X-Git-Url: https://git.sur5r.net/?a=blobdiff_plain;f=fs%2Fubifs%2Fsuper.c;h=9603163d8a04129f3017c500ad668344e5872168;hb=5a08cfee3967d6e8174d7de135af1daa8e4aea00;hp=39e3efec802d6929ac914ec88d10afa720e74d47;hpb=4185ae7a7146a77a792e5ba49845e5ca6e95b038;p=u-boot diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 39e3efec80..9603163d8a 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -1,21 +1,9 @@ +// SPDX-License-Identifier: GPL-2.0+ /* * This file is part of UBIFS. * * Copyright (C) 2006-2008 Nokia Corporation. * - * 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. - * - * 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. - * - * You should have received a copy of the GNU General Public License along with - * this program; if not, write to the Free Software Foundation, Inc., 51 - * Franklin St, Fifth Floor, Boston, MA 02110-1301 USA - * * Authors: Artem Bityutskiy (Битюцкий Артём) * Adrian Hunter */ @@ -26,108 +14,53 @@ * corresponding subsystems, but most of it is here. */ -#include "ubifs.h" +#ifndef __UBOOT__ +#include +#include +#include +#include +#include +#include +#include +#include #include +#include +#else + +#include +#include +#include +#include +#include +#include +#include +#include "ubifs.h" +#include +#include + +struct dentry; +struct file; +struct iattr; +struct kstat; +struct vfsmount; #define INODE_LOCKED_MAX 64 struct super_block *ubifs_sb; -static struct inode *inodes_locked_down[INODE_LOCKED_MAX]; - -/* shrinker.c */ -/* List of all UBIFS file-system instances */ -struct list_head ubifs_infos; - -/* linux/fs/super.c */ +static struct inode *inodes_locked_down[INODE_LOCKED_MAX]; -static int sb_set(struct super_block *sb, void *data) +int set_anon_super(struct super_block *s, void *data) { - dev_t *dev = data; - - sb->s_dev = *dev; return 0; } -/** - * sget - find or create a superblock - * @type: filesystem type superblock should belong to - * @test: comparison callback - * @set: setup callback - * @data: argument to each of them - */ -struct super_block *sget(struct file_system_type *type, - int (*test)(struct super_block *,void *), - int (*set)(struct super_block *,void *), - void *data) -{ - struct super_block *s = NULL; - int err; - - s = kzalloc(sizeof(struct super_block), GFP_USER); - if (!s) { - err = -ENOMEM; - return ERR_PTR(err); - } - - INIT_LIST_HEAD(&s->s_instances); - INIT_LIST_HEAD(&s->s_inodes); - s->s_time_gran = 1000000000; - - err = set(s, data); - if (err) { - return ERR_PTR(err); - } - s->s_type = type; - strncpy(s->s_id, type->name, sizeof(s->s_id)); - list_add(&s->s_instances, &type->fs_supers); - return s; -} - -/** - * validate_inode - validate inode. - * @c: UBIFS file-system description object - * @inode: the inode to validate - * - * This is a helper function for 'ubifs_iget()' which validates various fields - * of a newly built inode to make sure they contain sane values and prevent - * possible vulnerabilities. Returns zero if the inode is all right and - * a non-zero error code if not. - */ -static int validate_inode(struct ubifs_info *c, const struct inode *inode) -{ - int err; - const struct ubifs_inode *ui = ubifs_inode(inode); - - if (inode->i_size > c->max_inode_sz) { - ubifs_err("inode is too large (%lld)", - (long long)inode->i_size); - return 1; - } - - if (ui->compr_type < 0 || ui->compr_type >= UBIFS_COMPR_TYPES_CNT) { - ubifs_err("unknown compression type %d", ui->compr_type); - return 2; - } - - if (ui->data_len < 0 || ui->data_len > UBIFS_MAX_INO_DATA) - return 4; - - if (!ubifs_compr_present(ui->compr_type)) { - ubifs_warn("inode %lu uses '%s' compression, but it was not " - "compiled in", inode->i_ino, - ubifs_compr_name(ui->compr_type)); - } - - err = dbg_check_dir_size(c, inode); - return err; -} - struct inode *iget_locked(struct super_block *sb, unsigned long ino) { struct inode *inode; - inode = (struct inode *)malloc(sizeof(struct ubifs_inode)); + inode = (struct inode *)malloc_cache_aligned( + sizeof(struct ubifs_inode)); if (inode) { inode->i_ino = ino; inode->i_sb = sb; @@ -138,6 +71,10 @@ struct inode *iget_locked(struct super_block *sb, unsigned long ino) return inode; } +void iget_failed(struct inode *inode) +{ +} + int ubifs_iput(struct inode *inode) { list_del_init(&inode->i_sb_list); @@ -163,14 +100,14 @@ void iput(struct inode *inode) } if (i >= INODE_LOCKED_MAX) { - ubifs_err("Error, can't lock (save) more inodes while recovery!!!"); + dbg_gen("Error, can't lock (save) more inodes while recovery!!!"); return; } /* * Allocate and use new inode */ - ino = (struct inode *)malloc(sizeof(struct ubifs_inode)); + ino = (struct inode *)malloc_cache_aligned(sizeof(struct ubifs_inode)); memcpy(ino, inode, sizeof(struct ubifs_inode)); /* @@ -179,6 +116,125 @@ void iput(struct inode *inode) inodes_locked_down[i] = ino; } +/* from fs/inode.c */ +/** + * clear_nlink - directly zero an inode's link count + * @inode: inode + * + * This is a low-level filesystem helper to replace any + * direct filesystem manipulation of i_nlink. See + * drop_nlink() for why we care about i_nlink hitting zero. + */ +void clear_nlink(struct inode *inode) +{ + if (inode->i_nlink) { + inode->__i_nlink = 0; + atomic_long_inc(&inode->i_sb->s_remove_count); + } +} +EXPORT_SYMBOL(clear_nlink); + +/** + * set_nlink - directly set an inode's link count + * @inode: inode + * @nlink: new nlink (should be non-zero) + * + * This is a low-level filesystem helper to replace any + * direct filesystem manipulation of i_nlink. + */ +void set_nlink(struct inode *inode, unsigned int nlink) +{ + if (!nlink) { + clear_nlink(inode); + } else { + /* Yes, some filesystems do change nlink from zero to one */ + if (inode->i_nlink == 0) + atomic_long_dec(&inode->i_sb->s_remove_count); + + inode->__i_nlink = nlink; + } +} +EXPORT_SYMBOL(set_nlink); + +/* from include/linux/fs.h */ +static inline void i_uid_write(struct inode *inode, uid_t uid) +{ + inode->i_uid.val = uid; +} + +static inline void i_gid_write(struct inode *inode, gid_t gid) +{ + inode->i_gid.val = gid; +} + +void unlock_new_inode(struct inode *inode) +{ + return; +} +#endif + +/* + * Maximum amount of memory we may 'kmalloc()' without worrying that we are + * allocating too much. + */ +#define UBIFS_KMALLOC_OK (128*1024) + +/* Slab cache for UBIFS inodes */ +struct kmem_cache *ubifs_inode_slab; + +#ifndef __UBOOT__ +/* UBIFS TNC shrinker description */ +static struct shrinker ubifs_shrinker_info = { + .scan_objects = ubifs_shrink_scan, + .count_objects = ubifs_shrink_count, + .seeks = DEFAULT_SEEKS, +}; +#endif + +/** + * validate_inode - validate inode. + * @c: UBIFS file-system description object + * @inode: the inode to validate + * + * This is a helper function for 'ubifs_iget()' which validates various fields + * of a newly built inode to make sure they contain sane values and prevent + * possible vulnerabilities. Returns zero if the inode is all right and + * a non-zero error code if not. + */ +static int validate_inode(struct ubifs_info *c, const struct inode *inode) +{ + int err; + const struct ubifs_inode *ui = ubifs_inode(inode); + + if (inode->i_size > c->max_inode_sz) { + ubifs_err(c, "inode is too large (%lld)", + (long long)inode->i_size); + return 1; + } + + if (ui->compr_type >= UBIFS_COMPR_TYPES_CNT) { + ubifs_err(c, "unknown compression type %d", ui->compr_type); + return 2; + } + + if (ui->xattr_names + ui->xattr_cnt > XATTR_LIST_MAX) + return 3; + + if (ui->data_len < 0 || ui->data_len > UBIFS_MAX_INO_DATA) + return 4; + + if (ui->xattr && !S_ISREG(inode->i_mode)) + return 5; + + if (!ubifs_compr_present(ui->compr_type)) { + ubifs_warn(c, "inode %lu uses '%s' compression, but it was not compiled in", + inode->i_ino, ubifs_compr_name(ui->compr_type)); + } + + err = dbg_check_dir(c, inode); + return err; +} + struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) { int err; @@ -187,10 +243,13 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) struct ubifs_info *c = sb->s_fs_info; struct inode *inode; struct ubifs_inode *ui; +#ifdef __UBOOT__ int i; +#endif dbg_gen("inode %lu", inum); +#ifdef __UBOOT__ /* * U-Boot special handling of locked down inodes via recovery * e.g. ubifs_recover_size() @@ -211,6 +270,7 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) return inodes_locked_down[i]; } } +#endif inode = iget_locked(sb, inum); if (!inode) @@ -232,9 +292,9 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) goto out_ino; inode->i_flags |= (S_NOCMTIME | S_NOATIME); - inode->i_nlink = le32_to_cpu(ino->nlink); - inode->i_uid = le32_to_cpu(ino->uid); - inode->i_gid = le32_to_cpu(ino->gid); + set_nlink(inode, le32_to_cpu(ino->nlink)); + i_uid_write(inode, le32_to_cpu(ino->uid)); + i_gid_write(inode, le32_to_cpu(ino->gid)); inode->i_atime.tv_sec = (int64_t)le64_to_cpu(ino->atime_sec); inode->i_atime.tv_nsec = le32_to_cpu(ino->atime_nsec); inode->i_mtime.tv_sec = (int64_t)le64_to_cpu(ino->mtime_sec); @@ -248,12 +308,99 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) ui->flags = le32_to_cpu(ino->flags); ui->compr_type = le16_to_cpu(ino->compr_type); ui->creat_sqnum = le64_to_cpu(ino->creat_sqnum); + ui->xattr_cnt = le32_to_cpu(ino->xattr_cnt); + ui->xattr_size = le32_to_cpu(ino->xattr_size); + ui->xattr_names = le32_to_cpu(ino->xattr_names); ui->synced_i_size = ui->ui_size = inode->i_size; + ui->xattr = (ui->flags & UBIFS_XATTR_FL) ? 1 : 0; + err = validate_inode(c, inode); if (err) goto out_invalid; +#ifndef __UBOOT__ + switch (inode->i_mode & S_IFMT) { + case S_IFREG: + inode->i_mapping->a_ops = &ubifs_file_address_operations; + inode->i_op = &ubifs_file_inode_operations; + inode->i_fop = &ubifs_file_operations; + if (ui->xattr) { + ui->data = kmalloc(ui->data_len + 1, GFP_NOFS); + if (!ui->data) { + err = -ENOMEM; + goto out_ino; + } + memcpy(ui->data, ino->data, ui->data_len); + ((char *)ui->data)[ui->data_len] = '\0'; + } else if (ui->data_len != 0) { + err = 10; + goto out_invalid; + } + break; + case S_IFDIR: + inode->i_op = &ubifs_dir_inode_operations; + inode->i_fop = &ubifs_dir_operations; + if (ui->data_len != 0) { + err = 11; + goto out_invalid; + } + break; + case S_IFLNK: + inode->i_op = &ubifs_symlink_inode_operations; + if (ui->data_len <= 0 || ui->data_len > UBIFS_MAX_INO_DATA) { + err = 12; + goto out_invalid; + } + ui->data = kmalloc(ui->data_len + 1, GFP_NOFS); + if (!ui->data) { + err = -ENOMEM; + goto out_ino; + } + memcpy(ui->data, ino->data, ui->data_len); + ((char *)ui->data)[ui->data_len] = '\0'; + inode->i_link = ui->data; + break; + case S_IFBLK: + case S_IFCHR: + { + dev_t rdev; + union ubifs_dev_desc *dev; + + ui->data = kmalloc(sizeof(union ubifs_dev_desc), GFP_NOFS); + if (!ui->data) { + err = -ENOMEM; + goto out_ino; + } + + dev = (union ubifs_dev_desc *)ino->data; + if (ui->data_len == sizeof(dev->new)) + rdev = new_decode_dev(le32_to_cpu(dev->new)); + else if (ui->data_len == sizeof(dev->huge)) + rdev = huge_decode_dev(le64_to_cpu(dev->huge)); + else { + err = 13; + goto out_invalid; + } + memcpy(ui->data, ino->data, ui->data_len); + inode->i_op = &ubifs_file_inode_operations; + init_special_inode(inode, inode->i_mode, rdev); + break; + } + case S_IFSOCK: + case S_IFIFO: + inode->i_op = &ubifs_file_inode_operations; + init_special_inode(inode, inode->i_mode, 0); + if (ui->data_len != 0) { + err = 14; + goto out_invalid; + } + break; + default: + err = 15; + goto out_invalid; + } +#else if ((inode->i_mode & S_IFMT) == S_IFLNK) { if (ui->data_len <= 0 || ui->data_len > UBIFS_MAX_INO_DATA) { err = 12; @@ -267,113 +414,367 @@ struct inode *ubifs_iget(struct super_block *sb, unsigned long inum) memcpy(ui->data, ino->data, ui->data_len); ((char *)ui->data)[ui->data_len] = '\0'; } +#endif kfree(ino); - inode->i_state &= ~(I_LOCK | I_NEW); +#ifndef __UBOOT__ + ubifs_set_inode_flags(inode); +#endif + unlock_new_inode(inode); return inode; out_invalid: - ubifs_err("inode %lu validation failed, error %d", inode->i_ino, err); - dbg_dump_node(c, ino); - dbg_dump_inode(c, inode); + ubifs_err(c, "inode %lu validation failed, error %d", inode->i_ino, err); + ubifs_dump_node(c, ino); + ubifs_dump_inode(c, inode); err = -EINVAL; out_ino: kfree(ino); out: - ubifs_err("failed to read inode %lu, error %d", inode->i_ino, err); + ubifs_err(c, "failed to read inode %lu, error %d", inode->i_ino, err); + iget_failed(inode); return ERR_PTR(err); } -/** - * init_constants_early - initialize UBIFS constants. - * @c: UBIFS file-system description object - * - * This function initialize UBIFS constants which do not need the superblock to - * be read. It also checks that the UBI volume satisfies basic UBIFS - * requirements. Returns zero in case of success and a negative error code in - * case of failure. - */ -static int init_constants_early(struct ubifs_info *c) +static struct inode *ubifs_alloc_inode(struct super_block *sb) { - if (c->vi.corrupted) { - ubifs_warn("UBI volume is corrupted - read-only mode"); - c->ro_media = 1; - } + struct ubifs_inode *ui; - if (c->di.ro_mode) { - ubifs_msg("read-only UBI device"); - c->ro_media = 1; - } + ui = kmem_cache_alloc(ubifs_inode_slab, GFP_NOFS); + if (!ui) + return NULL; - if (c->vi.vol_type == UBI_STATIC_VOLUME) { - ubifs_msg("static UBI volume - read-only mode"); - c->ro_media = 1; - } + memset((void *)ui + sizeof(struct inode), 0, + sizeof(struct ubifs_inode) - sizeof(struct inode)); + mutex_init(&ui->ui_mutex); + spin_lock_init(&ui->ui_lock); + return &ui->vfs_inode; +}; - c->leb_cnt = c->vi.size; - c->leb_size = c->vi.usable_leb_size; - c->half_leb_size = c->leb_size / 2; - c->min_io_size = c->di.min_io_size; - c->min_io_shift = fls(c->min_io_size) - 1; +#ifndef __UBOOT__ +static void ubifs_i_callback(struct rcu_head *head) +{ + struct inode *inode = container_of(head, struct inode, i_rcu); + struct ubifs_inode *ui = ubifs_inode(inode); + kmem_cache_free(ubifs_inode_slab, ui); +} - if (c->leb_size < UBIFS_MIN_LEB_SZ) { - ubifs_err("too small LEBs (%d bytes), min. is %d bytes", - c->leb_size, UBIFS_MIN_LEB_SZ); - return -EINVAL; - } +static void ubifs_destroy_inode(struct inode *inode) +{ + struct ubifs_inode *ui = ubifs_inode(inode); - if (c->leb_cnt < UBIFS_MIN_LEB_CNT) { - ubifs_err("too few LEBs (%d), min. is %d", - c->leb_cnt, UBIFS_MIN_LEB_CNT); - return -EINVAL; - } + kfree(ui->data); + call_rcu(&inode->i_rcu, ubifs_i_callback); +} - if (!is_power_of_2(c->min_io_size)) { - ubifs_err("bad min. I/O size %d", c->min_io_size); - return -EINVAL; - } +/* + * Note, Linux write-back code calls this without 'i_mutex'. + */ +static int ubifs_write_inode(struct inode *inode, struct writeback_control *wbc) +{ + int err = 0; + struct ubifs_info *c = inode->i_sb->s_fs_info; + struct ubifs_inode *ui = ubifs_inode(inode); + + ubifs_assert(!ui->xattr); + if (is_bad_inode(inode)) + return 0; + mutex_lock(&ui->ui_mutex); /* - * UBIFS aligns all node to 8-byte boundary, so to make function in - * io.c simpler, assume minimum I/O unit size to be 8 bytes if it is - * less than 8. + * Due to races between write-back forced by budgeting + * (see 'sync_some_inodes()') and background write-back, the inode may + * have already been synchronized, do not do this again. This might + * also happen if it was synchronized in an VFS operation, e.g. + * 'ubifs_link()'. */ - if (c->min_io_size < 8) { - c->min_io_size = 8; - c->min_io_shift = 3; + if (!ui->dirty) { + mutex_unlock(&ui->ui_mutex); + return 0; } - c->ref_node_alsz = ALIGN(UBIFS_REF_NODE_SZ, c->min_io_size); - c->mst_node_alsz = ALIGN(UBIFS_MST_NODE_SZ, c->min_io_size); - /* - * Initialize node length ranges which are mostly needed for node - * length validation. + * As an optimization, do not write orphan inodes to the media just + * because this is not needed. */ - c->ranges[UBIFS_PAD_NODE].len = UBIFS_PAD_NODE_SZ; - c->ranges[UBIFS_SB_NODE].len = UBIFS_SB_NODE_SZ; - c->ranges[UBIFS_MST_NODE].len = UBIFS_MST_NODE_SZ; - c->ranges[UBIFS_REF_NODE].len = UBIFS_REF_NODE_SZ; - c->ranges[UBIFS_TRUN_NODE].len = UBIFS_TRUN_NODE_SZ; - c->ranges[UBIFS_CS_NODE].len = UBIFS_CS_NODE_SZ; + dbg_gen("inode %lu, mode %#x, nlink %u", + inode->i_ino, (int)inode->i_mode, inode->i_nlink); + if (inode->i_nlink) { + err = ubifs_jnl_write_inode(c, inode); + if (err) + ubifs_err(c, "can't write inode %lu, error %d", + inode->i_ino, err); + else + err = dbg_check_inode_size(c, inode, ui->ui_size); + } - c->ranges[UBIFS_INO_NODE].min_len = UBIFS_INO_NODE_SZ; - c->ranges[UBIFS_INO_NODE].max_len = UBIFS_MAX_INO_NODE_SZ; - c->ranges[UBIFS_ORPH_NODE].min_len = - UBIFS_ORPH_NODE_SZ + sizeof(__le64); - c->ranges[UBIFS_ORPH_NODE].max_len = c->leb_size; - c->ranges[UBIFS_DENT_NODE].min_len = UBIFS_DENT_NODE_SZ; - c->ranges[UBIFS_DENT_NODE].max_len = UBIFS_MAX_DENT_NODE_SZ; - c->ranges[UBIFS_XENT_NODE].min_len = UBIFS_XENT_NODE_SZ; - c->ranges[UBIFS_XENT_NODE].max_len = UBIFS_MAX_XENT_NODE_SZ; - c->ranges[UBIFS_DATA_NODE].min_len = UBIFS_DATA_NODE_SZ; - c->ranges[UBIFS_DATA_NODE].max_len = UBIFS_MAX_DATA_NODE_SZ; - /* - * Minimum indexing node size is amended later when superblock is - * read and the key length is known. - */ - c->ranges[UBIFS_IDX_NODE].min_len = UBIFS_IDX_NODE_SZ + UBIFS_BRANCH_SZ; - /* + ui->dirty = 0; + mutex_unlock(&ui->ui_mutex); + ubifs_release_dirty_inode_budget(c, ui); + return err; +} + +static void ubifs_evict_inode(struct inode *inode) +{ + int err; + struct ubifs_info *c = inode->i_sb->s_fs_info; + struct ubifs_inode *ui = ubifs_inode(inode); + + if (ui->xattr) + /* + * Extended attribute inode deletions are fully handled in + * 'ubifs_removexattr()'. These inodes are special and have + * limited usage, so there is nothing to do here. + */ + goto out; + + dbg_gen("inode %lu, mode %#x", inode->i_ino, (int)inode->i_mode); + ubifs_assert(!atomic_read(&inode->i_count)); + + truncate_inode_pages_final(&inode->i_data); + + if (inode->i_nlink) + goto done; + + if (is_bad_inode(inode)) + goto out; + + ui->ui_size = inode->i_size = 0; + err = ubifs_jnl_delete_inode(c, inode); + if (err) + /* + * Worst case we have a lost orphan inode wasting space, so a + * simple error message is OK here. + */ + ubifs_err(c, "can't delete inode %lu, error %d", + inode->i_ino, err); + +out: + if (ui->dirty) + ubifs_release_dirty_inode_budget(c, ui); + else { + /* We've deleted something - clean the "no space" flags */ + c->bi.nospace = c->bi.nospace_rp = 0; + smp_wmb(); + } +done: + clear_inode(inode); +} +#endif + +static void ubifs_dirty_inode(struct inode *inode, int flags) +{ + struct ubifs_inode *ui = ubifs_inode(inode); + + ubifs_assert(mutex_is_locked(&ui->ui_mutex)); + if (!ui->dirty) { + ui->dirty = 1; + dbg_gen("inode %lu", inode->i_ino); + } +} + +#ifndef __UBOOT__ +static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf) +{ + struct ubifs_info *c = dentry->d_sb->s_fs_info; + unsigned long long free; + __le32 *uuid = (__le32 *)c->uuid; + + free = ubifs_get_free_space(c); + dbg_gen("free space %lld bytes (%lld blocks)", + free, free >> UBIFS_BLOCK_SHIFT); + + buf->f_type = UBIFS_SUPER_MAGIC; + buf->f_bsize = UBIFS_BLOCK_SIZE; + buf->f_blocks = c->block_cnt; + buf->f_bfree = free >> UBIFS_BLOCK_SHIFT; + if (free > c->report_rp_size) + buf->f_bavail = (free - c->report_rp_size) >> UBIFS_BLOCK_SHIFT; + else + buf->f_bavail = 0; + buf->f_files = 0; + buf->f_ffree = 0; + buf->f_namelen = UBIFS_MAX_NLEN; + buf->f_fsid.val[0] = le32_to_cpu(uuid[0]) ^ le32_to_cpu(uuid[2]); + buf->f_fsid.val[1] = le32_to_cpu(uuid[1]) ^ le32_to_cpu(uuid[3]); + ubifs_assert(buf->f_bfree <= c->block_cnt); + return 0; +} + +static int ubifs_show_options(struct seq_file *s, struct dentry *root) +{ + struct ubifs_info *c = root->d_sb->s_fs_info; + + if (c->mount_opts.unmount_mode == 2) + seq_puts(s, ",fast_unmount"); + else if (c->mount_opts.unmount_mode == 1) + seq_puts(s, ",norm_unmount"); + + if (c->mount_opts.bulk_read == 2) + seq_puts(s, ",bulk_read"); + else if (c->mount_opts.bulk_read == 1) + seq_puts(s, ",no_bulk_read"); + + if (c->mount_opts.chk_data_crc == 2) + seq_puts(s, ",chk_data_crc"); + else if (c->mount_opts.chk_data_crc == 1) + seq_puts(s, ",no_chk_data_crc"); + + if (c->mount_opts.override_compr) { + seq_printf(s, ",compr=%s", + ubifs_compr_name(c->mount_opts.compr_type)); + } + + return 0; +} + +static int ubifs_sync_fs(struct super_block *sb, int wait) +{ + int i, err; + struct ubifs_info *c = sb->s_fs_info; + + /* + * Zero @wait is just an advisory thing to help the file system shove + * lots of data into the queues, and there will be the second + * '->sync_fs()' call, with non-zero @wait. + */ + if (!wait) + return 0; + + /* + * Synchronize write buffers, because 'ubifs_run_commit()' does not + * do this if it waits for an already running commit. + */ + for (i = 0; i < c->jhead_cnt; i++) { + err = ubifs_wbuf_sync(&c->jheads[i].wbuf); + if (err) + return err; + } + + /* + * Strictly speaking, it is not necessary to commit the journal here, + * synchronizing write-buffers would be enough. But committing makes + * UBIFS free space predictions much more accurate, so we want to let + * the user be able to get more accurate results of 'statfs()' after + * they synchronize the file system. + */ + err = ubifs_run_commit(c); + if (err) + return err; + + return ubi_sync(c->vi.ubi_num); +} +#endif + +/** + * init_constants_early - initialize UBIFS constants. + * @c: UBIFS file-system description object + * + * This function initialize UBIFS constants which do not need the superblock to + * be read. It also checks that the UBI volume satisfies basic UBIFS + * requirements. Returns zero in case of success and a negative error code in + * case of failure. + */ +static int init_constants_early(struct ubifs_info *c) +{ + if (c->vi.corrupted) { + ubifs_warn(c, "UBI volume is corrupted - read-only mode"); + c->ro_media = 1; + } + + if (c->di.ro_mode) { + ubifs_msg(c, "read-only UBI device"); + c->ro_media = 1; + } + + if (c->vi.vol_type == UBI_STATIC_VOLUME) { + ubifs_msg(c, "static UBI volume - read-only mode"); + c->ro_media = 1; + } + + c->leb_cnt = c->vi.size; + c->leb_size = c->vi.usable_leb_size; + c->leb_start = c->di.leb_start; + c->half_leb_size = c->leb_size / 2; + c->min_io_size = c->di.min_io_size; + c->min_io_shift = fls(c->min_io_size) - 1; + c->max_write_size = c->di.max_write_size; + c->max_write_shift = fls(c->max_write_size) - 1; + + if (c->leb_size < UBIFS_MIN_LEB_SZ) { + ubifs_err(c, "too small LEBs (%d bytes), min. is %d bytes", + c->leb_size, UBIFS_MIN_LEB_SZ); + return -EINVAL; + } + + if (c->leb_cnt < UBIFS_MIN_LEB_CNT) { + ubifs_err(c, "too few LEBs (%d), min. is %d", + c->leb_cnt, UBIFS_MIN_LEB_CNT); + return -EINVAL; + } + + if (!is_power_of_2(c->min_io_size)) { + ubifs_err(c, "bad min. I/O size %d", c->min_io_size); + return -EINVAL; + } + + /* + * Maximum write size has to be greater or equivalent to min. I/O + * size, and be multiple of min. I/O size. + */ + if (c->max_write_size < c->min_io_size || + c->max_write_size % c->min_io_size || + !is_power_of_2(c->max_write_size)) { + ubifs_err(c, "bad write buffer size %d for %d min. I/O unit", + c->max_write_size, c->min_io_size); + return -EINVAL; + } + + /* + * UBIFS aligns all node to 8-byte boundary, so to make function in + * io.c simpler, assume minimum I/O unit size to be 8 bytes if it is + * less than 8. + */ + if (c->min_io_size < 8) { + c->min_io_size = 8; + c->min_io_shift = 3; + if (c->max_write_size < c->min_io_size) { + c->max_write_size = c->min_io_size; + c->max_write_shift = c->min_io_shift; + } + } + + c->ref_node_alsz = ALIGN(UBIFS_REF_NODE_SZ, c->min_io_size); + c->mst_node_alsz = ALIGN(UBIFS_MST_NODE_SZ, c->min_io_size); + + /* + * Initialize node length ranges which are mostly needed for node + * length validation. + */ + c->ranges[UBIFS_PAD_NODE].len = UBIFS_PAD_NODE_SZ; + c->ranges[UBIFS_SB_NODE].len = UBIFS_SB_NODE_SZ; + c->ranges[UBIFS_MST_NODE].len = UBIFS_MST_NODE_SZ; + c->ranges[UBIFS_REF_NODE].len = UBIFS_REF_NODE_SZ; + c->ranges[UBIFS_TRUN_NODE].len = UBIFS_TRUN_NODE_SZ; + c->ranges[UBIFS_CS_NODE].len = UBIFS_CS_NODE_SZ; + + c->ranges[UBIFS_INO_NODE].min_len = UBIFS_INO_NODE_SZ; + c->ranges[UBIFS_INO_NODE].max_len = UBIFS_MAX_INO_NODE_SZ; + c->ranges[UBIFS_ORPH_NODE].min_len = + UBIFS_ORPH_NODE_SZ + sizeof(__le64); + c->ranges[UBIFS_ORPH_NODE].max_len = c->leb_size; + c->ranges[UBIFS_DENT_NODE].min_len = UBIFS_DENT_NODE_SZ; + c->ranges[UBIFS_DENT_NODE].max_len = UBIFS_MAX_DENT_NODE_SZ; + c->ranges[UBIFS_XENT_NODE].min_len = UBIFS_XENT_NODE_SZ; + c->ranges[UBIFS_XENT_NODE].max_len = UBIFS_MAX_XENT_NODE_SZ; + c->ranges[UBIFS_DATA_NODE].min_len = UBIFS_DATA_NODE_SZ; + c->ranges[UBIFS_DATA_NODE].max_len = UBIFS_MAX_DATA_NODE_SZ; + /* + * Minimum indexing node size is amended later when superblock is + * read and the key length is known. + */ + c->ranges[UBIFS_IDX_NODE].min_len = UBIFS_IDX_NODE_SZ + UBIFS_BRANCH_SZ; + /* * Maximum indexing node size is amended later when superblock is * read and the fanout is known. */ @@ -393,9 +794,33 @@ static int init_constants_early(struct ubifs_info *c) */ c->leb_overhead = c->leb_size % UBIFS_MAX_DATA_NODE_SZ; + /* Buffer size for bulk-reads */ + c->max_bu_buf_len = UBIFS_MAX_BULK_READ * UBIFS_MAX_DATA_NODE_SZ; + if (c->max_bu_buf_len > c->leb_size) + c->max_bu_buf_len = c->leb_size; return 0; } +/** + * bud_wbuf_callback - bud LEB write-buffer synchronization call-back. + * @c: UBIFS file-system description object + * @lnum: LEB the write-buffer was synchronized to + * @free: how many free bytes left in this LEB + * @pad: how many bytes were padded + * + * This is a callback function which is called by the I/O unit when the + * write-buffer is synchronized. We need this to correctly maintain space + * accounting in bud logical eraseblocks. This function returns zero in case of + * success and a negative error code in case of failure. + * + * This function actually belongs to the journal, but we keep it here because + * we want to keep it static. + */ +static int bud_wbuf_callback(struct ubifs_info *c, int lnum, int free, int pad) +{ + return ubifs_update_one_lp(c, lnum, free, pad, 0, 0); +} + /* * init_constants_sb - initialize UBIFS constants. * @c: UBIFS file-system description object @@ -426,8 +851,8 @@ static int init_constants_sb(struct ubifs_info *c) tmp = UBIFS_CS_NODE_SZ + UBIFS_REF_NODE_SZ * c->jhead_cnt; tmp = ALIGN(tmp, c->min_io_size); if (tmp > c->leb_size) { - dbg_err("too small LEB size %d, at least %d needed", - c->leb_size, tmp); + ubifs_err(c, "too small LEB size %d, at least %d needed", + c->leb_size, tmp); return -EINVAL; } @@ -441,8 +866,8 @@ static int init_constants_sb(struct ubifs_info *c) tmp /= c->leb_size; tmp += 1; if (c->log_lebs < tmp) { - dbg_err("too small log %d LEBs, required min. %d LEBs", - c->log_lebs, tmp); + ubifs_err(c, "too small log %d LEBs, required min. %d LEBs", + c->log_lebs, tmp); return -EINVAL; } @@ -451,11 +876,11 @@ static int init_constants_sb(struct ubifs_info *c) * be compressed and direntries are of the maximum size. * * Note, data, which may be stored in inodes is budgeted separately, so - * it is not included into 'c->inode_budget'. + * it is not included into 'c->bi.inode_budget'. */ - c->page_budget = UBIFS_MAX_DATA_NODE_SZ * UBIFS_BLOCKS_PER_PAGE; - c->inode_budget = UBIFS_INO_NODE_SZ; - c->dent_budget = UBIFS_MAX_DENT_NODE_SZ; + c->bi.page_budget = UBIFS_MAX_DATA_NODE_SZ * UBIFS_BLOCKS_PER_PAGE; + c->bi.inode_budget = UBIFS_INO_NODE_SZ; + c->bi.dent_budget = UBIFS_MAX_DENT_NODE_SZ; /* * When the amount of flash space used by buds becomes @@ -482,6 +907,8 @@ static int init_constants_sb(struct ubifs_info *c) if (err) return err; + /* Initialize effective LEB size used in budgeting calculations */ + c->idx_leb_size = c->leb_size - c->max_idx_node_sz; return 0; } @@ -497,7 +924,8 @@ static void init_constants_master(struct ubifs_info *c) { long long tmp64; - c->min_idx_lebs = ubifs_calc_min_idx_lebs(c); + c->bi.min_idx_lebs = ubifs_calc_min_idx_lebs(c); + c->report_rp_size = ubifs_reported_space(c, c->rp_size); /* * Calculate total amount of FS blocks. This number is not used @@ -514,6 +942,88 @@ static void init_constants_master(struct ubifs_info *c) c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT; } +/** + * take_gc_lnum - reserve GC LEB. + * @c: UBIFS file-system description object + * + * This function ensures that the LEB reserved for garbage collection is marked + * as "taken" in lprops. We also have to set free space to LEB size and dirty + * space to zero, because lprops may contain out-of-date information if the + * file-system was un-mounted before it has been committed. This function + * returns zero in case of success and a negative error code in case of + * failure. + */ +static int take_gc_lnum(struct ubifs_info *c) +{ + int err; + + if (c->gc_lnum == -1) { + ubifs_err(c, "no LEB for GC"); + return -EINVAL; + } + + /* And we have to tell lprops that this LEB is taken */ + err = ubifs_change_one_lp(c, c->gc_lnum, c->leb_size, 0, + LPROPS_TAKEN, 0, 0); + return err; +} + +/** + * alloc_wbufs - allocate write-buffers. + * @c: UBIFS file-system description object + * + * This helper function allocates and initializes UBIFS write-buffers. Returns + * zero in case of success and %-ENOMEM in case of failure. + */ +static int alloc_wbufs(struct ubifs_info *c) +{ + int i, err; + + c->jheads = kcalloc(c->jhead_cnt, sizeof(struct ubifs_jhead), + GFP_KERNEL); + if (!c->jheads) + return -ENOMEM; + + /* Initialize journal heads */ + for (i = 0; i < c->jhead_cnt; i++) { + INIT_LIST_HEAD(&c->jheads[i].buds_list); + err = ubifs_wbuf_init(c, &c->jheads[i].wbuf); + if (err) + return err; + + c->jheads[i].wbuf.sync_callback = &bud_wbuf_callback; + c->jheads[i].wbuf.jhead = i; + c->jheads[i].grouped = 1; + } + + /* + * Garbage Collector head does not need to be synchronized by timer. + * Also GC head nodes are not grouped. + */ + c->jheads[GCHD].wbuf.no_timer = 1; + c->jheads[GCHD].grouped = 0; + + return 0; +} + +/** + * free_wbufs - free write-buffers. + * @c: UBIFS file-system description object + */ +static void free_wbufs(struct ubifs_info *c) +{ + int i; + + if (c->jheads) { + for (i = 0; i < c->jhead_cnt; i++) { + kfree(c->jheads[i].wbuf.buf); + kfree(c->jheads[i].wbuf.inodes); + } + kfree(c->jheads); + c->jheads = NULL; + } +} + /** * free_orphans - free orphans. * @c: UBIFS file-system description object @@ -533,13 +1043,25 @@ static void free_orphans(struct ubifs_info *c) orph = list_entry(c->orph_list.next, struct ubifs_orphan, list); list_del(&orph->list); kfree(orph); - dbg_err("orphan list not empty at unmount"); + ubifs_err(c, "orphan list not empty at unmount"); } vfree(c->orph_buf); c->orph_buf = NULL; } +/** + * free_buds - free per-bud objects. + * @c: UBIFS file-system description object + */ +static void free_buds(struct ubifs_info *c) +{ + struct ubifs_bud *bud, *n; + + rbtree_postorder_for_each_entry_safe(bud, n, &c->buds, rb) + kfree(bud); +} + /** * check_volume_empty - check if the UBI volume is empty. * @c: UBIFS file-system description object @@ -555,7 +1077,7 @@ static int check_volume_empty(struct ubifs_info *c) c->empty = 1; for (lnum = 0; lnum < c->leb_cnt; lnum++) { - err = ubi_is_mapped(c->ubi, lnum); + err = ubifs_is_mapped(c, lnum); if (unlikely(err < 0)) return err; if (err == 1) { @@ -569,51 +1091,292 @@ static int check_volume_empty(struct ubifs_info *c) return 0; } +/* + * UBIFS mount options. + * + * Opt_fast_unmount: do not run a journal commit before un-mounting + * Opt_norm_unmount: run a journal commit before un-mounting + * Opt_bulk_read: enable bulk-reads + * Opt_no_bulk_read: disable bulk-reads + * Opt_chk_data_crc: check CRCs when reading data nodes + * Opt_no_chk_data_crc: do not check CRCs when reading data nodes + * Opt_override_compr: override default compressor + * Opt_err: just end of array marker + */ +enum { + Opt_fast_unmount, + Opt_norm_unmount, + Opt_bulk_read, + Opt_no_bulk_read, + Opt_chk_data_crc, + Opt_no_chk_data_crc, + Opt_override_compr, + Opt_err, +}; + +#ifndef __UBOOT__ +static const match_table_t tokens = { + {Opt_fast_unmount, "fast_unmount"}, + {Opt_norm_unmount, "norm_unmount"}, + {Opt_bulk_read, "bulk_read"}, + {Opt_no_bulk_read, "no_bulk_read"}, + {Opt_chk_data_crc, "chk_data_crc"}, + {Opt_no_chk_data_crc, "no_chk_data_crc"}, + {Opt_override_compr, "compr=%s"}, + {Opt_err, NULL}, +}; + /** - * mount_ubifs - mount UBIFS file-system. - * @c: UBIFS file-system description object + * parse_standard_option - parse a standard mount option. + * @option: the option to parse * - * This function mounts UBIFS file system. Returns zero in case of success and - * a negative error code in case of failure. + * Normally, standard mount options like "sync" are passed to file-systems as + * flags. However, when a "rootflags=" kernel boot parameter is used, they may + * be present in the options string. This function tries to deal with this + * situation and parse standard options. Returns 0 if the option was not + * recognized, and the corresponding integer flag if it was. * - * Note, the function does not de-allocate resources it it fails half way - * through, and the caller has to do this instead. + * UBIFS is only interested in the "sync" option, so do not check for anything + * else. */ -static int mount_ubifs(struct ubifs_info *c) +static int parse_standard_option(const char *option) { - struct super_block *sb = c->vfs_sb; - int err, mounted_read_only = (sb->s_flags & MS_RDONLY); - long long x; - size_t sz; - err = init_constants_early(c); - if (err) - return err; + pr_notice("UBIFS: parse %s\n", option); + if (!strcmp(option, "sync")) + return MS_SYNCHRONOUS; + return 0; +} - err = ubifs_debugging_init(c); - if (err) - return err; +/** + * ubifs_parse_options - parse mount parameters. + * @c: UBIFS file-system description object + * @options: parameters to parse + * @is_remount: non-zero if this is FS re-mount + * + * This function parses UBIFS mount options and returns zero in case success + * and a negative error code in case of failure. + */ +static int ubifs_parse_options(struct ubifs_info *c, char *options, + int is_remount) +{ + char *p; + substring_t args[MAX_OPT_ARGS]; - err = check_volume_empty(c); - if (err) - goto out_free; + if (!options) + return 0; + + while ((p = strsep(&options, ","))) { + int token; - if (c->empty && (mounted_read_only || c->ro_media)) { + if (!*p) + continue; + + token = match_token(p, tokens, args); + switch (token) { /* - * This UBI volume is empty, and read-only, or the file system - * is mounted read-only - we cannot format it. + * %Opt_fast_unmount and %Opt_norm_unmount options are ignored. + * We accept them in order to be backward-compatible. But this + * should be removed at some point. */ - ubifs_err("can't format empty UBI volume: read-only %s", - c->ro_media ? "UBI volume" : "mount"); - err = -EROFS; - goto out_free; + case Opt_fast_unmount: + c->mount_opts.unmount_mode = 2; + break; + case Opt_norm_unmount: + c->mount_opts.unmount_mode = 1; + break; + case Opt_bulk_read: + c->mount_opts.bulk_read = 2; + c->bulk_read = 1; + break; + case Opt_no_bulk_read: + c->mount_opts.bulk_read = 1; + c->bulk_read = 0; + break; + case Opt_chk_data_crc: + c->mount_opts.chk_data_crc = 2; + c->no_chk_data_crc = 0; + break; + case Opt_no_chk_data_crc: + c->mount_opts.chk_data_crc = 1; + c->no_chk_data_crc = 1; + break; + case Opt_override_compr: + { + char *name = match_strdup(&args[0]); + + if (!name) + return -ENOMEM; + if (!strcmp(name, "none")) + c->mount_opts.compr_type = UBIFS_COMPR_NONE; + else if (!strcmp(name, "lzo")) + c->mount_opts.compr_type = UBIFS_COMPR_LZO; + else if (!strcmp(name, "zlib")) + c->mount_opts.compr_type = UBIFS_COMPR_ZLIB; + else { + ubifs_err(c, "unknown compressor \"%s\"", name); //FIXME: is c ready? + kfree(name); + return -EINVAL; + } + kfree(name); + c->mount_opts.override_compr = 1; + c->default_compr = c->mount_opts.compr_type; + break; + } + default: + { + unsigned long flag; + struct super_block *sb = c->vfs_sb; + + flag = parse_standard_option(p); + if (!flag) { + ubifs_err(c, "unrecognized mount option \"%s\" or missing value", + p); + return -EINVAL; + } + sb->s_flags |= flag; + break; + } + } } - if (c->ro_media && !mounted_read_only) { - ubifs_err("cannot mount read-write - read-only media"); - err = -EROFS; - goto out_free; - } + return 0; +} +#endif + +/** + * destroy_journal - destroy journal data structures. + * @c: UBIFS file-system description object + * + * This function destroys journal data structures including those that may have + * been created by recovery functions. + */ +static void destroy_journal(struct ubifs_info *c) +{ + while (!list_empty(&c->unclean_leb_list)) { + struct ubifs_unclean_leb *ucleb; + + ucleb = list_entry(c->unclean_leb_list.next, + struct ubifs_unclean_leb, list); + list_del(&ucleb->list); + kfree(ucleb); + } + while (!list_empty(&c->old_buds)) { + struct ubifs_bud *bud; + + bud = list_entry(c->old_buds.next, struct ubifs_bud, list); + list_del(&bud->list); + kfree(bud); + } + ubifs_destroy_idx_gc(c); + ubifs_destroy_size_tree(c); + ubifs_tnc_close(c); + free_buds(c); +} + +/** + * bu_init - initialize bulk-read information. + * @c: UBIFS file-system description object + */ +static void bu_init(struct ubifs_info *c) +{ + ubifs_assert(c->bulk_read == 1); + + if (c->bu.buf) + return; /* Already initialized */ + +again: + c->bu.buf = kmalloc(c->max_bu_buf_len, GFP_KERNEL | __GFP_NOWARN); + if (!c->bu.buf) { + if (c->max_bu_buf_len > UBIFS_KMALLOC_OK) { + c->max_bu_buf_len = UBIFS_KMALLOC_OK; + goto again; + } + + /* Just disable bulk-read */ + ubifs_warn(c, "cannot allocate %d bytes of memory for bulk-read, disabling it", + c->max_bu_buf_len); + c->mount_opts.bulk_read = 1; + c->bulk_read = 0; + return; + } +} + +#ifndef __UBOOT__ +/** + * check_free_space - check if there is enough free space to mount. + * @c: UBIFS file-system description object + * + * This function makes sure UBIFS has enough free space to be mounted in + * read/write mode. UBIFS must always have some free space to allow deletions. + */ +static int check_free_space(struct ubifs_info *c) +{ + ubifs_assert(c->dark_wm > 0); + if (c->lst.total_free + c->lst.total_dirty < c->dark_wm) { + ubifs_err(c, "insufficient free space to mount in R/W mode"); + ubifs_dump_budg(c, &c->bi); + ubifs_dump_lprops(c); + return -ENOSPC; + } + return 0; +} +#endif + +/** + * mount_ubifs - mount UBIFS file-system. + * @c: UBIFS file-system description object + * + * This function mounts UBIFS file system. Returns zero in case of success and + * a negative error code in case of failure. + */ +static int mount_ubifs(struct ubifs_info *c) +{ + int err; + long long x; +#ifndef CONFIG_UBIFS_SILENCE_MSG + long long y; +#endif + size_t sz; + + c->ro_mount = !!(c->vfs_sb->s_flags & MS_RDONLY); + /* Suppress error messages while probing if MS_SILENT is set */ + c->probing = !!(c->vfs_sb->s_flags & MS_SILENT); +#ifdef __UBOOT__ + if (!c->ro_mount) { + printf("UBIFS: only ro mode in U-Boot allowed.\n"); + return -EACCES; + } +#endif + + err = init_constants_early(c); + if (err) + return err; + + err = ubifs_debugging_init(c); + if (err) + return err; + + err = check_volume_empty(c); + if (err) + goto out_free; + + if (c->empty && (c->ro_mount || c->ro_media)) { + /* + * This UBI volume is empty, and read-only, or the file system + * is mounted read-only - we cannot format it. + */ + ubifs_err(c, "can't format empty UBI volume: read-only %s", + c->ro_media ? "UBI volume" : "mount"); + err = -EROFS; + goto out_free; + } + + if (c->ro_media && !c->ro_mount) { + ubifs_err(c, "cannot mount read-write - read-only media"); + err = -EROFS; + goto out_free; + } /* * The requirement for the buffer is that it should fit indexing B-tree @@ -629,28 +1392,45 @@ static int mount_ubifs(struct ubifs_info *c) if (!c->sbuf) goto out_free; - /* - * We have to check all CRCs, even for data nodes, when we mount the FS - * (specifically, when we are replaying). - */ - c->always_chk_crc = 1; +#ifndef __UBOOT__ + if (!c->ro_mount) { + c->ileb_buf = vmalloc(c->leb_size); + if (!c->ileb_buf) + goto out_free; + } +#endif + + if (c->bulk_read == 1) + bu_init(c); + +#ifndef __UBOOT__ + if (!c->ro_mount) { + c->write_reserve_buf = kmalloc(COMPRESSED_DATA_NODE_BUF_SZ, + GFP_KERNEL); + if (!c->write_reserve_buf) + goto out_free; + } +#endif + + c->mounting = 1; err = ubifs_read_superblock(c); if (err) goto out_free; + c->probing = 0; + /* * Make sure the compressor which is set as default in the superblock * or overridden by mount options is actually compiled in. */ if (!ubifs_compr_present(c->default_compr)) { - ubifs_err("'compressor \"%s\" is not compiled in", + ubifs_err(c, "'compressor \"%s\" is not compiled in", ubifs_compr_name(c->default_compr)); + err = -ENOTSUPP; goto out_free; } - dbg_failure_mode_registration(c); - err = init_constants_sb(c); if (err) goto out_free; @@ -663,7 +1443,25 @@ static int mount_ubifs(struct ubifs_info *c) goto out_free; } + err = alloc_wbufs(c); + if (err) + goto out_cbuf; + sprintf(c->bgt_name, BGT_NAME_PATTERN, c->vi.ubi_num, c->vi.vol_id); +#ifndef __UBOOT__ + if (!c->ro_mount) { + /* Create background thread */ + c->bgt = kthread_create(ubifs_bg_thread, c, "%s", c->bgt_name); + if (IS_ERR(c->bgt)) { + err = PTR_ERR(c->bgt); + c->bgt = NULL; + ubifs_err(c, "cannot spawn \"%s\", error %d", + c->bgt_name, err); + goto out_wbufs; + } + wake_up_process(c->bgt); + } +#endif err = ubifs_read_master(c); if (err) @@ -672,207 +1470,696 @@ static int mount_ubifs(struct ubifs_info *c) init_constants_master(c); if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) { - ubifs_msg("recovery needed"); + ubifs_msg(c, "recovery needed"); c->need_recovery = 1; } - err = ubifs_lpt_init(c, 1, !mounted_read_only); +#ifndef __UBOOT__ + if (c->need_recovery && !c->ro_mount) { + err = ubifs_recover_inl_heads(c, c->sbuf); + if (err) + goto out_master; + } +#endif + + err = ubifs_lpt_init(c, 1, !c->ro_mount); if (err) - goto out_lpt; + goto out_master; + +#ifndef __UBOOT__ + if (!c->ro_mount && c->space_fixup) { + err = ubifs_fixup_free_space(c); + if (err) + goto out_lpt; + } + + if (!c->ro_mount && !c->need_recovery) { + /* + * Set the "dirty" flag so that if we reboot uncleanly we + * will notice this immediately on the next mount. + */ + c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY); + err = ubifs_write_master(c); + if (err) + goto out_lpt; + } +#endif - err = dbg_check_idx_size(c, c->old_idx_sz); + err = dbg_check_idx_size(c, c->bi.old_idx_sz); if (err) goto out_lpt; err = ubifs_replay_journal(c); if (err) - goto out_journal; + goto out_journal; + + /* Calculate 'min_idx_lebs' after journal replay */ + c->bi.min_idx_lebs = ubifs_calc_min_idx_lebs(c); + + err = ubifs_mount_orphans(c, c->need_recovery, c->ro_mount); + if (err) + goto out_orphans; + + if (!c->ro_mount) { +#ifndef __UBOOT__ + int lnum; + + err = check_free_space(c); + if (err) + goto out_orphans; + + /* Check for enough log space */ + lnum = c->lhead_lnum + 1; + if (lnum >= UBIFS_LOG_LNUM + c->log_lebs) + lnum = UBIFS_LOG_LNUM; + if (lnum == c->ltail_lnum) { + err = ubifs_consolidate_log(c); + if (err) + goto out_orphans; + } + + if (c->need_recovery) { + err = ubifs_recover_size(c); + if (err) + goto out_orphans; + err = ubifs_rcvry_gc_commit(c); + if (err) + goto out_orphans; + } else { + err = take_gc_lnum(c); + if (err) + goto out_orphans; + + /* + * GC LEB may contain garbage if there was an unclean + * reboot, and it should be un-mapped. + */ + err = ubifs_leb_unmap(c, c->gc_lnum); + if (err) + goto out_orphans; + } + + err = dbg_check_lprops(c); + if (err) + goto out_orphans; +#endif + } else if (c->need_recovery) { + err = ubifs_recover_size(c); + if (err) + goto out_orphans; + } else { + /* + * Even if we mount read-only, we have to set space in GC LEB + * to proper value because this affects UBIFS free space + * reporting. We do not want to have a situation when + * re-mounting from R/O to R/W changes amount of free space. + */ + err = take_gc_lnum(c); + if (err) + goto out_orphans; + } + +#ifndef __UBOOT__ + spin_lock(&ubifs_infos_lock); + list_add_tail(&c->infos_list, &ubifs_infos); + spin_unlock(&ubifs_infos_lock); +#endif + + if (c->need_recovery) { + if (c->ro_mount) + ubifs_msg(c, "recovery deferred"); + else { + c->need_recovery = 0; + ubifs_msg(c, "recovery completed"); + /* + * GC LEB has to be empty and taken at this point. But + * the journal head LEBs may also be accounted as + * "empty taken" if they are empty. + */ + ubifs_assert(c->lst.taken_empty_lebs > 0); + } + } else + ubifs_assert(c->lst.taken_empty_lebs > 0); + + err = dbg_check_filesystem(c); + if (err) + goto out_infos; + + err = dbg_debugfs_init_fs(c); + if (err) + goto out_infos; + + c->mounting = 0; + + ubifs_msg(c, "UBIFS: mounted UBI device %d, volume %d, name \"%s\"%s", + c->vi.ubi_num, c->vi.vol_id, c->vi.name, + c->ro_mount ? ", R/O mode" : ""); + x = (long long)c->main_lebs * c->leb_size; +#ifndef CONFIG_UBIFS_SILENCE_MSG + y = (long long)c->log_lebs * c->leb_size + c->max_bud_bytes; +#endif + ubifs_msg(c, "LEB size: %d bytes (%d KiB), min./max. I/O unit sizes: %d bytes/%d bytes", + c->leb_size, c->leb_size >> 10, c->min_io_size, + c->max_write_size); + ubifs_msg(c, "FS size: %lld bytes (%lld MiB, %d LEBs), journal size %lld bytes (%lld MiB, %d LEBs)", + x, x >> 20, c->main_lebs, + y, y >> 20, c->log_lebs + c->max_bud_cnt); + ubifs_msg(c, "reserved for root: %llu bytes (%llu KiB)", + c->report_rp_size, c->report_rp_size >> 10); + ubifs_msg(c, "media format: w%d/r%d (latest is w%d/r%d), UUID %pUB%s", + c->fmt_version, c->ro_compat_version, + UBIFS_FORMAT_VERSION, UBIFS_RO_COMPAT_VERSION, c->uuid, + c->big_lpt ? ", big LPT model" : ", small LPT model"); + + dbg_gen("default compressor: %s", ubifs_compr_name(c->default_compr)); + dbg_gen("data journal heads: %d", + c->jhead_cnt - NONDATA_JHEADS_CNT); + dbg_gen("log LEBs: %d (%d - %d)", + c->log_lebs, UBIFS_LOG_LNUM, c->log_last); + dbg_gen("LPT area LEBs: %d (%d - %d)", + c->lpt_lebs, c->lpt_first, c->lpt_last); + dbg_gen("orphan area LEBs: %d (%d - %d)", + c->orph_lebs, c->orph_first, c->orph_last); + dbg_gen("main area LEBs: %d (%d - %d)", + c->main_lebs, c->main_first, c->leb_cnt - 1); + dbg_gen("index LEBs: %d", c->lst.idx_lebs); + dbg_gen("total index bytes: %lld (%lld KiB, %lld MiB)", + c->bi.old_idx_sz, c->bi.old_idx_sz >> 10, + c->bi.old_idx_sz >> 20); + dbg_gen("key hash type: %d", c->key_hash_type); + dbg_gen("tree fanout: %d", c->fanout); + dbg_gen("reserved GC LEB: %d", c->gc_lnum); + dbg_gen("max. znode size %d", c->max_znode_sz); + dbg_gen("max. index node size %d", c->max_idx_node_sz); + dbg_gen("node sizes: data %zu, inode %zu, dentry %zu", + UBIFS_DATA_NODE_SZ, UBIFS_INO_NODE_SZ, UBIFS_DENT_NODE_SZ); + dbg_gen("node sizes: trun %zu, sb %zu, master %zu", + UBIFS_TRUN_NODE_SZ, UBIFS_SB_NODE_SZ, UBIFS_MST_NODE_SZ); + dbg_gen("node sizes: ref %zu, cmt. start %zu, orph %zu", + UBIFS_REF_NODE_SZ, UBIFS_CS_NODE_SZ, UBIFS_ORPH_NODE_SZ); + dbg_gen("max. node sizes: data %zu, inode %zu dentry %zu, idx %d", + UBIFS_MAX_DATA_NODE_SZ, UBIFS_MAX_INO_NODE_SZ, + UBIFS_MAX_DENT_NODE_SZ, ubifs_idx_node_sz(c, c->fanout)); + dbg_gen("dead watermark: %d", c->dead_wm); + dbg_gen("dark watermark: %d", c->dark_wm); + dbg_gen("LEB overhead: %d", c->leb_overhead); + x = (long long)c->main_lebs * c->dark_wm; + dbg_gen("max. dark space: %lld (%lld KiB, %lld MiB)", + x, x >> 10, x >> 20); + dbg_gen("maximum bud bytes: %lld (%lld KiB, %lld MiB)", + c->max_bud_bytes, c->max_bud_bytes >> 10, + c->max_bud_bytes >> 20); + dbg_gen("BG commit bud bytes: %lld (%lld KiB, %lld MiB)", + c->bg_bud_bytes, c->bg_bud_bytes >> 10, + c->bg_bud_bytes >> 20); + dbg_gen("current bud bytes %lld (%lld KiB, %lld MiB)", + c->bud_bytes, c->bud_bytes >> 10, c->bud_bytes >> 20); + dbg_gen("max. seq. number: %llu", c->max_sqnum); + dbg_gen("commit number: %llu", c->cmt_no); + + return 0; + +out_infos: + spin_lock(&ubifs_infos_lock); + list_del(&c->infos_list); + spin_unlock(&ubifs_infos_lock); +out_orphans: + free_orphans(c); +out_journal: + destroy_journal(c); +out_lpt: + ubifs_lpt_free(c, 0); +out_master: + kfree(c->mst_node); + kfree(c->rcvrd_mst_node); + if (c->bgt) + kthread_stop(c->bgt); +#ifndef __UBOOT__ +out_wbufs: +#endif + free_wbufs(c); +out_cbuf: + kfree(c->cbuf); +out_free: + kfree(c->write_reserve_buf); + kfree(c->bu.buf); + vfree(c->ileb_buf); + vfree(c->sbuf); + kfree(c->bottom_up_buf); + ubifs_debugging_exit(c); + return err; +} + +/** + * ubifs_umount - un-mount UBIFS file-system. + * @c: UBIFS file-system description object + * + * Note, this function is called to free allocated resourced when un-mounting, + * as well as free resources when an error occurred while we were half way + * through mounting (error path cleanup function). So it has to make sure the + * resource was actually allocated before freeing it. + */ +#ifndef __UBOOT__ +static void ubifs_umount(struct ubifs_info *c) +#else +void ubifs_umount(struct ubifs_info *c) +#endif +{ + dbg_gen("un-mounting UBI device %d, volume %d", c->vi.ubi_num, + c->vi.vol_id); + + dbg_debugfs_exit_fs(c); + spin_lock(&ubifs_infos_lock); + list_del(&c->infos_list); + spin_unlock(&ubifs_infos_lock); + +#ifndef __UBOOT__ + if (c->bgt) + kthread_stop(c->bgt); + + destroy_journal(c); +#endif + free_wbufs(c); + free_orphans(c); + ubifs_lpt_free(c, 0); + + kfree(c->cbuf); + kfree(c->rcvrd_mst_node); + kfree(c->mst_node); + kfree(c->write_reserve_buf); + kfree(c->bu.buf); + vfree(c->ileb_buf); + vfree(c->sbuf); + kfree(c->bottom_up_buf); + ubifs_debugging_exit(c); +#ifdef __UBOOT__ + /* Finally free U-Boot's global copy of superblock */ + if (ubifs_sb != NULL) { + free(ubifs_sb->s_fs_info); + free(ubifs_sb); + } +#endif +} + +#ifndef __UBOOT__ +/** + * ubifs_remount_rw - re-mount in read-write mode. + * @c: UBIFS file-system description object + * + * UBIFS avoids allocating many unnecessary resources when mounted in read-only + * mode. This function allocates the needed resources and re-mounts UBIFS in + * read-write mode. + */ +static int ubifs_remount_rw(struct ubifs_info *c) +{ + int err, lnum; + + if (c->rw_incompat) { + ubifs_err(c, "the file-system is not R/W-compatible"); + ubifs_msg(c, "on-flash format version is w%d/r%d, but software only supports up to version w%d/r%d", + c->fmt_version, c->ro_compat_version, + UBIFS_FORMAT_VERSION, UBIFS_RO_COMPAT_VERSION); + return -EROFS; + } + + mutex_lock(&c->umount_mutex); + dbg_save_space_info(c); + c->remounting_rw = 1; + c->ro_mount = 0; + + if (c->space_fixup) { + err = ubifs_fixup_free_space(c); + if (err) + goto out; + } + + err = check_free_space(c); + if (err) + goto out; + + if (c->old_leb_cnt != c->leb_cnt) { + struct ubifs_sb_node *sup; + + sup = ubifs_read_sb_node(c); + if (IS_ERR(sup)) { + err = PTR_ERR(sup); + goto out; + } + sup->leb_cnt = cpu_to_le32(c->leb_cnt); + err = ubifs_write_sb_node(c, sup); + kfree(sup); + if (err) + goto out; + } + + if (c->need_recovery) { + ubifs_msg(c, "completing deferred recovery"); + err = ubifs_write_rcvrd_mst_node(c); + if (err) + goto out; + err = ubifs_recover_size(c); + if (err) + goto out; + err = ubifs_clean_lebs(c, c->sbuf); + if (err) + goto out; + err = ubifs_recover_inl_heads(c, c->sbuf); + if (err) + goto out; + } else { + /* A readonly mount is not allowed to have orphans */ + ubifs_assert(c->tot_orphans == 0); + err = ubifs_clear_orphans(c); + if (err) + goto out; + } + + if (!(c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY))) { + c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY); + err = ubifs_write_master(c); + if (err) + goto out; + } + + c->ileb_buf = vmalloc(c->leb_size); + if (!c->ileb_buf) { + err = -ENOMEM; + goto out; + } + + c->write_reserve_buf = kmalloc(COMPRESSED_DATA_NODE_BUF_SZ, GFP_KERNEL); + if (!c->write_reserve_buf) { + err = -ENOMEM; + goto out; + } + + err = ubifs_lpt_init(c, 0, 1); + if (err) + goto out; + + /* Create background thread */ + c->bgt = kthread_create(ubifs_bg_thread, c, "%s", c->bgt_name); + if (IS_ERR(c->bgt)) { + err = PTR_ERR(c->bgt); + c->bgt = NULL; + ubifs_err(c, "cannot spawn \"%s\", error %d", + c->bgt_name, err); + goto out; + } + wake_up_process(c->bgt); + + c->orph_buf = vmalloc(c->leb_size); + if (!c->orph_buf) { + err = -ENOMEM; + goto out; + } + + /* Check for enough log space */ + lnum = c->lhead_lnum + 1; + if (lnum >= UBIFS_LOG_LNUM + c->log_lebs) + lnum = UBIFS_LOG_LNUM; + if (lnum == c->ltail_lnum) { + err = ubifs_consolidate_log(c); + if (err) + goto out; + } + + if (c->need_recovery) + err = ubifs_rcvry_gc_commit(c); + else + err = ubifs_leb_unmap(c, c->gc_lnum); + if (err) + goto out; + + dbg_gen("re-mounted read-write"); + c->remounting_rw = 0; + + if (c->need_recovery) { + c->need_recovery = 0; + ubifs_msg(c, "deferred recovery completed"); + } else { + /* + * Do not run the debugging space check if the were doing + * recovery, because when we saved the information we had the + * file-system in a state where the TNC and lprops has been + * modified in memory, but all the I/O operations (including a + * commit) were deferred. So the file-system was in + * "non-committed" state. Now the file-system is in committed + * state, and of course the amount of free space will change + * because, for example, the old index size was imprecise. + */ + err = dbg_check_space_info(c); + } + + mutex_unlock(&c->umount_mutex); + return err; + +out: + c->ro_mount = 1; + vfree(c->orph_buf); + c->orph_buf = NULL; + if (c->bgt) { + kthread_stop(c->bgt); + c->bgt = NULL; + } + free_wbufs(c); + kfree(c->write_reserve_buf); + c->write_reserve_buf = NULL; + vfree(c->ileb_buf); + c->ileb_buf = NULL; + ubifs_lpt_free(c, 1); + c->remounting_rw = 0; + mutex_unlock(&c->umount_mutex); + return err; +} + +/** + * ubifs_remount_ro - re-mount in read-only mode. + * @c: UBIFS file-system description object + * + * We assume VFS has stopped writing. Possibly the background thread could be + * running a commit, however kthread_stop will wait in that case. + */ +static void ubifs_remount_ro(struct ubifs_info *c) +{ + int i, err; + + ubifs_assert(!c->need_recovery); + ubifs_assert(!c->ro_mount); + + mutex_lock(&c->umount_mutex); + if (c->bgt) { + kthread_stop(c->bgt); + c->bgt = NULL; + } + + dbg_save_space_info(c); + + for (i = 0; i < c->jhead_cnt; i++) + ubifs_wbuf_sync(&c->jheads[i].wbuf); + + c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_DIRTY); + c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS); + c->mst_node->gc_lnum = cpu_to_le32(c->gc_lnum); + err = ubifs_write_master(c); + if (err) + ubifs_ro_mode(c, err); - err = ubifs_mount_orphans(c, c->need_recovery, mounted_read_only); + vfree(c->orph_buf); + c->orph_buf = NULL; + kfree(c->write_reserve_buf); + c->write_reserve_buf = NULL; + vfree(c->ileb_buf); + c->ileb_buf = NULL; + ubifs_lpt_free(c, 1); + c->ro_mount = 1; + err = dbg_check_space_info(c); if (err) - goto out_orphans; + ubifs_ro_mode(c, err); + mutex_unlock(&c->umount_mutex); +} - if (c->need_recovery) { - err = ubifs_recover_size(c); - if (err) - goto out_orphans; - } +static void ubifs_put_super(struct super_block *sb) +{ + int i; + struct ubifs_info *c = sb->s_fs_info; - spin_lock(&ubifs_infos_lock); - list_add_tail(&c->infos_list, &ubifs_infos); - spin_unlock(&ubifs_infos_lock); + ubifs_msg(c, "un-mount UBI device %d", c->vi.ubi_num); - if (c->need_recovery) { - if (mounted_read_only) - ubifs_msg("recovery deferred"); - else { - c->need_recovery = 0; - ubifs_msg("recovery completed"); - } + /* + * The following asserts are only valid if there has not been a failure + * of the media. For example, there will be dirty inodes if we failed + * to write them back because of I/O errors. + */ + if (!c->ro_error) { + ubifs_assert(c->bi.idx_growth == 0); + ubifs_assert(c->bi.dd_growth == 0); + ubifs_assert(c->bi.data_growth == 0); } - err = dbg_check_filesystem(c); - if (err) - goto out_infos; + /* + * The 'c->umount_lock' prevents races between UBIFS memory shrinker + * and file system un-mount. Namely, it prevents the shrinker from + * picking this superblock for shrinking - it will be just skipped if + * the mutex is locked. + */ + mutex_lock(&c->umount_mutex); + if (!c->ro_mount) { + /* + * First of all kill the background thread to make sure it does + * not interfere with un-mounting and freeing resources. + */ + if (c->bgt) { + kthread_stop(c->bgt); + c->bgt = NULL; + } - c->always_chk_crc = 0; + /* + * On fatal errors c->ro_error is set to 1, in which case we do + * not write the master node. + */ + if (!c->ro_error) { + int err; - ubifs_msg("mounted UBI device %d, volume %d, name \"%s\"", - c->vi.ubi_num, c->vi.vol_id, c->vi.name); - if (mounted_read_only) - ubifs_msg("mounted read-only"); - x = (long long)c->main_lebs * c->leb_size; - ubifs_msg("file system size: %lld bytes (%lld KiB, %lld MiB, %d " - "LEBs)", x, x >> 10, x >> 20, c->main_lebs); - x = (long long)c->log_lebs * c->leb_size + c->max_bud_bytes; - ubifs_msg("journal size: %lld bytes (%lld KiB, %lld MiB, %d " - "LEBs)", x, x >> 10, x >> 20, c->log_lebs + c->max_bud_cnt); - ubifs_msg("media format: w%d/r%d (latest is w%d/r%d)", - c->fmt_version, c->ro_compat_version, - UBIFS_FORMAT_VERSION, UBIFS_RO_COMPAT_VERSION); - ubifs_msg("default compressor: %s", ubifs_compr_name(c->default_compr)); - ubifs_msg("reserved for root: %llu bytes (%llu KiB)", - c->report_rp_size, c->report_rp_size >> 10); - - dbg_msg("compiled on: " __DATE__ " at " __TIME__); - dbg_msg("min. I/O unit size: %d bytes", c->min_io_size); - dbg_msg("LEB size: %d bytes (%d KiB)", - c->leb_size, c->leb_size >> 10); - dbg_msg("data journal heads: %d", - c->jhead_cnt - NONDATA_JHEADS_CNT); - dbg_msg("UUID: %02X%02X%02X%02X-%02X%02X" - "-%02X%02X-%02X%02X-%02X%02X%02X%02X%02X%02X", - c->uuid[0], c->uuid[1], c->uuid[2], c->uuid[3], - c->uuid[4], c->uuid[5], c->uuid[6], c->uuid[7], - c->uuid[8], c->uuid[9], c->uuid[10], c->uuid[11], - c->uuid[12], c->uuid[13], c->uuid[14], c->uuid[15]); - dbg_msg("big_lpt %d", c->big_lpt); - dbg_msg("log LEBs: %d (%d - %d)", - c->log_lebs, UBIFS_LOG_LNUM, c->log_last); - dbg_msg("LPT area LEBs: %d (%d - %d)", - c->lpt_lebs, c->lpt_first, c->lpt_last); - dbg_msg("orphan area LEBs: %d (%d - %d)", - c->orph_lebs, c->orph_first, c->orph_last); - dbg_msg("main area LEBs: %d (%d - %d)", - c->main_lebs, c->main_first, c->leb_cnt - 1); - dbg_msg("index LEBs: %d", c->lst.idx_lebs); - dbg_msg("total index bytes: %lld (%lld KiB, %lld MiB)", - c->old_idx_sz, c->old_idx_sz >> 10, c->old_idx_sz >> 20); - dbg_msg("key hash type: %d", c->key_hash_type); - dbg_msg("tree fanout: %d", c->fanout); - dbg_msg("reserved GC LEB: %d", c->gc_lnum); - dbg_msg("first main LEB: %d", c->main_first); - dbg_msg("max. znode size %d", c->max_znode_sz); - dbg_msg("max. index node size %d", c->max_idx_node_sz); - dbg_msg("node sizes: data %zu, inode %zu, dentry %zu", - UBIFS_DATA_NODE_SZ, UBIFS_INO_NODE_SZ, UBIFS_DENT_NODE_SZ); - dbg_msg("node sizes: trun %zu, sb %zu, master %zu", - UBIFS_TRUN_NODE_SZ, UBIFS_SB_NODE_SZ, UBIFS_MST_NODE_SZ); - dbg_msg("node sizes: ref %zu, cmt. start %zu, orph %zu", - UBIFS_REF_NODE_SZ, UBIFS_CS_NODE_SZ, UBIFS_ORPH_NODE_SZ); - dbg_msg("max. node sizes: data %zu, inode %zu dentry %zu", - UBIFS_MAX_DATA_NODE_SZ, UBIFS_MAX_INO_NODE_SZ, - UBIFS_MAX_DENT_NODE_SZ); - dbg_msg("dead watermark: %d", c->dead_wm); - dbg_msg("dark watermark: %d", c->dark_wm); - dbg_msg("LEB overhead: %d", c->leb_overhead); - x = (long long)c->main_lebs * c->dark_wm; - dbg_msg("max. dark space: %lld (%lld KiB, %lld MiB)", - x, x >> 10, x >> 20); - dbg_msg("maximum bud bytes: %lld (%lld KiB, %lld MiB)", - c->max_bud_bytes, c->max_bud_bytes >> 10, - c->max_bud_bytes >> 20); - dbg_msg("BG commit bud bytes: %lld (%lld KiB, %lld MiB)", - c->bg_bud_bytes, c->bg_bud_bytes >> 10, - c->bg_bud_bytes >> 20); - dbg_msg("current bud bytes %lld (%lld KiB, %lld MiB)", - c->bud_bytes, c->bud_bytes >> 10, c->bud_bytes >> 20); - dbg_msg("max. seq. number: %llu", c->max_sqnum); - dbg_msg("commit number: %llu", c->cmt_no); + /* Synchronize write-buffers */ + for (i = 0; i < c->jhead_cnt; i++) + ubifs_wbuf_sync(&c->jheads[i].wbuf); - return 0; + /* + * We are being cleanly unmounted which means the + * orphans were killed - indicate this in the master + * node. Also save the reserved GC LEB number. + */ + c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_DIRTY); + c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS); + c->mst_node->gc_lnum = cpu_to_le32(c->gc_lnum); + err = ubifs_write_master(c); + if (err) + /* + * Recovery will attempt to fix the master area + * next mount, so we just print a message and + * continue to unmount normally. + */ + ubifs_err(c, "failed to write master node, error %d", + err); + } else { +#ifndef __UBOOT__ + for (i = 0; i < c->jhead_cnt; i++) + /* Make sure write-buffer timers are canceled */ + hrtimer_cancel(&c->jheads[i].wbuf.timer); +#endif + } + } -out_infos: - spin_lock(&ubifs_infos_lock); - list_del(&c->infos_list); - spin_unlock(&ubifs_infos_lock); -out_orphans: - free_orphans(c); -out_journal: -out_lpt: - ubifs_lpt_free(c, 0); -out_master: - kfree(c->mst_node); - kfree(c->rcvrd_mst_node); - if (c->bgt) - kthread_stop(c->bgt); - kfree(c->cbuf); -out_free: - vfree(c->ileb_buf); - vfree(c->sbuf); - kfree(c->bottom_up_buf); - ubifs_debugging_exit(c); - return err; + ubifs_umount(c); +#ifndef __UBOOT__ + bdi_destroy(&c->bdi); +#endif + ubi_close_volume(c->ubi); + mutex_unlock(&c->umount_mutex); } +#endif -/** - * ubifs_umount - un-mount UBIFS file-system. - * @c: UBIFS file-system description object - * - * Note, this function is called to free allocated resourced when un-mounting, - * as well as free resources when an error occurred while we were half way - * through mounting (error path cleanup function). So it has to make sure the - * resource was actually allocated before freeing it. - */ -static void ubifs_umount(struct ubifs_info *c) +#ifndef __UBOOT__ +static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data) { - dbg_gen("un-mounting UBI device %d, volume %d", c->vi.ubi_num, - c->vi.vol_id); + int err; + struct ubifs_info *c = sb->s_fs_info; - spin_lock(&ubifs_infos_lock); - list_del(&c->infos_list); - spin_unlock(&ubifs_infos_lock); + sync_filesystem(sb); + dbg_gen("old flags %#lx, new flags %#x", sb->s_flags, *flags); - if (c->bgt) - kthread_stop(c->bgt); + err = ubifs_parse_options(c, data, 1); + if (err) { + ubifs_err(c, "invalid or unknown remount parameter"); + return err; + } - free_orphans(c); - ubifs_lpt_free(c, 0); + if (c->ro_mount && !(*flags & MS_RDONLY)) { + if (c->ro_error) { + ubifs_msg(c, "cannot re-mount R/W due to prior errors"); + return -EROFS; + } + if (c->ro_media) { + ubifs_msg(c, "cannot re-mount R/W - UBI volume is R/O"); + return -EROFS; + } + err = ubifs_remount_rw(c); + if (err) + return err; + } else if (!c->ro_mount && (*flags & MS_RDONLY)) { + if (c->ro_error) { + ubifs_msg(c, "cannot re-mount R/O due to prior errors"); + return -EROFS; + } + ubifs_remount_ro(c); + } - kfree(c->cbuf); - kfree(c->rcvrd_mst_node); - kfree(c->mst_node); - vfree(c->ileb_buf); - vfree(c->sbuf); - kfree(c->bottom_up_buf); - ubifs_debugging_exit(c); + if (c->bulk_read == 1) + bu_init(c); + else { + dbg_gen("disable bulk-read"); + kfree(c->bu.buf); + c->bu.buf = NULL; + } - /* Finally free U-Boot's global copy of superblock */ - free(ubifs_sb->s_fs_info); - free(ubifs_sb); + ubifs_assert(c->lst.taken_empty_lebs > 0); + return 0; } +#endif + +const struct super_operations ubifs_super_operations = { + .alloc_inode = ubifs_alloc_inode, +#ifndef __UBOOT__ + .destroy_inode = ubifs_destroy_inode, + .put_super = ubifs_put_super, + .write_inode = ubifs_write_inode, + .evict_inode = ubifs_evict_inode, + .statfs = ubifs_statfs, +#endif + .dirty_inode = ubifs_dirty_inode, +#ifndef __UBOOT__ + .remount_fs = ubifs_remount_fs, + .show_options = ubifs_show_options, + .sync_fs = ubifs_sync_fs, +#endif +}; /** * open_ubi - parse UBI device name string and open the UBI device. * @name: UBI volume name * @mode: UBI volume open mode * - * There are several ways to specify UBI volumes when mounting UBIFS: - * o ubiX_Y - UBI device number X, volume Y; - * o ubiY - UBI device number 0, volume Y; + * The primary method of mounting UBIFS is by specifying the UBI volume + * character device node path. However, UBIFS may also be mounted withoug any + * character device node using one of the following methods: + * + * o ubiX_Y - mount UBI device number X, volume Y; + * o ubiY - mount UBI device number 0, volume Y; * o ubiX:NAME - mount UBI device X, volume with name NAME; * o ubi:NAME - mount UBI device 0, volume with name NAME. * * Alternative '!' separator may be used instead of ':' (because some shells * like busybox may interpret ':' as an NFS host name separator). This function - * returns ubi volume object in case of success and a negative error code in - * case of failure. + * returns UBI volume description object in case of success and a negative + * error code in case of failure. */ static struct ubi_volume_desc *open_ubi(const char *name, int mode) { +#ifndef __UBOOT__ + struct ubi_volume_desc *ubi; +#endif int dev, vol; char *endptr; +#ifndef __UBOOT__ + /* First, try to open using the device node path method */ + ubi = ubi_open_volume_path(name, mode); + if (!IS_ERR(ubi)) + return ubi; +#endif + + /* Try the "nodev" method */ if (name[0] != 'u' || name[1] != 'b' || name[2] != 'i') return ERR_PTR(-EINVAL); @@ -904,78 +2191,108 @@ static struct ubi_volume_desc *open_ubi(const char *name, int mode) return ERR_PTR(-EINVAL); } -static int ubifs_fill_super(struct super_block *sb, void *data, int silent) +static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi) { - struct ubi_volume_desc *ubi = sb->s_fs_info; struct ubifs_info *c; - struct inode *root; - int err; c = kzalloc(sizeof(struct ubifs_info), GFP_KERNEL); - if (!c) - return -ENOMEM; + if (c) { + spin_lock_init(&c->cnt_lock); + spin_lock_init(&c->cs_lock); + spin_lock_init(&c->buds_lock); + spin_lock_init(&c->space_lock); + spin_lock_init(&c->orphan_lock); + init_rwsem(&c->commit_sem); + mutex_init(&c->lp_mutex); + mutex_init(&c->tnc_mutex); + mutex_init(&c->log_mutex); + mutex_init(&c->umount_mutex); + mutex_init(&c->bu_mutex); + mutex_init(&c->write_reserve_mutex); + init_waitqueue_head(&c->cmt_wq); + c->buds = RB_ROOT; + c->old_idx = RB_ROOT; + c->size_tree = RB_ROOT; + c->orph_tree = RB_ROOT; + INIT_LIST_HEAD(&c->infos_list); + INIT_LIST_HEAD(&c->idx_gc); + INIT_LIST_HEAD(&c->replay_list); + INIT_LIST_HEAD(&c->replay_buds); + INIT_LIST_HEAD(&c->uncat_list); + INIT_LIST_HEAD(&c->empty_list); + INIT_LIST_HEAD(&c->freeable_list); + INIT_LIST_HEAD(&c->frdi_idx_list); + INIT_LIST_HEAD(&c->unclean_leb_list); + INIT_LIST_HEAD(&c->old_buds); + INIT_LIST_HEAD(&c->orph_list); + INIT_LIST_HEAD(&c->orph_new); + c->no_chk_data_crc = 1; + + c->highest_inum = UBIFS_FIRST_INO; + c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM; + + ubi_get_volume_info(ubi, &c->vi); + ubi_get_device_info(c->vi.ubi_num, &c->di); + } + return c; +} - spin_lock_init(&c->cnt_lock); - spin_lock_init(&c->cs_lock); - spin_lock_init(&c->buds_lock); - spin_lock_init(&c->space_lock); - spin_lock_init(&c->orphan_lock); - init_rwsem(&c->commit_sem); - mutex_init(&c->lp_mutex); - mutex_init(&c->tnc_mutex); - mutex_init(&c->log_mutex); - mutex_init(&c->mst_mutex); - mutex_init(&c->umount_mutex); - init_waitqueue_head(&c->cmt_wq); - c->buds = RB_ROOT; - c->old_idx = RB_ROOT; - c->size_tree = RB_ROOT; - c->orph_tree = RB_ROOT; - INIT_LIST_HEAD(&c->infos_list); - INIT_LIST_HEAD(&c->idx_gc); - INIT_LIST_HEAD(&c->replay_list); - INIT_LIST_HEAD(&c->replay_buds); - INIT_LIST_HEAD(&c->uncat_list); - INIT_LIST_HEAD(&c->empty_list); - INIT_LIST_HEAD(&c->freeable_list); - INIT_LIST_HEAD(&c->frdi_idx_list); - INIT_LIST_HEAD(&c->unclean_leb_list); - INIT_LIST_HEAD(&c->old_buds); - INIT_LIST_HEAD(&c->orph_list); - INIT_LIST_HEAD(&c->orph_new); - - c->highest_inum = UBIFS_FIRST_INO; - c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM; - - ubi_get_volume_info(ubi, &c->vi); - ubi_get_device_info(c->vi.ubi_num, &c->di); +static int ubifs_fill_super(struct super_block *sb, void *data, int silent) +{ + struct ubifs_info *c = sb->s_fs_info; + struct inode *root; + int err; + c->vfs_sb = sb; +#ifndef __UBOOT__ /* Re-open the UBI device in read-write mode */ + c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READWRITE); +#else + /* U-Boot read only mode */ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READONLY); +#endif + if (IS_ERR(c->ubi)) { err = PTR_ERR(c->ubi); - goto out_free; + goto out; } - c->vfs_sb = sb; +#ifndef __UBOOT__ + /* + * UBIFS provides 'backing_dev_info' in order to disable read-ahead. For + * UBIFS, I/O is not deferred, it is done immediately in readpage, + * which means the user would have to wait not just for their own I/O + * but the read-ahead I/O as well i.e. completely pointless. + * + * Read-ahead will be disabled because @c->bdi.ra_pages is 0. + */ + c->bdi.name = "ubifs", + c->bdi.capabilities = 0; + err = bdi_init(&c->bdi); + if (err) + goto out_close; + err = bdi_register(&c->bdi, NULL, "ubifs_%d_%d", + c->vi.ubi_num, c->vi.vol_id); + if (err) + goto out_bdi; + err = ubifs_parse_options(c, data, 0); + if (err) + goto out_bdi; + + sb->s_bdi = &c->bdi; +#endif sb->s_fs_info = c; sb->s_magic = UBIFS_SUPER_MAGIC; sb->s_blocksize = UBIFS_BLOCK_SIZE; sb->s_blocksize_bits = UBIFS_BLOCK_SHIFT; - sb->s_dev = c->vi.cdev; sb->s_maxbytes = c->max_inode_sz = key_max_inode_size(c); if (c->max_inode_sz > MAX_LFS_FILESIZE) sb->s_maxbytes = c->max_inode_sz = MAX_LFS_FILESIZE; - - if (c->rw_incompat) { - ubifs_err("the file-system is not R/W-compatible"); - ubifs_msg("on-flash format version is w%d/r%d, but software " - "only supports up to version w%d/r%d", c->fmt_version, - c->ro_compat_version, UBIFS_FORMAT_VERSION, - UBIFS_RO_COMPAT_VERSION); - return -EROFS; - } + sb->s_op = &ubifs_super_operations; +#ifndef __UBOOT__ + sb->s_xattr = ubifs_xattr_handlers; +#endif mutex_lock(&c->umount_mutex); err = mount_ubifs(c); @@ -991,7 +2308,15 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent) goto out_umount; } +#ifndef __UBOOT__ + sb->s_root = d_make_root(root); + if (!sb->s_root) { + err = -ENOMEM; + goto out_umount; + } +#else sb->s_root = NULL; +#endif mutex_unlock(&c->umount_mutex); return 0; @@ -1000,24 +2325,130 @@ out_umount: ubifs_umount(c); out_unlock: mutex_unlock(&c->umount_mutex); +#ifndef __UBOOT__ +out_bdi: + bdi_destroy(&c->bdi); +out_close: +#endif ubi_close_volume(c->ubi); -out_free: - kfree(c); +out: return err; } static int sb_test(struct super_block *sb, void *data) { - dev_t *dev = data; + struct ubifs_info *c1 = data; + struct ubifs_info *c = sb->s_fs_info; + + return c->vi.cdev == c1->vi.cdev; +} + +static int sb_set(struct super_block *sb, void *data) +{ + sb->s_fs_info = data; + return set_anon_super(sb, NULL); +} + +static struct super_block *alloc_super(struct file_system_type *type, int flags) +{ + struct super_block *s; + int err; + + s = kzalloc(sizeof(struct super_block), GFP_USER); + if (!s) { + err = -ENOMEM; + return ERR_PTR(err); + } + +#ifndef __UBOOT__ + INIT_HLIST_NODE(&s->s_instances); +#endif + INIT_LIST_HEAD(&s->s_inodes); + s->s_time_gran = 1000000000; + s->s_flags = flags; + + return s; +} + +/** + * sget - find or create a superblock + * @type: filesystem type superblock should belong to + * @test: comparison callback + * @set: setup callback + * @flags: mount flags + * @data: argument to each of them + */ +struct super_block *sget(struct file_system_type *type, + int (*test)(struct super_block *,void *), + int (*set)(struct super_block *,void *), + int flags, + void *data) +{ + struct super_block *s = NULL; +#ifndef __UBOOT__ + struct super_block *old; +#endif + int err; - return sb->s_dev == *dev; +#ifndef __UBOOT__ +retry: + spin_lock(&sb_lock); + if (test) { + hlist_for_each_entry(old, &type->fs_supers, s_instances) { + if (!test(old, data)) + continue; + if (!grab_super(old)) + goto retry; + if (s) { + up_write(&s->s_umount); + destroy_super(s); + s = NULL; + } + return old; + } + } +#endif + if (!s) { + spin_unlock(&sb_lock); + s = alloc_super(type, flags); + if (!s) + return ERR_PTR(-ENOMEM); +#ifndef __UBOOT__ + goto retry; +#endif + } + + err = set(s, data); + if (err) { +#ifndef __UBOOT__ + spin_unlock(&sb_lock); + up_write(&s->s_umount); + destroy_super(s); +#endif + return ERR_PTR(err); + } + s->s_type = type; +#ifndef __UBOOT__ + strlcpy(s->s_id, type->name, sizeof(s->s_id)); + list_add_tail(&s->s_list, &super_blocks); + hlist_add_head(&s->s_instances, &type->fs_supers); + spin_unlock(&sb_lock); + get_filesystem(type); + register_shrinker(&s->s_shrink); +#else + strncpy(s->s_id, type->name, sizeof(s->s_id)); +#endif + return s; } -static int ubifs_get_sb(struct file_system_type *fs_type, int flags, - const char *name, void *data, struct vfsmount *mnt) +EXPORT_SYMBOL(sget); + + +static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags, + const char *name, void *data) { struct ubi_volume_desc *ubi; - struct ubi_volume_info vi; + struct ubifs_info *c; struct super_block *sb; int err; @@ -1030,34 +2461,36 @@ static int ubifs_get_sb(struct file_system_type *fs_type, int flags, */ ubi = open_ubi(name, UBI_READONLY); if (IS_ERR(ubi)) { - ubifs_err("cannot open \"%s\", error %d", - name, (int)PTR_ERR(ubi)); - return PTR_ERR(ubi); + pr_err("UBIFS error (pid: %d): cannot open \"%s\", error %d", + current->pid, name, (int)PTR_ERR(ubi)); + return ERR_CAST(ubi); + } + + c = alloc_ubifs_info(ubi); + if (!c) { + err = -ENOMEM; + goto out_close; } - ubi_get_volume_info(ubi, &vi); - dbg_gen("opened ubi%d_%d", vi.ubi_num, vi.vol_id); + dbg_gen("opened ubi%d_%d", c->vi.ubi_num, c->vi.vol_id); - sb = sget(fs_type, &sb_test, &sb_set, &vi.cdev); + sb = sget(fs_type, sb_test, sb_set, flags, c); if (IS_ERR(sb)) { err = PTR_ERR(sb); + kfree(c); goto out_close; } if (sb->s_root) { + struct ubifs_info *c1 = sb->s_fs_info; + kfree(c); /* A new mount point for already mounted UBIFS */ dbg_gen("this ubi volume is already mounted"); - if ((flags ^ sb->s_flags) & MS_RDONLY) { + if (!!(flags & MS_RDONLY) != c1->ro_mount) { err = -EBUSY; goto out_deact; } } else { - sb->s_flags = flags; - /* - * Pass 'ubi' to 'fill_super()' in sb->s_fs_info where it is - * replaced by 'c'. - */ - sb->s_fs_info = ubi; err = ubifs_fill_super(sb, data, flags & MS_SILENT ? 1 : 0); if (err) goto out_deact; @@ -1068,17 +2501,53 @@ static int ubifs_get_sb(struct file_system_type *fs_type, int flags, /* 'fill_super()' opens ubi again so we must close it here */ ubi_close_volume(ubi); +#ifdef __UBOOT__ ubifs_sb = sb; return 0; +#else + return dget(sb->s_root); +#endif out_deact: - up_write(&sb->s_umount); +#ifndef __UBOOT__ + deactivate_locked_super(sb); +#endif out_close: ubi_close_volume(ubi); - return err; + return ERR_PTR(err); +} + +static void kill_ubifs_super(struct super_block *s) +{ + struct ubifs_info *c = s->s_fs_info; +#ifndef __UBOOT__ + kill_anon_super(s); +#endif + kfree(c); } -int __init ubifs_init(void) +static struct file_system_type ubifs_fs_type = { + .name = "ubifs", + .owner = THIS_MODULE, + .mount = ubifs_mount, + .kill_sb = kill_ubifs_super, +}; +#ifndef __UBOOT__ +MODULE_ALIAS_FS("ubifs"); + +/* + * Inode slab cache constructor. + */ +static void inode_slab_ctor(void *obj) +{ + struct ubifs_inode *ui = obj; + inode_init_once(&ui->vfs_inode); +} + +static int __init ubifs_init(void) +#else +int ubifs_init(void) +#endif { int err; @@ -1134,42 +2603,88 @@ int __init ubifs_init(void) * UBIFS_BLOCK_SIZE. It is assumed that both are powers of 2. */ if (PAGE_CACHE_SIZE < UBIFS_BLOCK_SIZE) { - ubifs_err("VFS page cache size is %u bytes, but UBIFS requires" - " at least 4096 bytes", - (unsigned int)PAGE_CACHE_SIZE); + pr_err("UBIFS error (pid %d): VFS page cache size is %u bytes, but UBIFS requires at least 4096 bytes", + current->pid, (unsigned int)PAGE_CACHE_SIZE); return -EINVAL; } - err = -ENOMEM; +#ifndef __UBOOT__ + ubifs_inode_slab = kmem_cache_create("ubifs_inode_slab", + sizeof(struct ubifs_inode), 0, + SLAB_MEM_SPREAD | SLAB_RECLAIM_ACCOUNT, + &inode_slab_ctor); + if (!ubifs_inode_slab) + return -ENOMEM; + + err = register_shrinker(&ubifs_shrinker_info); + if (err) + goto out_slab; +#endif err = ubifs_compressors_init(); if (err) goto out_shrinker; +#ifndef __UBOOT__ + err = dbg_debugfs_init(); + if (err) + goto out_compr; + + err = register_filesystem(&ubifs_fs_type); + if (err) { + pr_err("UBIFS error (pid %d): cannot register file system, error %d", + current->pid, err); + goto out_dbg; + } +#endif return 0; +#ifndef __UBOOT__ +out_dbg: + dbg_debugfs_exit(); +out_compr: + ubifs_compressors_exit(); +#endif out_shrinker: +#ifndef __UBOOT__ + unregister_shrinker(&ubifs_shrinker_info); +out_slab: +#endif + kmem_cache_destroy(ubifs_inode_slab); return err; } +/* late_initcall to let compressors initialize first */ +late_initcall(ubifs_init); -/* - * ubifsmount... - */ +#ifndef __UBOOT__ +static void __exit ubifs_exit(void) +{ + ubifs_assert(list_empty(&ubifs_infos)); + ubifs_assert(atomic_long_read(&ubifs_clean_zn_cnt) == 0); -static struct file_system_type ubifs_fs_type = { - .name = "ubifs", - .owner = THIS_MODULE, - .get_sb = ubifs_get_sb, -}; + dbg_debugfs_exit(); + ubifs_compressors_exit(); + unregister_shrinker(&ubifs_shrinker_info); -int ubifs_mount(char *vol_name) + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); + kmem_cache_destroy(ubifs_inode_slab); + unregister_filesystem(&ubifs_fs_type); +} +module_exit(ubifs_exit); + +MODULE_LICENSE("GPL"); +MODULE_VERSION(__stringify(UBIFS_VERSION)); +MODULE_AUTHOR("Artem Bityutskiy, Adrian Hunter"); +MODULE_DESCRIPTION("UBIFS - UBI File System"); +#else +int uboot_ubifs_mount(char *vol_name) { + struct dentry *ret; int flags; - char name[80] = "ubi:"; - void *data; - struct vfsmount *mnt; - int ret; - struct ubifs_info *c; /* * First unmount if allready mounted @@ -1177,24 +2692,17 @@ int ubifs_mount(char *vol_name) if (ubifs_sb) ubifs_umount(ubifs_sb->s_fs_info); - INIT_LIST_HEAD(&ubifs_infos); - INIT_LIST_HEAD(&ubifs_fs_type.fs_supers); - /* * Mount in read-only mode */ flags = MS_RDONLY; - strcat(name, vol_name); - data = NULL; - mnt = NULL; - ret = ubifs_get_sb(&ubifs_fs_type, flags, name, data, mnt); - if (ret) { - printf("Error reading superblock on volume '%s'!\n", name); + ret = ubifs_mount(&ubifs_fs_type, flags, vol_name, NULL); + if (IS_ERR(ret)) { + printf("Error reading superblock on volume '%s' " \ + "errno=%d!\n", vol_name, (int)PTR_ERR(ret)); return -1; } - c = ubifs_sb->s_fs_info; - ubi_close_volume(c->ubi); - return 0; } +#endif