOMIC_WRITTEN_PAGE, and setting it as page private value to indicate page indexing status. By using this way, we can save memory and lookup time. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: invalidate temporary meta pageChao Yu2015-08-051-3/+6
| | | | | | | | | | | | | | | | To avoid meeting garbage data in next free node block at the end of warm node chain when doing recovery, we will try to zero out that invalid block. If the device is not support discard, our way for zeroing out block is: grabbing a temporary zeroed page in meta inode, then, issue write request with this page. But, we forget to release that temporary page, so our memory usage will increase without gaining any hit ratio benefit, so it's better to free it for saving memory. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: handle error cases in commit_inmem_pagesJaegeuk Kim2015-08-051-2/+8
| | | | | | | | This patch adds to handle error cases in commit_inmem_pages. If an error occurs, it stops to write the pages and return the error right away. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: shrink extent_cache entriesJaegeuk Kim2015-08-041-1/+2
| | | | | | | This patch registers shrinking extent_caches. Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: shrink nat_cache entriesJaegeuk Kim2015-08-041-2/+6
| | | | | | | This patch registers shrinking nat_cache entries. Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: call set_page_dirty to attach i_wb for cgroupJaegeuk Kim2015-07-251-0/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The cgroup attaches inode->i_wb via mark_inode_dirty and when set_page_writeback is called, __inc_wb_stat() updates i_wb's stat. So, we need to explicitly call set_page_dirty->__mark_inode_dirty in prior to any writebacking pages. This patch should resolve the following kernel panic reported by Andreas Reis. https://bugzilla.kernel.org/show_bug.cgi?id=101801 --- Comment #2 from Andreas Reis <andreas.reis@gmail.com> --- BUG: unable to handle kernel NULL pointer dereference at 00000000000000a8 IP: [<ffffffff8149deea>] __percpu_counter_add+0x1a/0x90 PGD 2951ff067 PUD 2df43f067 PMD 0 Oops: 0000 [#1] PREEMPT SMP Modules linked in: CPU: 7 PID: 10356 Comm: gcc Tainted: G W 4.2.0-1-cu #1 Hardware name: Gigabyte Technology Co., Ltd. G1.Sniper M5/G1.Sniper M5, BIOS T01 02/03/2015 task: ffff880295044f80 ti: ffff880295140000 task.ti: ffff880295140000 RIP: 0010:[<ffffffff8149deea>] [<ffffffff8149deea>] __percpu_counter_add+0x1a/0x90 RSP: 0018:ffff880295143ac8 EFLAGS: 00010082 RAX: 0000000000000003 RBX: ffffea000a526d40 RCX: 0000000000000001 RDX: 0000000000000020 RSI: 0000000000000001 RDI: 0000000000000088 RBP: ffff880295143ae8 R08: 0000000000000000 R09: ffff88008f69bb30 R10: 00000000fffffffa R11: 0000000000000000 R12: 0000000000000088 R13: 0000000000000001 R14: ffff88041d099000 R15: ffff880084a205d0 FS: 00007f8549374700(0000) GS:ffff88042f3c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00000000000000a8 CR3: 000000033e1d5000 CR4: 00000000001406e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Stack: 0000000000000000 ffffea000a526d40 ffff880084a20738 ffff880084a20750 ffff880295143b48 ffffffff811cc91e ffff880000000000 0000000000000296 0000000000000000 ffff880417090198 0000000000000000 ffffea000a526d40 Call Trace: [<ffffffff811cc91e>] __test_set_page_writeback+0xde/0x1d0 [<ffffffff813fee87>] do_write_data_page+0xe7/0x3a0 [<ffffffff813faeea>] gc_data_segment+0x5aa/0x640 [<ffffffff813fb0b8>] do_garbage_collect+0x138/0x150 [<ffffffff813fb3fe>] f2fs_gc+0x1be/0x3e0 [<ffffffff81405541>] f2fs_balance_fs+0x81/0x90 [<ffffffff813ee357>] f2fs_unlink+0x47/0x1d0 [<ffffffff81239329>] vfs_unlink+0x109/0x1b0 [<ffffffff8123e3d7>] do_unlinkat+0x287/0x2c0 [<ffffffff8123ebc6>] SyS_unlink+0x16/0x20 [<ffffffff81942e2e>] entry_SYSCALL_64_fastpath+0x12/0x71 Code: 41 5e 5d c3 0f 1f 00 66 2e 0f 1f 84 00 00 00 00 00 55 48 89 e5 41 55 49 89 f5 41 54 49 89 fc 53 48 83 ec 08 65 ff 05 e6 d9 b6 7e <48> 8b 47 20 48 63 ca 65 8b 18 48 63 db 48 01 f3 48 39 cb 7d 0a RIP [<ffffffff8149deea>] __percpu_counter_add+0x1a/0x90 RSP <ffff880295143ac8> CR2: 00000000000000a8 ---[ end trace 5132449a58ed93a3 ]--- note: gcc[10356] exited with preempt_count 2 Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: fix to return exact trimmed sizeJaegeuk Kim2015-06-021-1/+1
| | | | | | | | Now, we add all the candidates for trim commands and then finally issue discard commands. So, we should count the trimmed size in back-end. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: hide common code in f2fs_replace_blockChao Yu2015-06-021-1/+17
| | | | | | | | | | | | | This patch clean up codes through: 1.rename f2fs_replace_block to __f2fs_replace_block(). 2.introduce new f2fs_replace_block() to include __f2fs_replace_block() and some common related codes around __f2fs_replace_block(). Then, newly introduced function f2fs_replace_block can be used by following patch. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: introduce update_meta_pageChao Yu2015-06-011-10/+15
| | | | | | | | Add a help function update_meta_page() to update meta page with specified buffer. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs crypto: zero next free dnode blockChao Yu2015-06-011-1/+1
| | | | | | | | Now page cache of meta inode is used by garbage collection for encrypted page, it may contain random data, so we should zero it before issuing discard. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs crypto: shrink size of the f2fs_crypto_ctx structureJaegeuk Kim2015-06-011-1/+1
| | | | | | | | | | | | | | This patch integrates the below patch into f2fs. "ext4 crypto: shrink size of the ext4_crypto_ctx structure Some fields are only used when the crypto_ctx is being used on the read path, some are only used on the write path, and some are only used when the structure is on free list. Optimize memory use by using a union." Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: cleanup a confusing indentDan Carpenter2015-06-011-2/+2
| | | | | | | | The return was not indented far enough so it looked like it was supposed to go with the other if statement. Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: avoid buggy functionsJaegeuk Kim2015-06-011-0/+18
| | | | | | | This patch avoids to use a buggy function for now. It needs to fix them later. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: do not issue next dnode discard redundantlyJaegeuk Kim2015-06-011-1/+14
| | | | | | We have a discard map, so that we can avoid redundant discard issues. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs crypto: add encryption support in read/write pathsJaegeuk Kim2015-05-281-4/+20
| | | | | | | | | | This patch adds encryption support in read and write paths. Note that, in f2fs, we need to consider cleaning operation. In cleaning procedure, we must avoid encrypting and decrypting written blocks. So, this patch implements move_encrypted_block(). Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: introduce f2fs_replace_block() for reuseChao Yu2015-05-281-7/+24
| | | | | | | | | | Introduce a generic function replace_block base on recover_data_page, and export it. So with it we can operate file's meta data which is in CP/SSA area when we invoke fallocate with FALLOC_FL_COLLAPSE_RANGE flag. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: issue discard with finally produced len and minlenJaegeuk Kim2015-05-281-8/+6
| | | | | | | This patch determines to issue discard commands by comparing given minlen and the length of produced final candidates. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: introduce discard_map for f2fs_trim_fsJaegeuk Kim2015-05-281-28/+43
| | | | | | | | | | This patch adds a bitmap for discard issues from f2fs_trim_fs. There-in rule is to issue discard commands only for invalidated blocks after mount. Once mount is done, f2fs_trim_fs trims out whole invalid area. After ehn, it will not issue and discrads redundantly. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: add sbi and page pointer in f2fs_io_infoJaegeuk Kim2015-05-281-18/+20
| | | | | | | This patch adds f2fs_sb_info and page pointers in f2fs_io_info structure. With this change, we can reduce a lot of parameters for IO functions. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: add some tracepoints to debug volatile and atomic writesJaegeuk Kim2015-04-101-0/+4
| | | | Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: fix to cover sentry_lock for block allocationJaegeuk Kim2015-04-101-1/+1
| | | | | | | | | | | | | | | | | | | | In the following call stack, f2fs changes the bitmap for dirty segments and # of dirty sentries without grabbing sit_i->sentry_lock. This can result in mismatch on bitmap and # of dirty sentries, since if there are some direct_io operations. In allocate_data_block, - __allocate_new_segments - mutex_lock(&curseg->curseg_mutex); - s_ops->allocate_segment - new_curseg/change_curseg - reset_curseg - __set_sit_entry_type - __mark_sit_entry_dirty - set_bit(dirty_sentries_bitmap) - dirty_sentries++; Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: fix to issue small discard in real-time mode discardChao Yu2015-04-101-1/+1
| | | | | | | | | | | | Now in f2fs, we share functions and structures for batch mode and real-time mode discard. For real-time mode discard, in shared function add_discard_addrs, we will use uninitialized trim_minlen in struct cp_control to compare with length of contiguous free blocks to decide whether skipping discard fragmented freespace or not, this makes us ignore small discard sometimes. Fix it. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Reviewed-by : Changman Lee <cm224.lee@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: don't need to collect dirty sit entries and flush journal when there's ↵Wanpeng Li2015-04-101-3/+3
| | | | | | | | | | | | no dirty sit entries Don't need to collect dirty sit entries and flush sit journal to sit entries when there's no dirty sit entries. This patch check dirty_sentries earlier just like flush_nat_entries. Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com> Reviewed-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: enable rb-tree extent cacheChao Yu2015-03-031-0/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This patch enables rb-tree based extent cache in f2fs. When we mount with "-o extent_cache", f2fs will try to add recently accessed page-block mappings into rb-tree based extent cache as much as possible, instead of original one extent info cache. By this way, f2fs can support more effective cache between dnode page cache and disk. It will supply high hit ratio in the cache with fewer memory when dnode page cache are reclaimed in environment of low memory. Storage: Sandisk sd card 64g 1.append write file (offset: 0, size: 128M); 2.override write file (offset: 2M, size: 1M); 3.override write file (offset: 4M, size: 1M); ... 4.override write file (offset: 48M, size: 1M); ... 5.override write file (offset: 112M, size: 1M); 6.sync 7.echo 3 > /proc/sys/vm/drop_caches 8.read file (size:128M, unit: 4k, count: 32768) (time dd if=/mnt/f2fs/128m bs=4k count=32768) Extent Hit Ratio: before patched Hit Ratio 121 / 1071 1071 / 1071 Performance: before patched real 0m37.051s 0m35.556s user 0m0.040s 0m0.026s sys 0m2.990s 0m2.251s Memory Cost: before patched Tree Count: 0 1 (size: 24 bytes) Node Count: 0 45 (size: 1440 bytes) v3: o retest and given more details of test result. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: use spinlock for segmap_lock instead of rwlockChao Yu2015-02-111-3/+3
| | | | | | | | | | | | | | | | | | | | | | | rwlock can provide better concurrency when there are much more readers than writers because readers can hold the rwlock simultaneously. But now, for segmap_lock rwlock in struct free_segmap_info, there is only one reader 'mount' from below call path: ->f2fs_fill_super ->build_segment_manager ->build_dirty_segmap ->init_dirty_segmap ->find_next_inuse read_lock ... read_unlock Now that our concurrency can not be improved since there is no other reader for this lock, we do not need to use rwlock_t type for segmap_lock, let's replace it with spinlock_t type. Signed-off-by: Chao Yu <chao2.yu@samsung.com> Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
* f2fs: avoid variable length arrayJaegeuk Kim2015-02-111-2/+8
| | | | | | | Instead of using variable length array, this patch let preallocate memory for them. Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
OpenPOWER on IntegriCloud