[PATCH take 2 07/28] UBIFS: add file-system build

From: Artem Bityutskiy
Date: Tue May 06 2008 - 04:51:44 EST


The file-system build code contains most of the UBIFS initialization
and mount-related functionality implementation.

Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@xxxxxxxxx>
Signed-off-by: Adrian Hunter <ext-adrian.hunter@xxxxxxxxx>
---
fs/ubifs/build.c | 1374 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ubifs/super.c | 525 +++++++++++++++++++++
2 files changed, 1899 insertions(+), 0 deletions(-)

diff --git a/fs/ubifs/build.c b/fs/ubifs/build.c
new file mode 100644
index 0000000..b37ed6b
--- /dev/null
+++ b/fs/ubifs/build.c
@@ -0,0 +1,1374 @@
+/*
+ * 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
+ */
+
+/*
+ * This file implements UBIFS initialization, mount and un-mount. Some
+ * initialization stuff which is rather large and complex is placed at
+ * corresponding subsystems, but most of it is here.
+ */
+
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/module.h>
+#include <linux/ctype.h>
+#include <linux/random.h>
+#include <linux/kthread.h>
+#include <linux/parser.h>
+#include "ubifs.h"
+
+/* Slab cache for UBIFS inodes */
+struct kmem_cache *ubifs_inode_slab;
+
+/* UBIFS TNC shrinker description */
+static struct shrinker ubifs_shrinker_info = {
+ .shrink = ubifs_shrinker,
+ .seeks = DEFAULT_SEEKS,
+};
+
+/**
+ * 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("UBI volume is corrupted - read-only mode");
+ c->ro_media = 1;
+ }
+
+ if (c->di.ro_mode) {
+ ubifs_msg("read-only UBI device");
+ c->ro_media = 1;
+ }
+
+ if (c->vi.vol_type == UBI_STATIC_VOLUME) {
+ ubifs_msg("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->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;
+
+ 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;
+ }
+
+ 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;
+ }
+
+ if (!is_power_of_2(c->min_io_size)) {
+ ubifs_err("bad min. I/O size %d", 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;
+ }
+
+ 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.
+ */
+ c->ranges[UBIFS_IDX_NODE].max_len = INT_MAX;
+
+ /*
+ * Initialize dead and dark LEB space watermarks.
+ *
+ * Dead space is the space which cannot be used. Its watermark is
+ * equivalent to min. I/O unit or minimum node size if it is greater
+ * then min. I/O unit.
+ *
+ * Dark space is the space which might be used, or might not, depending
+ * on which node should be written to the LEB. Its watermark is
+ * equivalent to maximum UBIFS node size.
+ */
+ c->dead_wm = ALIGN(MIN_WRITE_SZ, c->min_io_size);
+ c->dark_wm = ALIGN(UBIFS_MAX_NODE_SZ, c->min_io_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_late - initialize UBIFS constants.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which initializes various UBIFS constants after
+ * the superblock has been read. It also checks various UBIFS parameters and
+ * makes sure they are all right. Returns zero in case of success and a
+ * negative error code in case of failure.
+ */
+static int init_constants_late(struct ubifs_info *c)
+{
+ int tmp, err;
+ uint64_t tmp64;
+
+ c->main_bytes = c->main_lebs * c->leb_size;
+
+ c->max_znode_sz = sizeof(struct ubifs_znode) +
+ c->fanout * sizeof(struct ubifs_zbranch);
+
+ tmp = ubifs_idx_node_sz(c, 1);
+ c->ranges[UBIFS_IDX_NODE].min_len = tmp;
+ c->min_idx_node_sz = ALIGN(tmp, 8);
+
+ tmp = ubifs_idx_node_sz(c, c->fanout);
+ c->ranges[UBIFS_IDX_NODE].max_len = tmp;
+ c->max_idx_node_sz = ALIGN(tmp, 8);
+
+ /* Make sure LEB size is large enough to fit full commit */
+ 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);
+ return -EINVAL;
+ }
+
+ /*
+ * Make sure that the log is large enough to fit reference nodes for
+ * all buds plus one reserved LEB.
+ */
+ tmp64 = c->max_bud_bytes;
+ tmp = do_div(tmp64, c->leb_size);
+ c->max_bud_cnt = tmp64 + !!tmp;
+ tmp = (c->ref_node_alsz * c->max_bud_cnt + c->leb_size - 1);
+ 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);
+ return -EINVAL;
+ }
+
+ /*
+ * When budgeting we assume worst-case scenarios when the pages are not
+ * 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'.
+ *
+ * c->page_budget is PAGE_CACHE_SIZE + UBIFS_CH_SZ * blocks_per_page
+ */
+ c->page_budget = PAGE_CACHE_SIZE + UBIFS_CH_SZ;
+ c->inode_budget = UBIFS_INO_NODE_SZ;
+ c->dent_budget = UBIFS_MAX_DENT_NODE_SZ;
+
+ /*
+ * When the amount of flash space used by buds becomes
+ * 'c->max_bud_bytes', UBIFS just blocks all writers and starts commit.
+ * The writers are unblocked when the commit is finished. To avoid
+ * writers to be blocked UBIFS initiates background commit in advance,
+ * when number of bud bytes becomes above the limit defined below.
+ */
+ c->bg_bud_bytes = (c->max_bud_bytes * 13) >> 4;
+
+ /*
+ * Ensure minimum journal size. All the bytes in the journal heads are
+ * considered to be used, when calculating the current journal usage.
+ * Consequently, if the journal is too small, UBIFS will treat it as
+ * always full.
+ */
+ tmp64 = (uint64_t)(c->jhead_cnt + 1) * c->leb_size + 1;
+ if (c->bg_bud_bytes < tmp64)
+ c->bg_bud_bytes = tmp64;
+ if (c->max_bud_bytes < tmp64 + c->leb_size)
+ c->max_bud_bytes = tmp64 + c->leb_size;
+
+ err = ubifs_calc_lpt_geom(c);
+ if (err)
+ return err;
+
+ c->min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+
+ /*
+ * Calculate total amount of FS blocks. This number is not used
+ * internally because it does not make much sense for UBIFS, but it is
+ * necessary to report something for the 'statfs()' call.
+ *
+ * Subtract the LEB reserved for GC and the LEB which is reserved for
+ * deletions.
+ *
+ * Review 'ubifs_calc_available()' if changing this calculation.
+ */
+ tmp64 = c->main_lebs - 2;
+ tmp64 *= c->leb_size - c->dark_wm;
+ tmp64 = ubifs_reported_space(c, tmp64);
+ c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT;
+
+ return 0;
+}
+
+/**
+ * care_about_gc_lnum - take care about reserved GC LEB.
+ * @c: UBIFS file-system description object
+ *
+ * This function ensures that the LEB reserved for garbage collection is
+ * unmapped and 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 care_about_gc_lnum(struct ubifs_info *c)
+{
+ int err;
+
+ if (c->gc_lnum == -1) {
+ ubifs_err("no LEB for GC");
+ return -EINVAL;
+ }
+
+ err = ubifs_leb_unmap(c, c->gc_lnum);
+ if (err)
+ return err;
+
+ /* 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 = kzalloc(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[BASEHD].wbuf.dtype = UBI_SHORTTERM;
+ /*
+ * Garbage Collector head likely contains long-term data and
+ * does not need to be synchronized by timer.
+ */
+ c->jheads[GCHD].wbuf.dtype = UBI_LONGTERM;
+ c->jheads[GCHD].wbuf.timeout = 0;
+
+ sprintf(c->bgt_name, "%s%d_%d", SYNCER_BG_NAME,
+ c->vi.ubi_num, c->vi.vol_id);
+
+ 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
+ */
+static void free_orphans(struct ubifs_info *c)
+{
+ struct ubifs_orphan *orph;
+
+ while (c->orph_dnext) {
+ orph = c->orph_dnext;
+ c->orph_dnext = orph->dnext;
+ list_del(&orph->list);
+ kfree(orph);
+ }
+
+ while (!list_empty(&c->orph_list)) {
+ 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");
+ }
+
+ 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 rb_node *this = c->buds.rb_node;
+ struct ubifs_bud *bud;
+
+ while (this) {
+ if (this->rb_left)
+ this = this->rb_left;
+ else if (this->rb_right)
+ this = this->rb_right;
+ else {
+ bud = rb_entry(this, struct ubifs_bud, rb);
+ this = rb_parent(this);
+ if (this) {
+ if (this->rb_left == &bud->rb)
+ this->rb_left = NULL;
+ else
+ this->rb_right = NULL;
+ }
+ kfree(bud);
+ }
+ }
+}
+
+/**
+ * check_volume_empty - check if the UBI volume is empty.
+ * @c: UBIFS file-system description object
+ *
+ * This function checks if the UBIFS volume is empty by looking if its LEBs are
+ * mapped or not. The result of checking is stored in the @c->empty variable.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int check_volume_empty(struct ubifs_info *c)
+{
+ int lnum, err;
+
+ c->empty = 1;
+ for (lnum = 0; lnum < c->leb_cnt; lnum++) {
+ err = ubi_is_mapped(c->ubi, lnum);
+ if (unlikely(err < 0))
+ return err;
+ if (err == 1) {
+ c->empty = 0;
+ break;
+ }
+
+ cond_resched();
+ }
+
+ return 0;
+}
+
+/**
+ * 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.
+ *
+ * Note, the function does not de-allocate resources it it fails half way
+ * through, and the caller has to do this instead.
+ */
+static int mount_ubifs(struct ubifs_info *c)
+{
+ struct super_block *sb = c->vfs_sb;
+ int err, mounted_read_only = (sb->s_flags & MS_RDONLY);
+ unsigned long long x;
+ size_t sz;
+
+ err = init_constants_early(c);
+ if (err)
+ return err;
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ c->dbg_buf = vmalloc(c->leb_size);
+ if (!c->dbg_buf)
+ return -ENOMEM;
+#endif
+
+ err = check_volume_empty(c);
+ if (err)
+ return err;
+
+ if (c->empty && (mounted_read_only || 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("can't format empty UBI volume: read-only %s",
+ c->ro_media ? "UBI volume" : "mount");
+ return -EROFS;
+ }
+
+ if (c->ro_media && !mounted_read_only) {
+ ubifs_err("cannot mount read-write - read-only media");
+ return -EROFS;
+ }
+
+ /*
+ * The requirement for the buffer is that it should fit indexing B-tree
+ * height amount of integers. We assume the height if the TNC tree will
+ * never exceed 64.
+ */
+ c->bottom_up_buf = kmalloc(BOTTOM_UP_HEIGHT * sizeof(int), GFP_KERNEL);
+ if (!c->bottom_up_buf)
+ return -ENOMEM;
+
+ c->sbuf = vmalloc(c->leb_size);
+ if (!c->sbuf)
+ return -ENOMEM;
+
+ if (!mounted_read_only) {
+ c->ileb_buf = vmalloc(c->leb_size);
+ if (!c->ileb_buf)
+ return -ENOMEM;
+ }
+
+ err = ubifs_read_superblock(c);
+ if (err)
+ return err;
+
+ /*
+ * Make sure the compressor which is set as the default on in the
+ * superblock was actually compiled in.
+ */
+ if (!ubifs_compr_present(c->default_compr)) {
+ ubifs_warn("'%s' compressor is set by superblock, but not "
+ "compiled in", ubifs_compr_name(c->default_compr));
+ c->default_compr = UBIFS_COMPR_NONE;
+ }
+
+ dbg_failure_mode_registration(c);
+
+ err = init_constants_late(c);
+ if (err)
+ return err;
+
+ sz = ALIGN(c->max_idx_node_sz, c->min_io_size);
+ sz = ALIGN(sz + c->max_idx_node_sz, c->min_io_size);
+ c->cbuf = kmalloc(sz, GFP_NOFS);
+ if (!c->cbuf)
+ return -ENOMEM;
+
+ if (!mounted_read_only) {
+ err = alloc_wbufs(c);
+ if (err)
+ return err;
+
+ /* Create background thread */
+ c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name);
+ if (!c->bgt)
+ c->bgt = ERR_PTR(-EINVAL);
+ if (IS_ERR(c->bgt)) {
+ err = PTR_ERR(c->bgt);
+ c->bgt = NULL;
+ ubifs_err("cannot spawn \"%s\", error %d",
+ c->bgt_name, err);
+ return err;
+ }
+ }
+
+ err = ubifs_read_master(c);
+ if (err)
+ return err;
+
+ if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) {
+ ubifs_msg("recovery needed");
+ c->need_recovery = 1;
+ if (!mounted_read_only) {
+ err = ubifs_recover_inl_heads(c, c->sbuf);
+ if (err)
+ return err;
+ }
+ } else if (!mounted_read_only) {
+ /*
+ * 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)
+ return err;
+ }
+
+ err = ubifs_lpt_init(c, 1, !mounted_read_only);
+ if (err)
+ return err;
+
+ err = dbg_check_idx_size(c, c->old_idx_sz);
+ if (err)
+ return err;
+
+ err = ubifs_replay_journal(c);
+ if (err)
+ return err;
+
+ if (!mounted_read_only) {
+ int lnum;
+
+ if (c->need_recovery)
+ err = ubifs_recover_gc_lnum(c);
+ else
+ err = care_about_gc_lnum(c);
+ if (err)
+ return err;
+ err = ubifs_mount_orphans(c, c->need_recovery);
+ if (err)
+ return err;
+
+ /* 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)
+ return err;
+ }
+
+ /* Check for enough free space */
+ if (ubifs_calc_available(c) <= 0) {
+ ubifs_err("insufficient available space");
+ return -EINVAL;
+ }
+
+ err = dbg_check_lprops(c);
+ if (err)
+ return err;
+ }
+
+ if (c->need_recovery) {
+ err = ubifs_recover_size(c);
+ if (err)
+ return err;
+ }
+
+ spin_lock(&ubifs_infos_lock);
+ list_add_tail(&c->infos_list, &ubifs_infos);
+ spin_unlock(&ubifs_infos_lock);
+
+ if (c->need_recovery) {
+ if (mounted_read_only)
+ ubifs_msg("recovery deferred");
+ else {
+ c->need_recovery = 0;
+ ubifs_msg("recovery completed");
+ }
+ }
+
+ ubifs_msg("mounted UBI device %d, volume %d", c->vi.ubi_num,
+ c->vi.vol_id);
+ if (mounted_read_only)
+ ubifs_msg("mounted read-only");
+ ubifs_msg("minimal I/O unit size: %d bytes", c->min_io_size);
+ ubifs_msg("logical eraseblock size: %d bytes (%d KiB)",
+ c->leb_size, c->leb_size / 1024);
+ x = (unsigned 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 = (unsigned 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("data journal heads: %d",
+ c->jhead_cnt - NONDATA_JHEADS_CNT);
+ ubifs_msg("default compressor: %s",
+ ubifs_compr_name(c->default_compr));
+
+ dbg_msg("compiled on: " __DATE__ " at " __TIME__);
+ dbg_msg("fast unmount: %d", c->fast_unmount);
+ 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("dead watermark: %d", c->dead_wm);
+ dbg_msg("dark watermark: %d", c->dark_wm);
+ x = 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);
+
+ return 0;
+}
+
+/**
+ * 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.
+ */
+void ubifs_umount(struct ubifs_info *c)
+{
+ dbg_gen("un-mounting UBI device %d, volume %d", c->vi.ubi_num,
+ c->vi.vol_id);
+
+ ubifs_destroy_size_tree(c);
+
+ if (c->bgt)
+ kthread_stop(c->bgt);
+
+ free_buds(c);
+ ubifs_destroy_idx_gc(c);
+ ubifs_tnc_close(c);
+
+ free_wbufs(c);
+ free_orphans(c);
+ ubifs_lpt_free(c, 0);
+
+ 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);
+ }
+
+ kfree(c->rcvrd_mst_node);
+ kfree(c->mst_node);
+ vfree(c->sbuf);
+ kfree(c->bottom_up_buf);
+ UBIFS_DBG(vfree(c->dbg_buf));
+ vfree(c->ileb_buf);
+ dbg_failure_mode_deregistration(c);
+}
+
+/**
+ * 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.
+ */
+int ubifs_remount_rw(struct ubifs_info *c)
+{
+ int err, lnum;
+
+ if (c->ro_media)
+ return -EINVAL;
+
+ mutex_lock(&c->umount_mutex);
+ c->remounting_rw = 1;
+
+ /* Check for enough free space */
+ if (ubifs_calc_available(c) <= 0) {
+ ubifs_err("insufficient available space");
+ err = -EINVAL;
+ 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);
+ if (err)
+ goto out;
+ }
+
+ if (c->need_recovery) {
+ ubifs_msg("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;
+ }
+
+ 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;
+ }
+
+ err = ubifs_lpt_init(c, 0, 1);
+ if (err)
+ goto out;
+
+ err = alloc_wbufs(c);
+ if (err)
+ goto out;
+
+ ubifs_create_buds_lists(c);
+
+ /* Create background thread */
+ c->bgt = kthread_create(ubifs_bg_thread, c, c->bgt_name);
+ if (!c->bgt)
+ c->bgt = ERR_PTR(-EINVAL);
+ if (IS_ERR(c->bgt)) {
+ err = PTR_ERR(c->bgt);
+ c->bgt = NULL;
+ ubifs_err("cannot spawn \"%s\", error %d",
+ c->bgt_name, err);
+ return err;
+ }
+
+ if (c->need_recovery)
+ err = ubifs_recover_gc_lnum(c);
+ else
+ err = care_about_gc_lnum(c);
+ if (err)
+ goto out;
+
+ err = ubifs_mount_orphans(c, c->need_recovery);
+ if (err)
+ 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) {
+ c->need_recovery = 0;
+ ubifs_msg("deferred recovery completed");
+ }
+
+ dbg_gen("re-mounted read-write");
+ c->vfs_sb->s_flags &= ~MS_RDONLY;
+ c->remounting_rw = 0;
+ mutex_unlock(&c->umount_mutex);
+ return 0;
+
+out:
+ free_orphans(c);
+ if (c->bgt) {
+ kthread_stop(c->bgt);
+ c->bgt = NULL;
+ }
+ free_wbufs(c);
+ vfree(c->ileb_buf);
+ c->ileb_buf = NULL;
+ ubifs_lpt_free(c, 1);
+ c->remounting_rw = 0;
+ mutex_unlock(&c->umount_mutex);
+ return err;
+}
+
+/**
+ * commit_on_unmount - commit the journal when un-mounting.
+ * @c: UBIFS file-system description object
+ *
+ * This function is called during un-mounting and it commits the journal unless
+ * the "fast unmount" mode is enabled. It also avoids committing the journal if
+ * it contains too few data.
+ *
+ * Sometimes recovery requires the journal to be committed at least once, and
+ * this function takes care about this.
+ */
+static void commit_on_unmount(struct ubifs_info *c)
+{
+ if (!c->fast_unmount) {
+ long long bud_bytes;
+
+ spin_lock(&c->buds_lock);
+ bud_bytes = c->bud_bytes;
+ spin_unlock(&c->buds_lock);
+ if (bud_bytes > c->leb_size)
+ ubifs_run_commit(c);
+ }
+
+ if (c->recovery_needs_commit)
+ ubifs_recovery_commit(c);
+}
+
+/**
+ * ubifs_remount_ro - re-mount in read-only mode.
+ * @c: UBIFS file-system description object
+ *
+ * We rely on VFS to have stopped writing. Possibly the background thread could
+ * be running a commit, however kthread_stop will wait in that case.
+ */
+void ubifs_remount_ro(struct ubifs_info *c)
+{
+ int i, err;
+
+ ubifs_assert(!c->need_recovery);
+ commit_on_unmount(c);
+
+ mutex_lock(&c->umount_mutex);
+ if (c->bgt) {
+ kthread_stop(c->bgt);
+ c->bgt = NULL;
+ }
+
+ for (i = 0; i < c->jhead_cnt; i++) {
+ ubifs_wbuf_sync(&c->jheads[i].wbuf);
+ del_timer_sync(&c->jheads[i].wbuf.timer);
+ }
+
+ if (!c->ro_media) {
+ 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);
+ }
+
+ ubifs_destroy_idx_gc(c);
+ free_wbufs(c);
+ free_orphans(c);
+ vfree(c->ileb_buf);
+ c->ileb_buf = NULL;
+ ubifs_lpt_free(c, 1);
+ mutex_unlock(&c->umount_mutex);
+}
+
+/**
+ * open_ubi - parse UBI device name string and open the UBI device.
+ * @c: UBIFS file-system description object
+ * @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;
+ * 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 zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int open_ubi(struct ubifs_info *c, const char *name, int mode)
+{
+ int dev, vol;
+ char *endptr;
+
+ if (name[0] != 'u' || name[1] != 'b' || name[2] != 'i')
+ return -EINVAL;
+
+ if ((name[3] == ':' || name[3] == '!') && name[4] != '\0') {
+ /* ubi:NAME method */
+ c->ubi = ubi_open_volume_nm(0, name + 4, mode);
+ if (IS_ERR(c->ubi))
+ return PTR_ERR(c->ubi);
+ } else if (isdigit(name[3])) {
+ dev = simple_strtoul(name + 3, &endptr, 0);
+ if (*endptr == '\0') {
+ /* ubiY method */
+ c->ubi = ubi_open_volume(0, dev, mode);
+ if (IS_ERR(c->ubi))
+ return PTR_ERR(c->ubi);
+ } else if (*endptr == '_' && isdigit(endptr[1])) {
+ /* ubiX_Y method */
+ vol = simple_strtoul(endptr + 1, &endptr, 0);
+ if (*endptr != '\0')
+ return -EINVAL;
+ c->ubi = ubi_open_volume(dev, vol, mode);
+ if (IS_ERR(c->ubi))
+ return PTR_ERR(c->ubi);
+ } else if ((*endptr == ':' || *endptr == '!') &&
+ endptr[1] != '\0') {
+ /* ubiX:NAME method */
+ c->ubi = ubi_open_volume_nm(dev, ++endptr, mode);
+ if (IS_ERR(c->ubi))
+ return PTR_ERR(c->ubi);
+ }
+ }
+
+ if (!c->ubi)
+ return -EINVAL;
+
+ ubi_get_volume_info(c->ubi, &c->vi);
+ ubi_get_device_info(c->vi.ubi_num, &c->di);
+ return 0;
+}
+
+static int sb_test(struct super_block *sb, void *data)
+{
+ dev_t *dev = data;
+
+ return sb->s_dev == *dev;
+}
+
+static int sb_set(struct super_block *sb, void *data)
+{
+ 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_err: just end of array marker
+ */
+enum {
+ Opt_fast_unmount,
+ Opt_norm_unmount,
+ Opt_err,
+};
+
+static match_table_t tokens = {
+ {Opt_fast_unmount, "fast_unmount"},
+ {Opt_norm_unmount, "norm_unmount"},
+ {Opt_err, NULL},
+};
+
+/**
+ * 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.
+ */
+int ubifs_parse_options(struct ubifs_info *c, char *options, int is_remount)
+{
+ char *p;
+ substring_t args[MAX_OPT_ARGS];
+
+ if (!options)
+ return 0;
+
+ while ((p = strsep(&options, ",")) != NULL) {
+ int token;
+
+ if (!*p)
+ continue;
+
+ token = match_token(p, tokens, args);
+ switch (token) {
+ case Opt_fast_unmount:
+ c->mount_opts.unmount_mode = 2;
+ c->fast_unmount = 1;
+ break;
+ case Opt_norm_unmount:
+ c->mount_opts.unmount_mode = 1;
+ c->fast_unmount = 0;
+ break;
+ default:
+ ubifs_err("unrecognized mount option \"%s\" "
+ "or missing value", p);
+ return -EINVAL;
+ }
+ }
+
+ return 0;
+}
+
+static int ubifs_get_sb(struct file_system_type *fs_type, int flags,
+ const char *name, void *data, struct vfsmount *mnt)
+{
+ int err;
+ struct super_block *sb;
+ struct ubifs_info *c;
+ struct inode *root;
+
+ dbg_gen("name %s, flags %#x", name, flags);
+
+ c = kzalloc(sizeof(struct ubifs_info), GFP_KERNEL);
+ if (!c)
+ return -ENOMEM;
+
+ 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;
+ get_random_bytes(&c->vfs_gen, sizeof(int));
+ c->lhead_lnum = c->ltail_lnum = UBIFS_LOG_LNUM;
+
+ err = ubifs_parse_options(c, data, 0);
+ if (err)
+ goto out_free;
+
+ /*
+ * Get UBI device number and volume ID. Mount it read-only so far
+ * because this might be a new mount point, and UBI allows only one
+ * read-write user at a time.
+ */
+ err = open_ubi(c, name, UBI_READONLY);
+ if (err) {
+ ubifs_err("cannot open \"%s\", error %d", name, err);
+ goto out_free;
+ }
+
+ dbg_gen("opened ubi%d_%d", c->vi.ubi_num, c->vi.vol_id);
+
+ sb = sget(fs_type, &sb_test, &sb_set, &c->vi.cdev);
+ if (IS_ERR(sb)) {
+ err = PTR_ERR(sb);
+ goto out_close;
+ }
+
+ if (sb->s_root) {
+ /* A new mount point for already mounted UBIFS */
+ dbg_gen("this ubi volume is already mounted");
+ err = simple_set_mnt(mnt, sb);
+ goto out_close;
+ }
+
+ /* Re-open the UBI device in read-write mode */
+ ubi_close_volume(c->ubi);
+ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READWRITE);
+ if (IS_ERR(c->ubi)) {
+ err = PTR_ERR(c->ubi);
+ goto out_free;
+ }
+
+ c->vfs_sb = sb;
+ 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 =
+ min_t(uint64_t, MAX_LFS_FILESIZE, UBIFS_MAX_INODE_SZ);
+ sb->s_op = &ubifs_super_operations;
+ sb->s_flags = flags;
+
+ mutex_lock(&c->umount_mutex);
+ err = mount_ubifs(c);
+ if (err) {
+ ubifs_assert(err < 0);
+ goto out_umount;
+ }
+
+ /* Read the root inode */
+ root = ubifs_iget(sb, UBIFS_ROOT_INO);
+ if (IS_ERR(root)) {
+ err = PTR_ERR(root);
+ goto out_umount;
+ }
+
+ sb->s_root = d_alloc_root(root);
+ if (!sb->s_root)
+ goto out_iput;
+
+ mutex_unlock(&c->umount_mutex);
+
+ /* We do not support atime */
+ sb->s_flags |= MS_ACTIVE | MS_NOATIME;
+ return simple_set_mnt(mnt, sb);
+
+out_iput:
+ iput(root);
+out_umount:
+ spin_lock(&ubifs_infos_lock);
+ if (c->infos_list.next)
+ list_del(&c->infos_list);
+ spin_unlock(&ubifs_infos_lock);
+ ubifs_umount(c);
+ mutex_unlock(&c->umount_mutex);
+ up_write(&sb->s_umount);
+ sb->s_root = NULL;
+ deactivate_super(sb);
+out_close:
+ ubi_close_volume(c->ubi);
+out_free:
+ kfree(c);
+ return err;
+}
+
+static void ubifs_kill_sb(struct super_block *sb)
+{
+ struct ubifs_info *c = sb->s_fs_info;
+
+ if (sb->s_root != NULL && !(sb->s_flags & MS_RDONLY))
+ commit_on_unmount(c);
+ /* The un-mount routine is actually done in put_super() */
+ generic_shutdown_super(sb);
+}
+
+static struct file_system_type ubifs_fs_type = {
+ .name = "ubifs",
+ .owner = THIS_MODULE,
+ .get_sb = ubifs_get_sb,
+ .kill_sb = ubifs_kill_sb
+};
+
+/*
+ * Inode slab cache constructor.
+ */
+static void inode_slab_ctor(struct kmem_cache *cachep, void *obj)
+{
+ struct ubifs_inode *inode = obj;
+ inode_init_once(&inode->vfs_inode);
+}
+
+static int __init ubifs_init(void)
+{
+ int err;
+
+ BUILD_BUG_ON(sizeof(struct ubifs_ch) != 24);
+
+ /* Make sure node sizes are 8-byte aligned */
+ BUILD_BUG_ON(UBIFS_CH_SZ & 7);
+ BUILD_BUG_ON(UBIFS_INO_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_DENT_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_XENT_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_DATA_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_TRUN_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_SB_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_MST_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_REF_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_CS_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_ORPH_NODE_SZ & 7);
+
+ BUILD_BUG_ON(UBIFS_MAX_DENT_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_MAX_XENT_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_MAX_DATA_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_MAX_INO_NODE_SZ & 7);
+ BUILD_BUG_ON(UBIFS_MAX_NODE_SZ & 7);
+ BUILD_BUG_ON(MIN_WRITE_SZ & 7);
+
+ /* Check min. node size */
+ BUILD_BUG_ON(UBIFS_INO_NODE_SZ < MIN_WRITE_SZ);
+ BUILD_BUG_ON(UBIFS_DENT_NODE_SZ < MIN_WRITE_SZ);
+ BUILD_BUG_ON(UBIFS_XENT_NODE_SZ < MIN_WRITE_SZ);
+ BUILD_BUG_ON(UBIFS_TRUN_NODE_SZ < MIN_WRITE_SZ);
+
+ BUILD_BUG_ON(UBIFS_MAX_DENT_NODE_SZ > UBIFS_MAX_NODE_SZ);
+ BUILD_BUG_ON(UBIFS_MAX_XENT_NODE_SZ > UBIFS_MAX_NODE_SZ);
+ BUILD_BUG_ON(UBIFS_MAX_DATA_NODE_SZ > UBIFS_MAX_NODE_SZ);
+ BUILD_BUG_ON(UBIFS_MAX_INO_NODE_SZ > UBIFS_MAX_NODE_SZ);
+
+ /* We do not support multiple pages per block ATM */
+ BUILD_BUG_ON(UBIFS_BLOCK_SIZE != PAGE_CACHE_SIZE);
+
+ /* Defined node sizes */
+ BUILD_BUG_ON(UBIFS_SB_NODE_SZ != 4096);
+ BUILD_BUG_ON(UBIFS_MST_NODE_SZ != 512);
+ BUILD_BUG_ON(UBIFS_INO_NODE_SZ != 160);
+ BUILD_BUG_ON(UBIFS_REF_NODE_SZ != 64);
+
+ err = bdi_init(&ubifs_backing_dev_info);
+ if (err)
+ return err;
+
+ err = register_filesystem(&ubifs_fs_type);
+ if (err) {
+ ubifs_err("cannot register file system, error %d", err);
+ goto out;
+ }
+
+ err = -ENOMEM;
+ 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)
+ goto out_reg;
+
+ register_shrinker(&ubifs_shrinker_info);
+
+ err = ubifs_compressors_init();
+ if (err)
+ goto out_compr;
+
+ return 0;
+
+out_compr:
+ unregister_shrinker(&ubifs_shrinker_info);
+ kmem_cache_destroy(ubifs_inode_slab);
+out_reg:
+ unregister_filesystem(&ubifs_fs_type);
+out:
+ bdi_destroy(&ubifs_backing_dev_info);
+ return err;
+}
+/* late_initcall to let compressors initialize first */
+late_initcall(ubifs_init);
+
+static void __exit ubifs_exit(void)
+{
+ ubifs_assert(list_empty(&ubifs_infos));
+ ubifs_assert(atomic_long_read(&ubifs_clean_zn_cnt) == 0);
+
+ ubifs_compressors_exit();
+ unregister_shrinker(&ubifs_shrinker_info);
+ kmem_cache_destroy(ubifs_inode_slab);
+ unregister_filesystem(&ubifs_fs_type);
+ bdi_destroy(&ubifs_backing_dev_info);
+}
+module_exit(ubifs_exit);
+
+MODULE_LICENSE("GPL");
+MODULE_VERSION(__stringify(UBIFS_VERSION));
+MODULE_AUTHOR("Artem Bityutskiy, Adrian Hunter");
+MODULE_DESCRIPTION("UBIFS - UBI File System");
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
new file mode 100644
index 0000000..86e3e60
--- /dev/null
+++ b/fs/ubifs/super.c
@@ -0,0 +1,525 @@
+/*
+ * 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
+ */
+
+/* This file implements VFS superblock operations */
+
+#include <linux/kthread.h>
+#include <linux/seq_file.h>
+#include <linux/mount.h>
+#include "ubifs.h"
+
+/**
+ * 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->xattr_cnt < 0)
+ return 3;
+
+ if (ui->xattr_size < 0)
+ return 4;
+
+ if (ui->xattr_names < 0 ||
+ ui->xattr_names + ui->xattr_cnt > XATTR_LIST_MAX)
+ return 5;
+
+ if (ui->data_len < 0 || ui->data_len > UBIFS_MAX_INO_DATA)
+ return 6;
+
+ 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 *ubifs_iget(struct super_block *sb, unsigned long inum)
+{
+ int err;
+ union ubifs_key key;
+ struct ubifs_ino_node *ino;
+ struct ubifs_info *c = sb->s_fs_info;
+ struct inode *inode;
+ struct ubifs_inode *ui;
+
+ dbg_gen("inode %lu", inum);
+
+ inode = iget_locked(sb, inum);
+ if (!inode)
+ return ERR_PTR(-ENOMEM);
+ if (!(inode->i_state & I_NEW))
+ return inode;
+ ui = ubifs_inode(inode);
+
+ ino = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
+ if (!ino) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ ino_key_init(c, &key, inode->i_ino);
+
+ err = ubifs_tnc_lookup(c, &key, ino);
+ if (err)
+ 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);
+ 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);
+ inode->i_mtime.tv_nsec = le32_to_cpu(ino->mtime_nsec);
+ inode->i_ctime.tv_sec = (int64_t)le64_to_cpu(ino->ctime_sec);
+ inode->i_ctime.tv_nsec = le32_to_cpu(ino->ctime_nsec);
+ inode->i_mode = le32_to_cpu(ino->mode);
+ inode->i_size = le64_to_cpu(ino->size);
+
+ ui->data_len = le32_to_cpu(ino->data_len);
+ 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 = le64_to_cpu(ino->xattr_size);
+ ui->xattr_names = le32_to_cpu(ino->xattr_names);
+
+ err = validate_inode(c, inode);
+ if (err)
+ goto out_invalid;
+
+ 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->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';
+ 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;
+ }
+
+ kfree(ino);
+ ubifs_set_inode_flags(inode);
+ 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);
+ err = -EINVAL;
+out_ino:
+ kfree(ino);
+out:
+ ubifs_err("failed to read inode %lu, error %d", inode->i_ino, err);
+ iget_failed(inode);
+ return ERR_PTR(err);
+}
+
+static struct inode *ubifs_alloc_inode(struct super_block *sb)
+{
+ struct ubifs_inode *ui;
+
+ ui = kmem_cache_alloc(ubifs_inode_slab, GFP_NOFS);
+ if (!ui)
+ return NULL;
+
+ memset((void *)ui + sizeof(struct inode), 0,
+ sizeof(struct ubifs_inode) - sizeof(struct inode));
+ mutex_init(&ui->budg_mutex);
+ return &ui->vfs_inode;
+};
+
+static void ubifs_destroy_inode(struct inode *inode)
+{
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ kfree(ui->data);
+ kmem_cache_free(ubifs_inode_slab, inode);
+}
+
+static void ubifs_put_super(struct super_block *sb)
+{
+ int i;
+ struct ubifs_info *c = sb->s_fs_info;
+
+ ubifs_msg("un-mount UBI device %d, volume %d", c->vi.ubi_num,
+ c->vi.vol_id);
+ /*
+ * 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.
+ */
+ ubifs_assert(atomic_long_read(&c->dirty_pg_cnt) == 0);
+ ubifs_assert(atomic_long_read(&c->dirty_ino_cnt) == 0);
+ ubifs_assert(c->budg_idx_growth == 0);
+ ubifs_assert(c->budg_data_growth == 0);
+
+ /*
+ * 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);
+
+ spin_lock(&ubifs_infos_lock);
+ list_del(&c->infos_list);
+ spin_unlock(&ubifs_infos_lock);
+
+ if (!(c->vfs_sb->s_flags & MS_RDONLY)) {
+ /*
+ * 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;
+ }
+
+ /* Synchronize write-buffers */
+ if (c->jheads)
+ for (i = 0; i < c->jhead_cnt; i++) {
+ ubifs_wbuf_sync(&c->jheads[i].wbuf);
+ del_timer_sync(&c->jheads[i].wbuf.timer);
+ }
+
+ /*
+ * On fatal errors c->ro_media is set to 1, in which case we do
+ * not write the master node.
+ */
+ if (!c->ro_media) {
+ /*
+ * We are being cleanly unmounted which means the
+ * orphans were killed - indicate this in the master
+ * node. Also save the reserved GC LEB number.
+ */
+ int err;
+
+ 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("failed to write master node, "
+ "error %d", err);
+ }
+ }
+
+ ubifs_umount(c);
+ ubi_close_volume(c->ubi);
+ mutex_unlock(&c->umount_mutex);
+ kfree(c);
+}
+
+/*
+ * Note, Linux write-back code calls this without 'i_mutex'.
+ */
+static int ubifs_write_inode(struct inode *inode, int wait)
+{
+ int err;
+ struct ubifs_info *c = inode->i_sb->s_fs_info;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ struct ubifs_budget_req req = {.dd_growth = c->inode_budget,
+ .dirtied_ino_d = ui->data_len};
+
+ ubifs_assert(!ui->xattr);
+ if (is_bad_inode(inode))
+ return 0;
+
+ mutex_lock(&ui->budg_mutex);
+
+ /*
+ * Due to races between write-back forced by budgeting
+ * (see 'sync_some_inodes()') and pdflush write-back, the inode may
+ * have already been synchronized, do not do this again.
+ *
+ * This might also happen if it was synchronized in e.g. ubifs_link()',
+ * etc.
+ */
+ if (!ui->dirty) {
+ mutex_unlock(&ui->budg_mutex);
+ return 0;
+ }
+
+ ubifs_assert(ui->budgeted);
+ dbg_gen("inode %lu", inode->i_ino);
+
+ err = ubifs_jrn_write_inode(c, inode, 0, IS_SYNC(inode));
+ if (err)
+ ubifs_err("can't write inode %lu, error %d", inode->i_ino, err);
+
+ ui->dirty = 0;
+ UBIFS_DBG(ui->budgeted = 0);
+ atomic_long_dec(&c->dirty_ino_cnt);
+
+ ubifs_release_budget(c, &req);
+ mutex_unlock(&ui->budg_mutex);
+
+ return err;
+}
+
+static void ubifs_delete_inode(struct inode *inode)
+{
+ struct ubifs_info *c = inode->i_sb->s_fs_info;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ struct ubifs_budget_req req = {.dd_growth = c->inode_budget,
+ .dirtied_ino_d = ui->data_len};
+ int err;
+
+ 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.
+ */
+ ubifs_assert(!ui->dirty);
+ goto out;
+ }
+
+ dbg_gen("inode %lu", inode->i_ino);
+ ubifs_assert(!atomic_read(&inode->i_count));
+ ubifs_assert(inode->i_nlink == 0);
+
+ truncate_inode_pages(&inode->i_data, 0);
+ if (is_bad_inode(inode))
+ goto out;
+
+ mutex_lock(&ui->budg_mutex);
+
+ inode->i_size = 0;
+
+ err = ubifs_jrn_write_inode(c, inode, 1, IS_SYNC(inode));
+ if (err)
+ /*
+ * Worst case we have a lost orphan inode wasting space, so a
+ * simple error message is ok here.
+ */
+ ubifs_err("can't write inode %lu, error %d", inode->i_ino, err);
+
+ if (ui->dirty) {
+ ubifs_assert(ui->budgeted);
+ atomic_long_dec(&c->dirty_ino_cnt);
+ ui->dirty = 0;
+ UBIFS_DBG(ui->budgeted = 0);
+ ubifs_release_budget(c, &req);
+ }
+
+ mutex_unlock(&ui->budg_mutex);
+out:
+ clear_inode(inode);
+}
+
+static void ubifs_dirty_inode(struct inode *inode)
+{
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ ubifs_assert(mutex_is_locked(&ui->budg_mutex));
+ if (!ui->dirty) {
+ struct ubifs_info *c = inode->i_sb->s_fs_info;
+
+ ui->dirty = 1;
+ atomic_long_inc(&c->dirty_ino_cnt);
+ dbg_gen("inode %lu", inode->i_ino);
+ }
+}
+
+static int ubifs_statfs(struct dentry *dentry, struct kstatfs *buf)
+{
+ struct ubifs_info *c = dentry->d_sb->s_fs_info;
+ unsigned long long free;
+
+ free = ubifs_budg_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;
+
+ return 0;
+}
+
+static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
+{
+ int err;
+ struct ubifs_info *c = sb->s_fs_info;
+
+ dbg_gen("old flags %#lx, new flags %#x", sb->s_flags, *flags);
+
+ err = ubifs_parse_options(c, data, 1);
+ if (err) {
+ ubifs_err("invalid or unknown remount parameter");
+ return err;
+ }
+ if ((sb->s_flags & MS_RDONLY) && !(*flags & MS_RDONLY)) {
+ err = ubifs_remount_rw(c);
+ if (err)
+ return err;
+ } else if (!(sb->s_flags & MS_RDONLY) && (*flags & MS_RDONLY))
+ ubifs_remount_ro(c);
+
+ return 0;
+}
+
+static int ubifs_show_options(struct seq_file *s, struct vfsmount *mnt)
+{
+ struct ubifs_info *c = mnt->mnt_sb->s_fs_info;
+
+ if (c->mount_opts.unmount_mode == 2)
+ seq_printf(s, ",fast_unmount");
+ else if (c->mount_opts.unmount_mode == 1)
+ seq_printf(s, ",norm_unmount");
+
+ return 0;
+}
+
+static int ubifs_sync_fs(struct super_block *sb, int wait)
+{
+ struct ubifs_info *c = sb->s_fs_info;
+ int i, ret = 0, err;
+
+ if (c->jheads)
+ for (i = 0; i < c->jhead_cnt; i++) {
+ err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
+ if (err && !ret)
+ ret = err;
+ }
+ /*
+ * We ought to call sync for c->ubi but it does not have one. If it had
+ * it would in turn call mtd->sync, however mtd operations are
+ * synchronous anyway, so we don't lose any sleep here.
+ */
+ return ret;
+}
+
+struct super_operations ubifs_super_operations = {
+ .alloc_inode = ubifs_alloc_inode,
+ .destroy_inode = ubifs_destroy_inode,
+ .put_super = ubifs_put_super,
+ .write_inode = ubifs_write_inode,
+ .delete_inode = ubifs_delete_inode,
+ .statfs = ubifs_statfs,
+ .dirty_inode = ubifs_dirty_inode,
+ .remount_fs = ubifs_remount_fs,
+ .show_options = ubifs_show_options,
+ .sync_fs = ubifs_sync_fs,
+};
--
1.5.4.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/