Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932646Ab2JWCc0 (ORCPT ); Mon, 22 Oct 2012 22:32:26 -0400 Received: from mailout3.samsung.com ([203.254.224.33]:48362 "EHLO mailout3.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753800Ab2JWCcX (ORCPT ); Mon, 22 Oct 2012 22:32:23 -0400 X-AuditID: cbfee61b-b7f616d00000319b-7a-508601b5ab27 From: Jaegeuk Kim To: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org, gregkh@linuxfoundation.org, viro@zeniv.linux.org.uk, arnd@arndb.de, tytso@mit.edu, chur.lee@samsung.com, cm224.lee@samsung.com, jooyoung.hwang@samsung.com References: <001001cdb0c5$2ac96520$805c2f60$%kim@samsung.com> In-reply-to: <001001cdb0c5$2ac96520$805c2f60$%kim@samsung.com> Subject: [PATCH 14/16 v2] f2fs: add garbage collection functions Date: Tue, 23 Oct 2012 11:32:20 +0900 Message-id: <001e01cdb0c6$a07381b0$e15a8510$%kim@samsung.com> MIME-version: 1.0 Content-type: text/plain; charset=UTF-8 Content-transfer-encoding: 7bit X-Mailer: Microsoft Office Outlook 12.0 Thread-index: Ac2wxSqFSBGNbwToQsqWze4D+LI2xQAAWLeg Content-language: ko DLP-Filter: Pass X-MTR: 20000000000000000@CPGS X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFjrEIsWRmVeSWpSXmKPExsVy+t8zI92tjG0BBpP2C1rs2XuSxeLyrjls DkwenzfJBTBGcdmkpOZklqUW6dslcGUcXLqMqWDePqaKFe3H2RsYWxqYuhg5OSQETCTOPT3F CmGLSVy4t56ti5GLQ0hgGaPE8SVzGWGK/hxcyAyRmM4ocebLDXYI5x+jxP7+z0AOBwebgLbE 5v0GIHERgYdA8WdL2UG6hQRsJSb9X8wMYnMK2Ek8e3QRbKqwgIPElccH2UBsFgFVieNTv7OA 2LxA9e/nLWSGsAUlfky+BxZnFlCXmDRvETOELS+xec1bZpC9EkDxR391QcIiAkYSp06fZIIo EZHY9+IdI8R4AYlvkw+xQJTLSmw6APaLhMAidonJP4+zQDwpKXFwxQ2WCYzis5BsnoVk8ywk m2chWbGAkWUVo2hqQXJBcVJ6rpFecWJucWleul5yfu4mRkhESe9gXNVgcYhRgINRiYe3Y39r gBBrYllxZe4hRgkOZiURXuUAoBBvSmJlVWpRfnxRaU5q8SFGH6DLJzJLiSbnA6M9ryTe0NjY xMzE1MTc0tTcFIewkjhvs0dKgJBAemJJanZqakFqEcw4Jg5OqQbGTMnpc83Ne+e9/7HmrH1M VcJWw6bVL1dHTjrRulNl4pbo9+xSkv4G+/TD5misf6h6Y65+0qNd1xdvqyg06HLIP+F9fgNb W8Xyowl7ti3ackDqbXOw71zFzC/fOR18t96xeOsi4VqiFPs+VvBRWNW5NhepL8eDmpn+5yl2 zGWWmlrx603UHZk+JZbijERDLeai4kQAXIOox9UCAAA= X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFvrIIsWRmVeSWpSXmKPExsVy+t9jQd2tjG0BBmd3clrs2XuSxeLyrjls DkwenzfJBTBGNTDaZKQmpqQWKaTmJeenZOal2yp5B8c7x5uaGRjqGlpamCsp5CXmptoqufgE 6Lpl5gCNVlIoS8wpBQoFJBYXK+nbYZoQGuKmawHTGKHrGxIE12NkgAYS1jFmHFy6jKlg3j6m ihXtx9kbGFsamLoYOTkkBEwk/hxcyAxhi0lcuLeerYuRi0NIYDqjxJkvN9ghnH+MEvv7PwM5 HBxsAtoSm/cbgMRFBB4CxZ8tZQfpFhKwlZj0fzHYJE4BO4lnjy4ygtjCAg4SVx4fZAOxWQRU JY5P/c4CYvMC1b+fB7GZV0BQ4sfke2BxZgF1iUnzFjFD2PISm9e8ZQbZKwEUf/RXFyQsImAk cer0SSaIEhGJfS/eMU5gFJyFZNIsJJNmIZk0C0nLAkaWVYyiqQXJBcVJ6blGesWJucWleel6 yfm5mxjB8fpMegfjqgaLQ4wCHIxKPLwd+1sDhFgTy4orcw8xSnAwK4nwKgcAhXhTEiurUovy 44tKc1KLDzH6AD06kVlKNDkfmErySuINjU3MjCyNzCyMTMzNcQgrifM2e6QECAmkJ5akZqem FqQWwYxj4uCUamDMydyhufaV5mQOeXGXRb/Z7+/Ud9iSkW19/NsiEfFZISpPhL8/0j77rJ/p 9mWLJPbZZ6/uV7qSfv5ivP8+nesXxf7xHK76fiCf9/UZZXYhi8gpx18fqGi9Xrir4sHXu37+ c4571l/uNCqNun3bv7tP696eRrFL/mkS77XNt6xZyHJ5P1/y4zwhJZbijERDLeai4kQA/Cqw egQDAAA= X-CFilter-Loop: Reflected Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 39577 Lines: 1407 This adds on-demand and background cleaning functions. - The basic background cleaning policy is trying to do cleaning jobs as much as possible whenever the system is idle. Once the background cleaning is done, the cleaner sleeps an amount of time not to interfere with VFS calls. The time is dynamically adjusted according to the status of whole segments, which is decreased when the following conditions are satisfied. . GC is not conducted currently, and . IO subsystem is idle by checking the number of requets in bdev's request list, and . There are enough dirty segments. Otherwise, the time is increased incrementally until to the maximum time. Note that, min and max times are 10 secs and 30 secs by default. - F2FS adopts a default victim selection policy where background cleaning uses a cost-benefit algorithm, while on-demand cleaning uses a greedy algorithm. - The method of moving data during the cleaning is slightly different between background and on-demand cleaning schemes. In the case of background cleaning, F2FS loads the data, and marks them as dirty. Then, F2FS expects that the data will be moved by flusher or VM. In the case of on-demand cleaning, F2FS should move the data right away. - In order to identify valid blocks in a victim segment, F2FS scans the bitmap of the segment managed as an SIT entry. Signed-off-by: Jaegeuk Kim --- fs/f2fs/gc.c | 1139 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/f2fs/gc.h | 203 +++++++++++ 2 files changed, 1342 insertions(+) create mode 100644 fs/f2fs/gc.c create mode 100644 fs/f2fs/gc.h diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c new file mode 100644 index 0000000..753b05e --- /dev/null +++ b/fs/f2fs/gc.c @@ -0,0 +1,1139 @@ +/** + * fs/f2fs/gc.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 +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "f2fs.h" +#include "node.h" +#include "segment.h" +#include "gc.h" + +static LIST_HEAD(f2fs_stat_list); +static struct kmem_cache *winode_slab; + +static int gc_thread_func(void *data) +{ + struct f2fs_sb_info *sbi = data; + wait_queue_head_t *wq = &sbi->gc_thread->gc_wait_queue_head; + long wait_ms; + + wait_ms = GC_THREAD_MIN_SLEEP_TIME; + + do { + if (try_to_freeze()) + continue; + else + wait_event_interruptible_timeout(*wq, + kthread_should_stop(), + msecs_to_jiffies(wait_ms)); + if (kthread_should_stop()) + break; + + f2fs_balance_fs(sbi); + + if (!test_opt(sbi, BG_GC)) + continue; + + /* + * [GC triggering condition] + * 0. GC is not conducted currently. + * 1. There are enough dirty segments. + * 2. IO subsystem is idle by checking the # of writeback pages. + * 3. IO subsystem is idle by checking the # of requests in + * bdev's request list. + * + * Note) We have to avoid triggering GCs too much frequently. + * Because it is possible that some segments can be + * invalidated soon after by user update or deletion. + * So, I'd like to wait some time to collect dirty segments. + */ + if (!mutex_trylock(&sbi->gc_mutex)) + continue; + + if (!is_idle(sbi)) { + wait_ms = increase_sleep_time(wait_ms); + mutex_unlock(&sbi->gc_mutex); + continue; + } + + if (has_enough_invalid_blocks(sbi)) + wait_ms = decrease_sleep_time(wait_ms); + else + wait_ms = increase_sleep_time(wait_ms); + + sbi->bg_gc++; + + if (f2fs_gc(sbi, 1) == GC_NONE) + wait_ms = GC_THREAD_NOGC_SLEEP_TIME; + else if (wait_ms == GC_THREAD_NOGC_SLEEP_TIME) + wait_ms = GC_THREAD_MAX_SLEEP_TIME; + + } while (!kthread_should_stop()); + return 0; +} + +int start_gc_thread(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_kthread *gc_th = NULL; + + gc_th = kmalloc(sizeof(struct f2fs_gc_kthread), GFP_KERNEL); + if (!gc_th) + return -ENOMEM; + + sbi->gc_thread = gc_th; + init_waitqueue_head(&sbi->gc_thread->gc_wait_queue_head); + sbi->gc_thread->f2fs_gc_task = kthread_run(gc_thread_func, sbi, + GC_THREAD_NAME); + if (IS_ERR(gc_th->f2fs_gc_task)) { + kfree(gc_th); + return -ENOMEM; + } + return 0; +} + +void stop_gc_thread(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_kthread *gc_th = sbi->gc_thread; + if (!gc_th) + return; + kthread_stop(gc_th->f2fs_gc_task); + kfree(gc_th); + sbi->gc_thread = NULL; +} + +static int select_gc_type(int gc_type) +{ + return (gc_type == BG_GC) ? GC_CB : GC_GREEDY; +} + +static void select_policy(struct f2fs_sb_info *sbi, int gc_type, + int type, struct victim_sel_policy *p) +{ + struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); + + if (IS_SSR_TYPE(type)) { + p->alloc_mode = SSR; + p->gc_mode = GC_GREEDY; + p->type = GET_SSR_TYPE(type); + p->dirty_segmap = dirty_i->dirty_segmap[p->type]; + p->ofs_unit = 1; + } else { + p->alloc_mode = LFS; + p->gc_mode = select_gc_type(gc_type); + p->type = 0; + p->dirty_segmap = dirty_i->dirty_segmap[DIRTY]; + p->ofs_unit = sbi->segs_per_sec; + } + p->offset = sbi->last_victim[p->gc_mode]; +} + +static unsigned int get_max_cost(struct f2fs_sb_info *sbi, + struct victim_sel_policy *p) +{ + if (p->gc_mode == GC_GREEDY) + return (1 << sbi->log_blocks_per_seg) * p->ofs_unit; + else if (p->gc_mode == GC_CB) + return UINT_MAX; + else /* No other gc_mode */ + return 0; +} + +static unsigned int check_bg_victims(struct f2fs_sb_info *sbi) +{ + struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); + unsigned int segno; + + /* + * If the gc_type is FG_GC, we can select victim segments + * selected by background GC before. + * Those segments guarantee they have small valid blocks. + */ + segno = find_next_bit(dirty_i->victim_segmap[BG_GC], + TOTAL_SEGS(sbi), 0); + if (segno < TOTAL_SEGS(sbi)) { + clear_bit(segno, dirty_i->victim_segmap[BG_GC]); + return segno; + } + return NULL_SEGNO; +} + +static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno) +{ + struct sit_info *sit_i = SIT_I(sbi); + unsigned int secno = GET_SECNO(sbi, segno); + unsigned int start = secno * sbi->segs_per_sec; + unsigned long long mtime = 0; + unsigned int vblocks; + unsigned char age = 0; + unsigned char u; + unsigned int i; + + for (i = 0; i < sbi->segs_per_sec; i++) + mtime += get_seg_entry(sbi, start + i)->mtime; + vblocks = get_valid_blocks(sbi, segno, sbi->segs_per_sec); + + mtime = div_u64(mtime, sbi->segs_per_sec); + vblocks = div_u64(vblocks, sbi->segs_per_sec); + + u = (vblocks * 100) >> sbi->log_blocks_per_seg; + + /* Handle if the system time is changed by user */ + if (mtime < sit_i->min_mtime) + sit_i->min_mtime = mtime; + if (mtime > sit_i->max_mtime) + sit_i->max_mtime = mtime; + if (sit_i->max_mtime != sit_i->min_mtime) + age = 100 - div64_u64(100 * (mtime - sit_i->min_mtime), + sit_i->max_mtime - sit_i->min_mtime); + + return UINT_MAX - ((100 * (100 - u) * age) / (100 + u)); +} + +static unsigned int get_gc_cost(struct f2fs_sb_info *sbi, unsigned int segno, + struct victim_sel_policy *p) +{ + if (p->alloc_mode == SSR) + return get_seg_entry(sbi, segno)->ckpt_valid_blocks; + + /* alloc_mode == LFS */ + if (p->gc_mode == GC_GREEDY) + return get_valid_blocks(sbi, segno, sbi->segs_per_sec); + else + return get_cb_cost(sbi, segno); +} + +/** + * This function is called from two pathes. + * One is garbage collection and the other is SSR segment selection. + * When it is called during GC, it just gets a victim segment + * and it does not remove it from dirty seglist. + * When it is called from SSR segment selection, it finds a segment + * which has minimum valid blocks and removes it from dirty seglist. + */ +static int get_victim_by_default(struct f2fs_sb_info *sbi, + unsigned int *result, int gc_type, int type) +{ + struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); + struct victim_sel_policy p; + unsigned int segno; + int nsearched = 0; + + select_policy(sbi, gc_type, type, &p); + + p.min_segno = NULL_SEGNO; + p.min_cost = get_max_cost(sbi, &p); + + mutex_lock(&dirty_i->seglist_lock); + + if (p.alloc_mode == LFS && gc_type == FG_GC) { + p.min_segno = check_bg_victims(sbi); + if (p.min_segno != NULL_SEGNO) + goto got_it; + } + + while (1) { + unsigned long cost; + + segno = find_next_bit(p.dirty_segmap, + TOTAL_SEGS(sbi), p.offset); + if (segno >= TOTAL_SEGS(sbi)) { + if (sbi->last_victim[p.gc_mode]) { + sbi->last_victim[p.gc_mode] = 0; + p.offset = 0; + continue; + } + break; + } + p.offset = ((segno / p.ofs_unit) * p.ofs_unit) + p.ofs_unit; + + if (test_bit(segno, dirty_i->victim_segmap[FG_GC])) + continue; + if (gc_type == BG_GC && + test_bit(segno, dirty_i->victim_segmap[BG_GC])) + continue; + if (IS_CURSEC(sbi, GET_SECNO(sbi, segno))) + continue; + + cost = get_gc_cost(sbi, segno, &p); + + if (p.min_cost > cost) { + p.min_segno = segno; + p.min_cost = cost; + } + + if (cost == get_max_cost(sbi, &p)) + continue; + + if (nsearched++ >= MAX_VICTIM_SEARCH) { + sbi->last_victim[p.gc_mode] = segno; + break; + } + } +got_it: + if (p.min_segno != NULL_SEGNO) { + *result = (p.min_segno / p.ofs_unit) * p.ofs_unit; + if (p.alloc_mode == LFS) { + int i; + for (i = 0; i < p.ofs_unit; i++) + set_bit(*result + i, + dirty_i->victim_segmap[gc_type]); + } + } + mutex_unlock(&dirty_i->seglist_lock); + + return (p.min_segno == NULL_SEGNO) ? 0 : 1; +} + +static const struct victim_selection default_v_ops = { + .get_victim = get_victim_by_default, +}; + +static struct inode *find_gc_inode(nid_t ino, struct list_head *ilist) +{ + struct list_head *this; + struct inode_entry *ie; + + list_for_each(this, ilist) { + ie = list_entry(this, struct inode_entry, list); + if (ie->inode->i_ino == ino) + return ie->inode; + } + return NULL; +} + +static void add_gc_inode(struct inode *inode, struct list_head *ilist) +{ + struct list_head *this; + struct inode_entry *new_ie, *ie; + + list_for_each(this, ilist) { + ie = list_entry(this, struct inode_entry, list); + if (ie->inode == inode) { + iput(inode); + return; + } + } +repeat: + new_ie = kmem_cache_alloc(winode_slab, GFP_NOFS); + if (!new_ie) { + cond_resched(); + goto repeat; + } + new_ie->inode = inode; + list_add_tail(&new_ie->list, ilist); +} + +static void put_gc_inode(struct list_head *ilist) +{ + struct inode_entry *ie, *next_ie; + list_for_each_entry_safe(ie, next_ie, ilist, list) { + iput(ie->inode); + list_del(&ie->list); + kmem_cache_free(winode_slab, ie); + } +} + +static int check_valid_map(struct f2fs_sb_info *sbi, + unsigned int segno, int offset) +{ + struct sit_info *sit_i = SIT_I(sbi); + struct seg_entry *sentry; + int ret; + + mutex_lock(&sit_i->sentry_lock); + sentry = get_seg_entry(sbi, segno); + ret = f2fs_test_bit(offset, sentry->cur_valid_map); + mutex_unlock(&sit_i->sentry_lock); + return ret ? GC_OK : GC_NEXT; +} + +/** + * This function compares node address got in summary with that in NAT. + * On validity, copy that node with cold status, otherwise (invalid node) + * ignore that. + */ +static int gc_node_segment(struct f2fs_sb_info *sbi, + struct f2fs_summary *sum, unsigned int segno, int gc_type) +{ + bool initial = true; + struct f2fs_summary *entry; + int off; + +next_step: + entry = sum; + for (off = 0; off < sbi->blocks_per_seg; off++, entry++) { + nid_t nid = le32_to_cpu(entry->nid); + struct page *node_page; + int err; + + /* + * It makes sure that free segments are able to write + * all the dirty node pages before CP after this CP. + * So let's check the space of dirty node pages. + */ + if (should_do_checkpoint(sbi)) { + mutex_lock(&sbi->cp_mutex); + block_operations(sbi); + return GC_BLOCKED; + } + + err = check_valid_map(sbi, segno, off); + if (err == GC_ERROR) + return err; + else if (err == GC_NEXT) + continue; + + if (initial) { + ra_node_page(sbi, nid); + continue; + } + node_page = get_node_page(sbi, nid); + if (IS_ERR(node_page)) + continue; + + /* set page dirty and write it */ + if (!PageWriteback(node_page)) + set_page_dirty(node_page); + f2fs_put_page(node_page, 1); + gc_stat_inc_node_blk_count(sbi, 1); + } + if (initial) { + initial = false; + goto next_step; + } + + if (gc_type == FG_GC) { + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = LONG_MAX, + .for_reclaim = 0, + }; + sync_node_pages(sbi, 0, &wbc); + } + return GC_DONE; +} + +/** + * Calculate start block index that this node page contains + */ +block_t start_bidx_of_node(unsigned int node_ofs) +{ + block_t start_bidx; + unsigned int bidx, indirect_blks; + int dec; + + indirect_blks = 2 * NIDS_PER_BLOCK + 4; + + start_bidx = 1; + if (node_ofs == 0) { + start_bidx = 0; + } else if (node_ofs <= 2) { + bidx = node_ofs - 1; + } else if (node_ofs <= indirect_blks) { + dec = (node_ofs - 4) / (NIDS_PER_BLOCK + 1); + bidx = node_ofs - 2 - dec; + } else { + dec = (node_ofs - indirect_blks - 3) / (NIDS_PER_BLOCK + 1); + bidx = node_ofs - 5 - dec; + } + + if (start_bidx) + start_bidx = bidx * ADDRS_PER_BLOCK + ADDRS_PER_INODE; + return start_bidx; +} + +static int check_dnode(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, + struct node_info *dni, block_t blkaddr, unsigned int *nofs) +{ + struct page *node_page; + nid_t nid; + unsigned int ofs_in_node; + block_t source_blkaddr; + + nid = le32_to_cpu(sum->nid); + ofs_in_node = le16_to_cpu(sum->ofs_in_node); + + node_page = get_node_page(sbi, nid); + if (IS_ERR(node_page)) + return GC_NEXT; + + get_node_info(sbi, nid, dni); + + if (sum->version != dni->version) { + f2fs_put_page(node_page, 1); + return GC_NEXT; + } + + *nofs = ofs_of_node(node_page); + source_blkaddr = datablock_addr(node_page, ofs_in_node); + f2fs_put_page(node_page, 1); + + if (source_blkaddr != blkaddr) + return GC_NEXT; + return GC_OK; +} + +static void move_data_page(struct inode *inode, struct page *page, int gc_type) +{ + if (page->mapping != inode->i_mapping) + goto out; + + if (inode != page->mapping->host) + goto out; + + if (PageWriteback(page)) + goto out; + + if (gc_type == BG_GC) { + set_page_dirty(page); + set_cold_data(page); + } else { + struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb); + mutex_lock_op(sbi, DATA_WRITE); + if (clear_page_dirty_for_io(page) && + S_ISDIR(inode->i_mode)) { + dec_page_count(sbi, F2FS_DIRTY_DENTS); + inode_dec_dirty_dents(inode); + } + set_cold_data(page); + do_write_data_page(page); + mutex_unlock_op(sbi, DATA_WRITE); + clear_cold_data(page); + } +out: + f2fs_put_page(page, 1); +} + +/** + * This function tries to get parent node of victim data block, and identifies + * data block validity. If the block is valid, copy that with cold status and + * modify parent node. + * If the parent node is not valid or the data block address is different, + * the victim data block is ignored. + */ +static int gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, + struct list_head *ilist, unsigned int segno, int gc_type) +{ + struct super_block *sb = sbi->sb; + struct f2fs_summary *entry; + block_t start_addr; + int err, off; + int phase = 0; + + start_addr = START_BLOCK(sbi, segno); + +next_step: + entry = sum; + for (off = 0; off < sbi->blocks_per_seg; off++, entry++) { + struct page *data_page; + struct inode *inode; + struct node_info dni; /* dnode info for the data */ + unsigned int ofs_in_node, nofs; + block_t start_bidx; + + /* + * It makes sure that free segments are able to write + * all the dirty node pages before CP after this CP. + * So let's check the space of dirty node pages. + */ + if (should_do_checkpoint(sbi)) { + mutex_lock(&sbi->cp_mutex); + block_operations(sbi); + err = GC_BLOCKED; + goto stop; + } + + err = check_valid_map(sbi, segno, off); + if (err == GC_ERROR) + goto stop; + else if (err == GC_NEXT) + continue; + + if (phase == 0) { + ra_node_page(sbi, le32_to_cpu(entry->nid)); + continue; + } + + /* Get an inode by ino with checking validity */ + err = check_dnode(sbi, entry, &dni, start_addr + off, &nofs); + if (err == GC_ERROR) + goto stop; + else if (err == GC_NEXT) + continue; + + if (phase == 1) { + ra_node_page(sbi, dni.ino); + continue; + } + + start_bidx = start_bidx_of_node(nofs); + ofs_in_node = le16_to_cpu(entry->ofs_in_node); + + if (phase == 2) { + inode = f2fs_iget_nowait(sb, dni.ino); + if (IS_ERR(inode)) + continue; + + data_page = find_data_page(inode, + start_bidx + ofs_in_node); + if (IS_ERR(data_page)) + goto next_iput; + + f2fs_put_page(data_page, 0); + add_gc_inode(inode, ilist); + } else { + inode = find_gc_inode(dni.ino, ilist); + if (inode) { + data_page = get_lock_data_page(inode, + start_bidx + ofs_in_node); + if (IS_ERR(data_page)) + continue; + move_data_page(inode, data_page, gc_type); + gc_stat_inc_data_blk_count(sbi, 1); + } + } + continue; +next_iput: + iput(inode); + } + if (++phase < 4) + goto next_step; + err = GC_DONE; +stop: + if (gc_type == FG_GC) + f2fs_submit_bio(sbi, DATA, true); + return err; +} + +static int __get_victim(struct f2fs_sb_info *sbi, unsigned int *result, + int gc_type, int type) +{ + struct sit_info *sit_i = SIT_I(sbi); + int ret; + mutex_lock(&sit_i->sentry_lock); + ret = DIRTY_I(sbi)->v_ops->get_victim(sbi, result, gc_type, type); + mutex_unlock(&sit_i->sentry_lock); + return ret; +} + +static int do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int segno, + struct list_head *ilist, int gc_type) +{ + struct page *sum_page; + struct f2fs_summary_block *sum; + int ret = GC_DONE; + + /* read segment summary of victim */ + sum_page = get_sum_page(sbi, segno); + if (IS_ERR(sum_page)) + return GC_ERROR; + + /* + * CP needs to lock sum_page. In this time, we don't need + * to lock this page, because this summary page is not gone anywhere. + * Also, this page is not gonna be updated before GC is done. + */ + unlock_page(sum_page); + sum = page_address(sum_page); + + switch (GET_SUM_TYPE((&sum->footer))) { + case SUM_TYPE_NODE: + ret = gc_node_segment(sbi, sum->entries, segno, gc_type); + break; + case SUM_TYPE_DATA: + ret = gc_data_segment(sbi, sum->entries, ilist, segno, gc_type); + break; + } + gc_stat_inc_seg_count(sbi, GET_SUM_TYPE((&sum->footer))); + gc_stat_inc_call_count(sbi->gc_info); + + f2fs_put_page(sum_page, 0); + return ret; +} + +int f2fs_gc(struct f2fs_sb_info *sbi, int nGC) +{ + unsigned int segno; + int old_free_secs, cur_free_secs; + int gc_status, nfree; + struct list_head ilist; + int gc_type = BG_GC; + + INIT_LIST_HEAD(&ilist); +gc_more: + nfree = 0; + gc_status = GC_NONE; + + if (has_not_enough_free_secs(sbi)) + old_free_secs = reserved_sections(sbi); + else + old_free_secs = free_sections(sbi); + + while (sbi->sb->s_flags & MS_ACTIVE) { + int i; + if (has_not_enough_free_secs(sbi)) + gc_type = FG_GC; + + cur_free_secs = free_sections(sbi) + nfree; + + /* We got free space successfully. */ + if (nGC < cur_free_secs - old_free_secs) + break; + + if (!__get_victim(sbi, &segno, gc_type, NO_CHECK_TYPE)) + break; + + for (i = 0; i < sbi->segs_per_sec; i++) { + /* + * do_garbage_collect will give us three gc_status: + * GC_ERROR, GC_DONE, and GC_BLOCKED. + * If GC is finished uncleanly, we have to return + * the victim to dirty segment list. + */ + gc_status = do_garbage_collect(sbi, segno + i, + &ilist, gc_type); + if (gc_status != GC_DONE) + goto stop; + nfree++; + } + } +stop: + if (has_not_enough_free_secs(sbi) || gc_status == GC_BLOCKED) { + write_checkpoint(sbi, (gc_status == GC_BLOCKED), false); + if (nfree) + goto gc_more; + } + sbi->last_gc_status = gc_status; + mutex_unlock(&sbi->gc_mutex); + + put_gc_inode(&ilist); + BUG_ON(!list_empty(&ilist)); + return gc_status; +} + +#ifdef CONFIG_F2FS_STAT_FS +void f2fs_update_stat(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_info *gc_i = sbi->gc_info; + struct f2fs_stat_info *si = gc_i->stat_info; + int i; + + /* valid check of the segment numbers */ + si->hit_ext = sbi->read_hit_ext; + si->total_ext = sbi->total_hit_ext; + si->ndirty_node = get_pages(sbi, F2FS_DIRTY_NODES); + si->ndirty_dent = get_pages(sbi, F2FS_DIRTY_DENTS); + si->ndirty_dirs = sbi->n_dirty_dirs; + si->ndirty_meta = get_pages(sbi, F2FS_DIRTY_META); + si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg; + si->rsvd_segs = reserved_segments(sbi); + si->overp_segs = overprovision_segments(sbi); + si->valid_count = valid_user_blocks(sbi); + si->valid_node_count = valid_node_count(sbi); + si->valid_inode_count = valid_inode_count(sbi); + si->utilization = utilization(sbi); + + si->free_segs = free_segments(sbi); + si->free_secs = free_sections(sbi); + si->prefree_count = prefree_segments(sbi); + si->dirty_count = dirty_segments(sbi); + si->node_pages = sbi->node_inode->i_mapping->nrpages; + si->meta_pages = sbi->meta_inode->i_mapping->nrpages; + si->nats = NM_I(sbi)->nat_cnt; + si->sits = SIT_I(sbi)->dirty_sentries; + si->fnids = NM_I(sbi)->fcnt; + si->bg_gc = sbi->bg_gc; + si->util_free = (int)(free_user_blocks(sbi) >> sbi->log_blocks_per_seg) + * 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg) + / 2; + si->util_valid = (int)(written_block_count(sbi) >> + sbi->log_blocks_per_seg) + * 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg) + / 2; + si->util_invalid = 50 - si->util_free - si->util_valid; + for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_NODE; i++) { + struct curseg_info *curseg = CURSEG_I(sbi, i); + si->curseg[i] = curseg->segno; + si->cursec[i] = curseg->segno / sbi->segs_per_sec; + si->curzone[i] = si->cursec[i] / sbi->secs_per_zone; + } + + for (i = 0; i < 2; i++) { + si->segment_count[i] = sbi->segment_count[i]; + si->block_count[i] = sbi->block_count[i]; + } +} + +/** + * This function calculates BDF of every segments + */ +void f2fs_update_gc_metric(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_info *gc_i = sbi->gc_info; + struct f2fs_stat_info *si = gc_i->stat_info; + unsigned int blks_per_sec, hblks_per_sec, total_vblocks, bimodal, dist; + struct sit_info *sit_i = SIT_I(sbi); + unsigned int segno, vblocks; + int ndirty = 0; + + bimodal = 0; + total_vblocks = 0; + blks_per_sec = sbi->segs_per_sec * (1 << sbi->log_blocks_per_seg); + hblks_per_sec = blks_per_sec / 2; + mutex_lock(&sit_i->sentry_lock); + for (segno = 0; segno < TOTAL_SEGS(sbi); segno += sbi->segs_per_sec) { + vblocks = get_valid_blocks(sbi, segno, sbi->segs_per_sec); + dist = abs(vblocks - hblks_per_sec); + bimodal += dist * dist; + + if (vblocks > 0 && vblocks < blks_per_sec) { + total_vblocks += vblocks; + ndirty++; + } + } + mutex_unlock(&sit_i->sentry_lock); + dist = sbi->total_sections * hblks_per_sec * hblks_per_sec / 100; + si->bimodal = bimodal / dist; + if (si->dirty_count) + si->avg_vblocks = total_vblocks / ndirty; + else + si->avg_vblocks = 0; +} + +static int f2fs_read_gc(char *page, char **start, off_t off, + int count, int *eof, void *data) +{ + struct f2fs_gc_info *gc_i, *next; + struct f2fs_stat_info *si; + char *buf = page; + int i = 0; + + list_for_each_entry_safe(gc_i, next, &f2fs_stat_list, stat_list) { + int j; + si = gc_i->stat_info; + + mutex_lock(&si->stat_list); + if (!si->sbi) { + mutex_unlock(&si->stat_list); + continue; + } + f2fs_update_stat(si->sbi); + + buf += sprintf(buf, "=====[ partition info. #%d ]=====\n", i++); + buf += sprintf(buf, "[SB: 1] [CP: 2] [NAT: %d] [SIT: %d] ", + si->nat_area_segs, si->sit_area_segs); + buf += sprintf(buf, "[SSA: %d] [MAIN: %d", + si->ssa_area_segs, si->main_area_segs); + buf += sprintf(buf, "(OverProv:%d Resv:%d)]\n\n", + si->overp_segs, si->rsvd_segs); + buf += sprintf(buf, "Utilization: %d%% (%d valid blocks)\n", + si->utilization, si->valid_count); + buf += sprintf(buf, " - Node: %u (Inode: %u, ", + si->valid_node_count, si->valid_inode_count); + buf += sprintf(buf, "Other: %u)\n - Data: %u\n", + si->valid_node_count - si->valid_inode_count, + si->valid_count - si->valid_node_count); + buf += sprintf(buf, "\nMain area: %d segs, %d secs %d zones\n", + si->main_area_segs, si->main_area_sections, + si->main_area_zones); + buf += sprintf(buf, " - COLD data: %d, %d, %d\n", + si->curseg[CURSEG_COLD_DATA], + si->cursec[CURSEG_COLD_DATA], + si->curzone[CURSEG_COLD_DATA]); + buf += sprintf(buf, " - WARM data: %d, %d, %d\n", + si->curseg[CURSEG_WARM_DATA], + si->cursec[CURSEG_WARM_DATA], + si->curzone[CURSEG_WARM_DATA]); + buf += sprintf(buf, " - HOT data: %d, %d, %d\n", + si->curseg[CURSEG_HOT_DATA], + si->cursec[CURSEG_HOT_DATA], + si->curzone[CURSEG_HOT_DATA]); + buf += sprintf(buf, " - Dir dnode: %d, %d, %d\n", + si->curseg[CURSEG_HOT_NODE], + si->cursec[CURSEG_HOT_NODE], + si->curzone[CURSEG_HOT_NODE]); + buf += sprintf(buf, " - File dnode: %d, %d, %d\n", + si->curseg[CURSEG_WARM_NODE], + si->cursec[CURSEG_WARM_NODE], + si->curzone[CURSEG_WARM_NODE]); + buf += sprintf(buf, " - Indir nodes: %d, %d, %d\n", + si->curseg[CURSEG_COLD_NODE], + si->cursec[CURSEG_COLD_NODE], + si->curzone[CURSEG_COLD_NODE]); + buf += sprintf(buf, "\n - Valid: %d\n - Dirty: %d\n", + si->main_area_segs - si->dirty_count - + si->prefree_count - si->free_segs, + si->dirty_count); + buf += sprintf(buf, " - Prefree: %d\n - Free: %d (%d)\n\n", + si->prefree_count, + si->free_segs, + si->free_secs); + buf += sprintf(buf, "GC calls: %d (BG: %d)\n", + si->call_count, si->bg_gc); + buf += sprintf(buf, " - data segments : %d\n", si->data_segs); + buf += sprintf(buf, " - node segments : %d\n", si->node_segs); + buf += sprintf(buf, "Try to move %d blocks\n", si->tot_blks); + buf += sprintf(buf, " - data blocks : %d\n", si->data_blks); + buf += sprintf(buf, " - node blocks : %d\n", si->node_blks); + buf += sprintf(buf, "\nExtent Hit Ratio: %d / %d\n", + si->hit_ext, si->total_ext); + buf += sprintf(buf, "\nBalancing F2FS Async:\n"); + buf += sprintf(buf, " - nodes %4d in %4d\n", + si->ndirty_node, si->node_pages); + buf += sprintf(buf, " - dents %4d in dirs:%4d\n", + si->ndirty_dent, si->ndirty_dirs); + buf += sprintf(buf, " - meta %4d in %4d\n", + si->ndirty_meta, si->meta_pages); + buf += sprintf(buf, " - NATs %5d > %lu\n", + si->nats, NM_WOUT_THRESHOLD); + buf += sprintf(buf, " - SITs: %5d\n - free_nids: %5d\n", + si->sits, si->fnids); + buf += sprintf(buf, "\nDistribution of User Blocks:"); + buf += sprintf(buf, " [ valid | invalid | free ]\n"); + buf += sprintf(buf, " ["); + for (j = 0; j < si->util_valid; j++) + buf += sprintf(buf, "-"); + buf += sprintf(buf, "|"); + for (j = 0; j < si->util_invalid; j++) + buf += sprintf(buf, "-"); + buf += sprintf(buf, "|"); + for (j = 0; j < si->util_free; j++) + buf += sprintf(buf, "-"); + buf += sprintf(buf, "]\n\n"); + buf += sprintf(buf, "SSR: %u blocks in %u segments\n", + si->block_count[SSR], si->segment_count[SSR]); + buf += sprintf(buf, "LFS: %u blocks in %u segments\n", + si->block_count[LFS], si->segment_count[LFS]); + mutex_unlock(&si->stat_list); + } + return buf - page; +} + +static int f2fs_read_sit(char *page, char **start, off_t off, + int count, int *eof, void *data) +{ + struct f2fs_gc_info *gc_i, *next; + struct f2fs_stat_info *si; + char *buf = page; + + list_for_each_entry_safe(gc_i, next, &f2fs_stat_list, stat_list) { + si = gc_i->stat_info; + + mutex_lock(&si->stat_list); + if (!si->sbi) { + mutex_unlock(&si->stat_list); + continue; + } + f2fs_update_gc_metric(si->sbi); + + buf += sprintf(buf, "BDF: %u, avg. vblocks: %u\n", + si->bimodal, si->avg_vblocks); + mutex_unlock(&si->stat_list); + } + return buf - page; +} + +static int f2fs_read_mem(char *page, char **start, off_t off, + int count, int *eof, void *data) +{ + struct f2fs_gc_info *gc_i, *next; + struct f2fs_stat_info *si; + char *buf = page; + + list_for_each_entry_safe(gc_i, next, &f2fs_stat_list, stat_list) { + struct f2fs_sb_info *sbi = gc_i->stat_info->sbi; + unsigned npages; + unsigned base_mem = 0, cache_mem = 0; + + si = gc_i->stat_info; + mutex_lock(&si->stat_list); + if (!si->sbi) { + mutex_unlock(&si->stat_list); + continue; + } + base_mem += sizeof(struct f2fs_sb_info) + sbi->sb->s_blocksize; + base_mem += 2 * sizeof(struct f2fs_inode_info); + base_mem += sizeof(*sbi->ckpt); + + /* build sm */ + base_mem += sizeof(struct f2fs_sm_info); + + /* build sit */ + base_mem += sizeof(struct sit_info); + base_mem += TOTAL_SEGS(sbi) * sizeof(struct seg_entry); + base_mem += f2fs_bitmap_size(TOTAL_SEGS(sbi)); + base_mem += 2 * SIT_VBLOCK_MAP_SIZE * TOTAL_SEGS(sbi); + if (sbi->segs_per_sec > 1) + base_mem += sbi->total_sections * + sizeof(struct sec_entry); + base_mem += __bitmap_size(sbi, SIT_BITMAP); + + /* build free segmap */ + base_mem += sizeof(struct free_segmap_info); + base_mem += f2fs_bitmap_size(TOTAL_SEGS(sbi)); + base_mem += f2fs_bitmap_size(sbi->total_sections); + + /* build curseg */ + base_mem += sizeof(struct curseg_info) * NR_CURSEG_TYPE; + base_mem += PAGE_CACHE_SIZE * NR_CURSEG_TYPE; + + /* build dirty segmap */ + base_mem += sizeof(struct dirty_seglist_info); + base_mem += NR_DIRTY_TYPE * f2fs_bitmap_size(TOTAL_SEGS(sbi)); + base_mem += 2 * f2fs_bitmap_size(TOTAL_SEGS(sbi)); + + /* buld nm */ + base_mem += sizeof(struct f2fs_nm_info); + base_mem += __bitmap_size(sbi, NAT_BITMAP); + + /* build gc */ + base_mem += sizeof(struct f2fs_gc_info); + base_mem += sizeof(struct f2fs_gc_kthread); + + /* free nids */ + cache_mem += NM_I(sbi)->fcnt; + cache_mem += NM_I(sbi)->nat_cnt; + npages = sbi->node_inode->i_mapping->nrpages; + cache_mem += npages << PAGE_CACHE_SHIFT; + npages = sbi->meta_inode->i_mapping->nrpages; + cache_mem += npages << PAGE_CACHE_SHIFT; + cache_mem += sbi->n_orphans * sizeof(struct orphan_inode_entry); + cache_mem += sbi->n_dirty_dirs * sizeof(struct dir_inode_entry); + + buf += sprintf(buf, "%u KB = static: %u + cached: %u\n", + (base_mem + cache_mem) >> 10, + base_mem >> 10, + cache_mem >> 10); + mutex_unlock(&si->stat_list); + } + return buf - page; +} + +int f2fs_stat_init(struct f2fs_sb_info *sbi) +{ + struct proc_dir_entry *entry; + + entry = create_proc_entry("f2fs_stat", 0, sbi->s_proc); + if (!entry) + return -ENOMEM; + entry->read_proc = f2fs_read_gc; + entry->write_proc = NULL; + + entry = create_proc_entry("f2fs_sit_stat", 0, sbi->s_proc); + if (!entry) { + remove_proc_entry("f2fs_stat", sbi->s_proc); + return -ENOMEM; + } + entry->read_proc = f2fs_read_sit; + entry->write_proc = NULL; + entry = create_proc_entry("f2fs_mem_stat", 0, sbi->s_proc); + if (!entry) { + remove_proc_entry("f2fs_sit_stat", sbi->s_proc); + remove_proc_entry("f2fs_stat", sbi->s_proc); + return -ENOMEM; + } + entry->read_proc = f2fs_read_mem; + entry->write_proc = NULL; + return 0; +} + +void f2fs_stat_exit(struct f2fs_sb_info *sbi) +{ + if (sbi->s_proc) { + remove_proc_entry("f2fs_stat", sbi->s_proc); + remove_proc_entry("f2fs_sit_stat", sbi->s_proc); + remove_proc_entry("f2fs_mem_stat", sbi->s_proc); + } +} +#endif + +int build_gc_manager(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_info *gc_i; + struct f2fs_checkpoint *ckp = F2FS_CKPT(sbi); +#ifdef CONFIG_F2FS_STAT_FS + struct f2fs_super_block *raw_super = F2FS_RAW_SUPER(sbi); + struct f2fs_stat_info *si; +#endif + + gc_i = kzalloc(sizeof(struct f2fs_gc_info), GFP_KERNEL); + if (!gc_i) + return -ENOMEM; + + sbi->gc_info = gc_i; + gc_i->rsvd_segment_count = le32_to_cpu(ckp->rsvd_segment_count); + gc_i->overp_segment_count = le32_to_cpu(ckp->overprov_segment_count); + + DIRTY_I(sbi)->v_ops = &default_v_ops; + +#ifdef CONFIG_F2FS_STAT_FS + gc_i->stat_info = kzalloc(sizeof(struct f2fs_stat_info), + GFP_KERNEL); + if (!gc_i->stat_info) + return -ENOMEM; + si = gc_i->stat_info; + mutex_init(&si->stat_list); + list_add_tail(&gc_i->stat_list, &f2fs_stat_list); + + si->all_area_segs = le32_to_cpu(raw_super->segment_count); + si->sit_area_segs = le32_to_cpu(raw_super->segment_count_sit); + si->nat_area_segs = le32_to_cpu(raw_super->segment_count_nat); + si->ssa_area_segs = le32_to_cpu(raw_super->segment_count_ssa); + si->main_area_segs = le32_to_cpu(raw_super->segment_count_main); + si->main_area_sections = le32_to_cpu(raw_super->section_count); + si->main_area_zones = si->main_area_sections / + le32_to_cpu(raw_super->secs_per_zone); + si->sbi = sbi; +#endif + return 0; +} + +void destroy_gc_manager(struct f2fs_sb_info *sbi) +{ + struct f2fs_gc_info *gc_i = sbi->gc_info; +#ifdef CONFIG_F2FS_STAT_FS + struct f2fs_stat_info *si = gc_i->stat_info; +#endif + if (!gc_i) + return; + +#ifdef CONFIG_F2FS_STAT_FS + list_del(&gc_i->stat_list); + mutex_lock(&si->stat_list); + si->sbi = NULL; + mutex_unlock(&si->stat_list); + kfree(gc_i->stat_info); +#endif + sbi->gc_info = NULL; + kfree(gc_i); +} + +int create_gc_caches(void) +{ + winode_slab = f2fs_kmem_cache_create("f2fs_gc_inodes", + sizeof(struct inode_entry), NULL); + if (!winode_slab) + return -ENOMEM; + return 0; +} + +void destroy_gc_caches(void) +{ + kmem_cache_destroy(winode_slab); +} diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h new file mode 100644 index 0000000..29b345d --- /dev/null +++ b/fs/f2fs/gc.h @@ -0,0 +1,203 @@ +/** + * fs/f2fs/gc.h + * + * 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. + */ +#define GC_THREAD_NAME "f2fs_gc_task" +#define GC_THREAD_MIN_WB_PAGES 1 /* + * a threshold to determine + * whether IO subsystem is idle + * or not + */ +#define GC_THREAD_MIN_SLEEP_TIME 10000 /* milliseconds */ +#define GC_THREAD_MAX_SLEEP_TIME 30000 +#define GC_THREAD_NOGC_SLEEP_TIME 10000 +#define LIMIT_INVALID_BLOCK 40 /* percentage over total user space */ +#define LIMIT_FREE_BLOCK 40 /* percentage over invalid + free space */ + +/* Search max. number of dirty segments to select a victim segment */ +#define MAX_VICTIM_SEARCH 20 + +enum { + GC_NONE = 0, + GC_ERROR, + GC_OK, + GC_NEXT, + GC_BLOCKED, + GC_DONE, +}; + +#ifdef CONFIG_F2FS_STAT_FS +struct f2fs_stat_info { + struct f2fs_sb_info *sbi; + struct mutex stat_list; + int all_area_segs; + int sit_area_segs; + int nat_area_segs; + int ssa_area_segs; + int main_area_segs; + int main_area_sections; + int main_area_zones; + int hit_ext, total_ext; + int ndirty_node; + int ndirty_dent; + int ndirty_dirs; + int ndirty_meta; + int nats, sits, fnids; + int total_count; + int utilization; + int bg_gc; + unsigned int valid_count; + unsigned int valid_node_count; + unsigned int valid_inode_count; + unsigned int bimodal, avg_vblocks; + int util_free, util_valid, util_invalid; + int rsvd_segs, overp_segs; + int dirty_count; + int node_pages; + int meta_pages; + int prefree_count; + int call_count; + int tot_segs; + int node_segs; + int data_segs; + int free_segs; + int free_secs; + int tot_blks; + int data_blks; + int node_blks; + int curseg[6]; + int cursec[6]; + int curzone[6]; + + unsigned int segment_count[2]; + unsigned int block_count[2]; +}; + +#define GC_STAT_I(gi) ((gi)->stat_info) + +#define gc_stat_inc_call_count(gi) ((GC_STAT_I(gi))->call_count++) + +#define gc_stat_inc_seg_count(sbi, type) \ + do { \ + struct f2fs_gc_info *gi = sbi->gc_info; \ + GC_STAT_I(gi)->tot_segs++; \ + if (type == SUM_TYPE_DATA) \ + GC_STAT_I(gi)->data_segs++; \ + else \ + GC_STAT_I(gi)->node_segs++; \ + } while (0) + +#define gc_stat_inc_tot_blk_count(gi, blks) \ + ((GC_STAT_I(gi)->tot_blks) += (blks)) + +#define gc_stat_inc_data_blk_count(sbi, blks) \ + do { \ + struct f2fs_gc_info *gi = sbi->gc_info; \ + gc_stat_inc_tot_blk_count(gi, blks); \ + GC_STAT_I(gi)->data_blks += (blks); \ + } while (0) + +#define gc_stat_inc_node_blk_count(sbi, blks) \ + do { \ + struct f2fs_gc_info *gi = sbi->gc_info; \ + gc_stat_inc_tot_blk_count(gi, blks); \ + GC_STAT_I(gi)->node_blks += (blks); \ + } while (0) + +#else +#define gc_stat_inc_call_count(gi) +#define gc_stat_inc_seg_count(gi, type) +#define gc_stat_inc_tot_blk_count(gi, blks) +#define gc_stat_inc_data_blk_count(gi, blks) +#define gc_stat_inc_node_blk_count(sbi, blks) +#endif + +struct f2fs_gc_kthread { + struct task_struct *f2fs_gc_task; + wait_queue_head_t gc_wait_queue_head; +}; + +struct inode_entry { + struct list_head list; + struct inode *inode; +}; + +/** + * inline functions + */ +static inline block_t free_user_blocks(struct f2fs_sb_info *sbi) +{ + if (free_segments(sbi) < overprovision_segments(sbi)) + return 0; + else + return (free_segments(sbi) - overprovision_segments(sbi)) + << sbi->log_blocks_per_seg; +} + +static inline block_t limit_invalid_user_blocks(struct f2fs_sb_info *sbi) +{ + return (long)(sbi->user_block_count * LIMIT_INVALID_BLOCK) / 100; +} + +static inline block_t limit_free_user_blocks(struct f2fs_sb_info *sbi) +{ + block_t reclaimable_user_blocks = sbi->user_block_count - + written_block_count(sbi); + return (long)(reclaimable_user_blocks * LIMIT_FREE_BLOCK) / 100; +} + +static inline long increase_sleep_time(long wait) +{ + wait += GC_THREAD_MIN_SLEEP_TIME; + if (wait > GC_THREAD_MAX_SLEEP_TIME) + wait = GC_THREAD_MAX_SLEEP_TIME; + return wait; +} + +static inline long decrease_sleep_time(long wait) +{ + wait -= GC_THREAD_MIN_SLEEP_TIME; + if (wait <= GC_THREAD_MIN_SLEEP_TIME) + wait = GC_THREAD_MIN_SLEEP_TIME; + return wait; +} + +static inline bool has_enough_invalid_blocks(struct f2fs_sb_info *sbi) +{ + block_t invalid_user_blocks = sbi->user_block_count - + written_block_count(sbi); + /* + * Background GC is triggered with the following condition. + * 1. There are a number of invalid blocks. + * 2. There is not enough free space. + */ + if (invalid_user_blocks > limit_invalid_user_blocks(sbi) && + free_user_blocks(sbi) < limit_free_user_blocks(sbi)) + return true; + return false; +} + +static inline int is_idle(struct f2fs_sb_info *sbi) +{ + struct block_device *bdev = sbi->sb->s_bdev; + struct request_queue *q = bdev_get_queue(bdev); + struct request_list *rl = &q->root_rl; + return !(rl->count[BLK_RW_SYNC]) && !(rl->count[BLK_RW_ASYNC]); +} + +static bool should_do_checkpoint(struct f2fs_sb_info *sbi) +{ + unsigned int pages_per_sec = sbi->segs_per_sec * + (1 << sbi->log_blocks_per_seg); + int node_secs = ((get_pages(sbi, F2FS_DIRTY_NODES) + pages_per_sec - 1) + >> sbi->log_blocks_per_seg) / sbi->segs_per_sec; + int dent_secs = ((get_pages(sbi, F2FS_DIRTY_DENTS) + pages_per_sec - 1) + >> sbi->log_blocks_per_seg) / sbi->segs_per_sec; + return free_sections(sbi) <= (node_secs + 2 * dent_secs + 2); +} -- 1.7.9.5 --- Jaegeuk Kim Samsung -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/