diff options
Diffstat (limited to 'fs/nilfs2/dat.c')
| -rw-r--r-- | fs/nilfs2/dat.c | 240 |
1 files changed, 128 insertions, 112 deletions
diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index dffedb2f8817..674380837ab9 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -1,18 +1,9 @@ +// SPDX-License-Identifier: GPL-2.0+ /* - * dat.c - NILFS disk address translation. + * NILFS disk address translation. * * Copyright (C) 2006-2008 Nippon Telegraph and Telephone Corporation. * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License as published by - * the Free Software Foundation; either version 2 of the License, or - * (at your option) any later version. - * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * * Written by Koji Sato. */ @@ -49,8 +40,21 @@ static inline struct nilfs_dat_info *NILFS_DAT_I(struct inode *dat) static int nilfs_dat_prepare_entry(struct inode *dat, struct nilfs_palloc_req *req, int create) { - return nilfs_palloc_get_entry_block(dat, req->pr_entry_nr, - create, &req->pr_entry_bh); + int ret; + + ret = nilfs_palloc_get_entry_block(dat, req->pr_entry_nr, + create, &req->pr_entry_bh); + if (unlikely(ret == -ENOENT)) { + nilfs_err(dat->i_sb, + "DAT doesn't have a block to manage vblocknr = %llu", + (unsigned long long)req->pr_entry_nr); + /* + * Return internal code -EINVAL to notify bmap layer of + * metadata corruption. + */ + ret = -EINVAL; + } + return ret; } static void nilfs_dat_commit_entry(struct inode *dat, @@ -71,7 +75,7 @@ int nilfs_dat_prepare_alloc(struct inode *dat, struct nilfs_palloc_req *req) { int ret; - ret = nilfs_palloc_prepare_alloc_entry(dat, req); + ret = nilfs_palloc_prepare_alloc_entry(dat, req, true); if (ret < 0) return ret; @@ -85,15 +89,15 @@ int nilfs_dat_prepare_alloc(struct inode *dat, struct nilfs_palloc_req *req) void nilfs_dat_commit_alloc(struct inode *dat, struct nilfs_palloc_req *req) { struct nilfs_dat_entry *entry; - void *kaddr; + size_t offset; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); entry->de_start = cpu_to_le64(NILFS_CNO_MIN); entry->de_end = cpu_to_le64(NILFS_CNO_MAX); entry->de_blocknr = cpu_to_le64(0); - kunmap_atomic(kaddr); + kunmap_local(entry); nilfs_palloc_commit_alloc_entry(dat, req); nilfs_dat_commit_entry(dat, req); @@ -109,41 +113,44 @@ static void nilfs_dat_commit_free(struct inode *dat, struct nilfs_palloc_req *req) { struct nilfs_dat_entry *entry; - void *kaddr; + size_t offset; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); entry->de_start = cpu_to_le64(NILFS_CNO_MIN); entry->de_end = cpu_to_le64(NILFS_CNO_MIN); entry->de_blocknr = cpu_to_le64(0); - kunmap_atomic(kaddr); + kunmap_local(entry); nilfs_dat_commit_entry(dat, req); + + if (unlikely(req->pr_desc_bh == NULL || req->pr_bitmap_bh == NULL)) { + nilfs_error(dat->i_sb, + "state inconsistency probably due to duplicate use of vblocknr = %llu", + (unsigned long long)req->pr_entry_nr); + return; + } nilfs_palloc_commit_free_entry(dat, req); } int nilfs_dat_prepare_start(struct inode *dat, struct nilfs_palloc_req *req) { - int ret; - - ret = nilfs_dat_prepare_entry(dat, req, 0); - WARN_ON(ret == -ENOENT); - return ret; + return nilfs_dat_prepare_entry(dat, req, 0); } void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req, sector_t blocknr) { struct nilfs_dat_entry *entry; - void *kaddr; + size_t offset; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); entry->de_start = cpu_to_le64(nilfs_mdt_cno(dat)); entry->de_blocknr = cpu_to_le64(blocknr); - kunmap_atomic(kaddr); + kunmap_local(entry); nilfs_dat_commit_entry(dat, req); } @@ -151,21 +158,21 @@ void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req, int nilfs_dat_prepare_end(struct inode *dat, struct nilfs_palloc_req *req) { struct nilfs_dat_entry *entry; + __u64 start; sector_t blocknr; - void *kaddr; + size_t offset; int ret; ret = nilfs_dat_prepare_entry(dat, req, 0); - if (ret < 0) { - WARN_ON(ret == -ENOENT); + if (ret < 0) return ret; - } - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); + start = le64_to_cpu(entry->de_start); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(entry); if (blocknr == 0) { ret = nilfs_palloc_prepare_free_entry(dat, req); @@ -174,6 +181,15 @@ int nilfs_dat_prepare_end(struct inode *dat, struct nilfs_palloc_req *req) return ret; } } + if (unlikely(start > nilfs_mdt_cno(dat))) { + nilfs_err(dat->i_sb, + "vblocknr = %llu has abnormal lifetime: start cno (= %llu) > current cno (= %llu)", + (unsigned long long)req->pr_entry_nr, + (unsigned long long)start, + (unsigned long long)nilfs_mdt_cno(dat)); + nilfs_dat_abort_entry(dat, req); + return -EINVAL; + } return 0; } @@ -184,11 +200,11 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req, struct nilfs_dat_entry *entry; __u64 start, end; sector_t blocknr; - void *kaddr; + size_t offset; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); end = start = le64_to_cpu(entry->de_start); if (!dead) { end = nilfs_mdt_cno(dat); @@ -196,7 +212,7 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req, } entry->de_end = cpu_to_le64(end); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(entry); if (blocknr == 0) nilfs_dat_commit_free(dat, req); @@ -209,14 +225,14 @@ void nilfs_dat_abort_end(struct inode *dat, struct nilfs_palloc_req *req) struct nilfs_dat_entry *entry; __u64 start; sector_t blocknr; - void *kaddr; + size_t offset; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, - req->pr_entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, req->pr_entry_nr, + req->pr_entry_bh); + entry = kmap_local_folio(req->pr_entry_bh->b_folio, offset); start = le64_to_cpu(entry->de_start); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(entry); if (start == nilfs_mdt_cno(dat) && blocknr == 0) nilfs_palloc_abort_free_entry(dat, req); @@ -255,18 +271,16 @@ void nilfs_dat_abort_update(struct inode *dat, } /** - * nilfs_dat_mark_dirty - - * @dat: DAT file inode + * nilfs_dat_mark_dirty - mark the DAT block buffer containing the specified + * virtual block address entry as dirty + * @dat: DAT file inode * @vblocknr: virtual block number * - * Description: - * - * Return Value: On success, 0 is returned. On error, one of the following - * negative error codes is returned. - * - * %-EIO - I/O error. - * - * %-ENOMEM - Insufficient amount of memory available. + * Return: 0 on success, or one of the following negative error codes on + * failure: + * * %-EINVAL - Invalid DAT entry (internal code). + * * %-EIO - I/O error (including metadata corruption). + * * %-ENOMEM - Insufficient memory available. */ int nilfs_dat_mark_dirty(struct inode *dat, __u64 vblocknr) { @@ -289,14 +303,11 @@ int nilfs_dat_mark_dirty(struct inode *dat, __u64 vblocknr) * Description: nilfs_dat_freev() frees the virtual block numbers specified by * @vblocknrs and @nitems. * - * Return Value: On success, 0 is returned. On error, one of the following - * negative error codes is returned. - * - * %-EIO - I/O error. - * - * %-ENOMEM - Insufficient amount of memory available. - * - * %-ENOENT - The virtual block number have not been allocated. + * Return: 0 on success, or one of the following negative error codes on + * failure: + * * %-EIO - I/O error (including metadata corruption). + * * %-ENOENT - The virtual block number have not been allocated. + * * %-ENOMEM - Insufficient memory available. */ int nilfs_dat_freev(struct inode *dat, __u64 *vblocknrs, size_t nitems) { @@ -312,18 +323,16 @@ int nilfs_dat_freev(struct inode *dat, __u64 *vblocknrs, size_t nitems) * Description: nilfs_dat_move() changes the block number associated with * @vblocknr to @blocknr. * - * Return Value: On success, 0 is returned. On error, one of the following - * negative error codes is returned. - * - * %-EIO - I/O error. - * - * %-ENOMEM - Insufficient amount of memory available. + * Return: 0 on success, or one of the following negative error codes on + * failure: + * * %-EIO - I/O error (including metadata corruption). + * * %-ENOMEM - Insufficient memory available. */ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) { struct buffer_head *entry_bh; struct nilfs_dat_entry *entry; - void *kaddr; + size_t offset; int ret; ret = nilfs_palloc_get_entry_block(dat, vblocknr, 0, &entry_bh); @@ -346,21 +355,21 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) } } - kaddr = kmap_atomic(entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, vblocknr, entry_bh); + entry = kmap_local_folio(entry_bh->b_folio, offset); if (unlikely(entry->de_blocknr == cpu_to_le64(0))) { - nilfs_msg(dat->i_sb, KERN_CRIT, - "%s: invalid vblocknr = %llu, [%llu, %llu)", - __func__, (unsigned long long)vblocknr, - (unsigned long long)le64_to_cpu(entry->de_start), - (unsigned long long)le64_to_cpu(entry->de_end)); - kunmap_atomic(kaddr); + nilfs_crit(dat->i_sb, + "%s: invalid vblocknr = %llu, [%llu, %llu)", + __func__, (unsigned long long)vblocknr, + (unsigned long long)le64_to_cpu(entry->de_start), + (unsigned long long)le64_to_cpu(entry->de_end)); + kunmap_local(entry); brelse(entry_bh); return -EINVAL; } WARN_ON(blocknr == 0); entry->de_blocknr = cpu_to_le64(blocknr); - kunmap_atomic(kaddr); + kunmap_local(entry); mark_buffer_dirty(entry_bh); nilfs_mdt_mark_dirty(dat); @@ -377,24 +386,21 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) * @blocknrp: pointer to a block number * * Description: nilfs_dat_translate() maps the virtual block number @vblocknr - * to the corresponding block number. - * - * Return Value: On success, 0 is returned and the block number associated - * with @vblocknr is stored in the place pointed by @blocknrp. On error, one - * of the following negative error codes is returned. - * - * %-EIO - I/O error. - * - * %-ENOMEM - Insufficient amount of memory available. + * to the corresponding block number. The block number associated with + * @vblocknr is stored in the place pointed to by @blocknrp. * - * %-ENOENT - A block number associated with @vblocknr does not exist. + * Return: 0 on success, or one of the following negative error codes on + * failure: + * * %-EIO - I/O error (including metadata corruption). + * * %-ENOENT - A block number associated with @vblocknr does not exist. + * * %-ENOMEM - Insufficient memory available. */ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) { struct buffer_head *entry_bh, *bh; struct nilfs_dat_entry *entry; sector_t blocknr; - void *kaddr; + size_t offset; int ret; ret = nilfs_palloc_get_entry_block(dat, vblocknr, 0, &entry_bh); @@ -410,8 +416,8 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) } } - kaddr = kmap_atomic(entry_bh->b_page); - entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); + offset = nilfs_palloc_entry_offset(dat, vblocknr, entry_bh); + entry = kmap_local_folio(entry_bh->b_folio, offset); blocknr = le64_to_cpu(entry->de_blocknr); if (blocknr == 0) { ret = -ENOENT; @@ -420,7 +426,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) *blocknrp = blocknr; out: - kunmap_atomic(kaddr); + kunmap_local(entry); brelse(entry_bh); return ret; } @@ -429,11 +435,12 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, size_t nvi) { struct buffer_head *entry_bh; - struct nilfs_dat_entry *entry; + struct nilfs_dat_entry *entry, *first_entry; struct nilfs_vinfo *vinfo = buf; __u64 first, last; - void *kaddr; + size_t offset; unsigned long entries_per_block = NILFS_MDT(dat)->mi_entries_per_block; + unsigned int entry_size = NILFS_MDT(dat)->mi_entry_size; int i, j, n, ret; for (i = 0; i < nvi; i += n) { @@ -441,23 +448,28 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, 0, &entry_bh); if (ret < 0) return ret; - kaddr = kmap_atomic(entry_bh->b_page); - /* last virtual block number in this block */ + first = vinfo->vi_vblocknr; - do_div(first, entries_per_block); + first = div64_ul(first, entries_per_block); first *= entries_per_block; + /* first virtual block number in this block */ + last = first + entries_per_block - 1; + /* last virtual block number in this block */ + + offset = nilfs_palloc_entry_offset(dat, first, entry_bh); + first_entry = kmap_local_folio(entry_bh->b_folio, offset); for (j = i, n = 0; j < nvi && vinfo->vi_vblocknr >= first && vinfo->vi_vblocknr <= last; j++, n++, vinfo = (void *)vinfo + visz) { - entry = nilfs_palloc_block_get_entry( - dat, vinfo->vi_vblocknr, entry_bh, kaddr); + entry = (void *)first_entry + + (vinfo->vi_vblocknr - first) * entry_size; vinfo->vi_start = le64_to_cpu(entry->de_start); vinfo->vi_end = le64_to_cpu(entry->de_end); vinfo->vi_blocknr = le64_to_cpu(entry->de_blocknr); } - kunmap_atomic(kaddr); + kunmap_local(first_entry); brelse(entry_bh); } @@ -470,6 +482,8 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, * @entry_size: size of a dat entry * @raw_inode: on-disk dat inode * @inodep: buffer to store the inode + * + * Return: 0 on success, or a negative error code on failure. */ int nilfs_dat_read(struct super_block *sb, size_t entry_size, struct nilfs_inode *raw_inode, struct inode **inodep) @@ -480,11 +494,11 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size, int err; if (entry_size > sb->s_blocksize) { - nilfs_msg(sb, KERN_ERR, "too large DAT entry size: %zu bytes", + nilfs_err(sb, "too large DAT entry size: %zu bytes", entry_size); return -EINVAL; } else if (entry_size < NILFS_MIN_DAT_ENTRY_SIZE) { - nilfs_msg(sb, KERN_ERR, "too small DAT entry size: %zu bytes", + nilfs_err(sb, "too small DAT entry size: %zu bytes", entry_size); return -EINVAL; } @@ -492,7 +506,7 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size, dat = nilfs_iget_locked(sb, NULL, NILFS_DAT_INO); if (unlikely(!dat)) return -ENOMEM; - if (!(dat->i_state & I_NEW)) + if (!(inode_state_read_once(dat) & I_NEW)) goto out; err = nilfs_mdt_init(dat, NILFS_MDT_GFP, sizeof(*di)); @@ -506,7 +520,9 @@ int nilfs_dat_read(struct super_block *sb, size_t entry_size, di = NILFS_DAT_I(dat); lockdep_set_class(&di->mi.mi_sem, &dat_lock_key); nilfs_palloc_setup_cache(dat, &di->palloc_cache); - nilfs_mdt_setup_shadow_map(dat, &di->shadow); + err = nilfs_mdt_setup_shadow_map(dat, &di->shadow); + if (err) + goto failed; err = nilfs_read_inode_common(dat, raw_inode); if (err) |
