summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorStefan Roese <sr@denx.de>2009-03-19 15:35:05 +0100
committerWolfgang Denk <wd@denx.de>2009-03-20 22:39:15 +0100
commit9eefe2a2b37a838558e3d213a9f5519503d0c180 (patch)
treecd6ce359f020615d685d353fc3a20e7e07ac05fc
parentb1b4e89a0f3b75854c39a62cae41bad56d210adf (diff)
downloadu-boot-imx-9eefe2a2b37a838558e3d213a9f5519503d0c180.zip
u-boot-imx-9eefe2a2b37a838558e3d213a9f5519503d0c180.tar.gz
u-boot-imx-9eefe2a2b37a838558e3d213a9f5519503d0c180.tar.bz2
UBIFS: Implement read-only UBIFS support in U-Boot
The U-Boot UBIFS implementation is largely a direct copy from the current Linux version (2.6.29-rc6). As already done in the UBI version we have an "abstraction layer" to redefine or remove some OS calls (e.g. mutex_lock() ...). This makes it possible to use the original Linux code with very little changes. And by this we can better update to later Linux versions. I removed some of the Linux features that are not used in the U-Boot version (e.g. garbage-collection, write support). Signed-off-by: Stefan Roese <sr@denx.de> CC: Artem Bityutskiy <dedekind@infradead.org> CC: Adrian Hunter <ext-Adrian.Hunter@nokia.com>
-rw-r--r--Makefile3
-rw-r--r--fs/Makefile1
-rw-r--r--fs/ubifs/Makefile52
-rw-r--r--fs/ubifs/budget.c113
-rw-r--r--fs/ubifs/crc16.c60
-rw-r--r--fs/ubifs/crc16.h30
-rw-r--r--fs/ubifs/debug.c156
-rw-r--r--fs/ubifs/debug.h392
-rw-r--r--fs/ubifs/io.c316
-rw-r--r--fs/ubifs/key.h557
-rw-r--r--fs/ubifs/log.c104
-rw-r--r--fs/ubifs/lprops.c842
-rw-r--r--fs/ubifs/lpt.c1105
-rw-r--r--fs/ubifs/lpt_commit.c171
-rw-r--r--fs/ubifs/master.c341
-rw-r--r--fs/ubifs/misc.h310
-rw-r--r--fs/ubifs/orphan.c316
-rw-r--r--fs/ubifs/recovery.c1249
-rw-r--r--fs/ubifs/replay.c1070
-rw-r--r--fs/ubifs/sb.c324
-rw-r--r--fs/ubifs/scan.c362
-rw-r--r--fs/ubifs/super.c1189
-rw-r--r--fs/ubifs/tnc.c2767
-rw-r--r--fs/ubifs/tnc_commit.c1102
-rw-r--r--fs/ubifs/tnc_misc.c435
-rw-r--r--fs/ubifs/ubifs-media.h751
-rw-r--r--fs/ubifs/ubifs.c684
-rw-r--r--fs/ubifs/ubifs.h2173
-rw-r--r--include/linux/math64.h85
-rw-r--r--include/ubi_uboot.h9
30 files changed, 17067 insertions, 2 deletions
diff --git a/Makefile b/Makefile
index 199a224..215ee00 100644
--- a/Makefile
+++ b/Makefile
@@ -222,7 +222,8 @@ LIBS += cpu/ixp/npe/libnpe.a
endif
LIBS += lib_$(ARCH)/lib$(ARCH).a
LIBS += fs/cramfs/libcramfs.a fs/fat/libfat.a fs/fdos/libfdos.a fs/jffs2/libjffs2.a \
- fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a fs/yaffs2/libyaffs2.a
+ fs/reiserfs/libreiserfs.a fs/ext2/libext2fs.a fs/yaffs2/libyaffs2.a \
+ fs/ubifs/libubifs.a
LIBS += net/libnet.a
LIBS += disk/libdisk.a
LIBS += drivers/bios_emulator/libatibiosemu.a
diff --git a/fs/Makefile b/fs/Makefile
index 8bbd563..22aad12 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -29,6 +29,7 @@ subdirs-$(CONFIG_CMD_FDOS) += fdos
subdirs-$(CONFIG_CMD_JFFS2) += jffs2
subdirs-$(CONFIG_CMD_REISER) += reiserfs
subdirs-$(CONFIG_YAFFS2) += yaffs2
+subdirs-$(CONFIG_CMD_UBIFS) += ubifs
SUBDIRS := $(subdirs-y)
diff --git a/fs/ubifs/Makefile b/fs/ubifs/Makefile
new file mode 100644
index 0000000..8328843
--- /dev/null
+++ b/fs/ubifs/Makefile
@@ -0,0 +1,52 @@
+#
+# (C) Copyright 2006
+# Wolfgang Denk, DENX Software Engineering, wd@denx.de.
+#
+# (C) Copyright 2003
+# Pavel Bartusek, Sysgo Real-Time Solutions AG, pba@sysgo.de
+#
+#
+# See file CREDITS for list of people who contributed to this
+# project.
+#
+# This program is free software; you can redistribute it and/or
+# modify it under the terms of the GNU General Public License as
+# published by the Free Software Foundation; either version 2 of
+# the License, or (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# 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., 59 Temple Place, Suite 330, Boston,
+# MA 02111-1307 USA
+#
+
+include $(TOPDIR)/config.mk
+
+LIB = $(obj)libubifs.a
+
+COBJS-$(CONFIG_CMD_UBIFS) := ubifs.o io.o super.o sb.o master.o lpt.o
+COBJS-$(CONFIG_CMD_UBIFS) += lpt_commit.o scan.o lprops.o
+COBJS-$(CONFIG_CMD_UBIFS) += tnc.o tnc_misc.o debug.o crc16.o budget.o
+COBJS-$(CONFIG_CMD_UBIFS) += log.o orphan.o recovery.o replay.o
+
+SRCS := $(AOBJS:.o=.S) $(COBJS-y:.o=.c)
+OBJS := $(addprefix $(obj),$(AOBJS) $(COBJS-y))
+
+all: $(LIB) $(AOBJS)
+
+$(LIB): $(obj).depend $(OBJS)
+ $(AR) $(ARFLAGS) $@ $(OBJS)
+
+#########################################################################
+
+# defines $(obj).depend target
+include $(SRCTREE)/rules.mk
+
+sinclude $(obj).depend
+
+#########################################################################
diff --git a/fs/ubifs/budget.c b/fs/ubifs/budget.c
new file mode 100644
index 0000000..85377ea
--- /dev/null
+++ b/fs/ubifs/budget.c
@@ -0,0 +1,113 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements the budgeting sub-system which is responsible for UBIFS
+ * space management.
+ *
+ * Factors such as compression, wasted space at the ends of LEBs, space in other
+ * journal heads, the effect of updates on the index, and so on, make it
+ * impossible to accurately predict the amount of space needed. Consequently
+ * approximations are used.
+ */
+
+#include "ubifs.h"
+#include <linux/math64.h>
+
+/**
+ * ubifs_calc_min_idx_lebs - calculate amount of eraseblocks for the index.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates and returns the number of eraseblocks which should
+ * be kept for index usage.
+ */
+int ubifs_calc_min_idx_lebs(struct ubifs_info *c)
+{
+ int idx_lebs, eff_leb_size = c->leb_size - c->max_idx_node_sz;
+ long long idx_size;
+
+ idx_size = c->old_idx_sz + c->budg_idx_growth + c->budg_uncommitted_idx;
+
+ /* And make sure we have thrice the index size of space reserved */
+ idx_size = idx_size + (idx_size << 1);
+
+ /*
+ * We do not maintain 'old_idx_size' as 'old_idx_lebs'/'old_idx_bytes'
+ * pair, nor similarly the two variables for the new index size, so we
+ * have to do this costly 64-bit division on fast-path.
+ */
+ idx_size += eff_leb_size - 1;
+ idx_lebs = div_u64(idx_size, eff_leb_size);
+ /*
+ * The index head is not available for the in-the-gaps method, so add an
+ * extra LEB to compensate.
+ */
+ idx_lebs += 1;
+ if (idx_lebs < MIN_INDEX_LEBS)
+ idx_lebs = MIN_INDEX_LEBS;
+ return idx_lebs;
+}
+
+/**
+ * ubifs_reported_space - calculate reported free space.
+ * @c: the UBIFS file-system description object
+ * @free: amount of free space
+ *
+ * This function calculates amount of free space which will be reported to
+ * user-space. User-space application tend to expect that if the file-system
+ * (e.g., via the 'statfs()' call) reports that it has N bytes available, they
+ * are able to write a file of size N. UBIFS attaches node headers to each data
+ * node and it has to write indexing nodes as well. This introduces additional
+ * overhead, and UBIFS has to report slightly less free space to meet the above
+ * expectations.
+ *
+ * This function assumes free space is made up of uncompressed data nodes and
+ * full index nodes (one per data node, tripled because we always allow enough
+ * space to write the index thrice).
+ *
+ * Note, the calculation is pessimistic, which means that most of the time
+ * UBIFS reports less space than it actually has.
+ */
+long long ubifs_reported_space(const struct ubifs_info *c, long long free)
+{
+ int divisor, factor, f;
+
+ /*
+ * Reported space size is @free * X, where X is UBIFS block size
+ * divided by UBIFS block size + all overhead one data block
+ * introduces. The overhead is the node header + indexing overhead.
+ *
+ * Indexing overhead calculations are based on the following formula:
+ * I = N/(f - 1) + 1, where I - number of indexing nodes, N - number
+ * of data nodes, f - fanout. Because effective UBIFS fanout is twice
+ * as less than maximum fanout, we assume that each data node
+ * introduces 3 * @c->max_idx_node_sz / (@c->fanout/2 - 1) bytes.
+ * Note, the multiplier 3 is because UBIFS reserves thrice as more space
+ * for the index.
+ */
+ f = c->fanout > 3 ? c->fanout >> 1 : 2;
+ factor = UBIFS_BLOCK_SIZE;
+ divisor = UBIFS_MAX_DATA_NODE_SZ;
+ divisor += (c->max_idx_node_sz * 3) / (f - 1);
+ free *= factor;
+ return div_u64(free, divisor);
+}
diff --git a/fs/ubifs/crc16.c b/fs/ubifs/crc16.c
new file mode 100644
index 0000000..443ccf8
--- /dev/null
+++ b/fs/ubifs/crc16.c
@@ -0,0 +1,60 @@
+/*
+ * crc16.c
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2. See the file COPYING for more details.
+ */
+
+#include <linux/types.h>
+#include "crc16.h"
+
+/** CRC table for the CRC-16. The poly is 0x8005 (x^16 + x^15 + x^2 + 1) */
+u16 const crc16_table[256] = {
+ 0x0000, 0xC0C1, 0xC181, 0x0140, 0xC301, 0x03C0, 0x0280, 0xC241,
+ 0xC601, 0x06C0, 0x0780, 0xC741, 0x0500, 0xC5C1, 0xC481, 0x0440,
+ 0xCC01, 0x0CC0, 0x0D80, 0xCD41, 0x0F00, 0xCFC1, 0xCE81, 0x0E40,
+ 0x0A00, 0xCAC1, 0xCB81, 0x0B40, 0xC901, 0x09C0, 0x0880, 0xC841,
+ 0xD801, 0x18C0, 0x1980, 0xD941, 0x1B00, 0xDBC1, 0xDA81, 0x1A40,
+ 0x1E00, 0xDEC1, 0xDF81, 0x1F40, 0xDD01, 0x1DC0, 0x1C80, 0xDC41,
+ 0x1400, 0xD4C1, 0xD581, 0x1540, 0xD701, 0x17C0, 0x1680, 0xD641,
+ 0xD201, 0x12C0, 0x1380, 0xD341, 0x1100, 0xD1C1, 0xD081, 0x1040,
+ 0xF001, 0x30C0, 0x3180, 0xF141, 0x3300, 0xF3C1, 0xF281, 0x3240,
+ 0x3600, 0xF6C1, 0xF781, 0x3740, 0xF501, 0x35C0, 0x3480, 0xF441,
+ 0x3C00, 0xFCC1, 0xFD81, 0x3D40, 0xFF01, 0x3FC0, 0x3E80, 0xFE41,
+ 0xFA01, 0x3AC0, 0x3B80, 0xFB41, 0x3900, 0xF9C1, 0xF881, 0x3840,
+ 0x2800, 0xE8C1, 0xE981, 0x2940, 0xEB01, 0x2BC0, 0x2A80, 0xEA41,
+ 0xEE01, 0x2EC0, 0x2F80, 0xEF41, 0x2D00, 0xEDC1, 0xEC81, 0x2C40,
+ 0xE401, 0x24C0, 0x2580, 0xE541, 0x2700, 0xE7C1, 0xE681, 0x2640,
+ 0x2200, 0xE2C1, 0xE381, 0x2340, 0xE101, 0x21C0, 0x2080, 0xE041,
+ 0xA001, 0x60C0, 0x6180, 0xA141, 0x6300, 0xA3C1, 0xA281, 0x6240,
+ 0x6600, 0xA6C1, 0xA781, 0x6740, 0xA501, 0x65C0, 0x6480, 0xA441,
+ 0x6C00, 0xACC1, 0xAD81, 0x6D40, 0xAF01, 0x6FC0, 0x6E80, 0xAE41,
+ 0xAA01, 0x6AC0, 0x6B80, 0xAB41, 0x6900, 0xA9C1, 0xA881, 0x6840,
+ 0x7800, 0xB8C1, 0xB981, 0x7940, 0xBB01, 0x7BC0, 0x7A80, 0xBA41,
+ 0xBE01, 0x7EC0, 0x7F80, 0xBF41, 0x7D00, 0xBDC1, 0xBC81, 0x7C40,
+ 0xB401, 0x74C0, 0x7580, 0xB541, 0x7700, 0xB7C1, 0xB681, 0x7640,
+ 0x7200, 0xB2C1, 0xB381, 0x7340, 0xB101, 0x71C0, 0x7080, 0xB041,
+ 0x5000, 0x90C1, 0x9181, 0x5140, 0x9301, 0x53C0, 0x5280, 0x9241,
+ 0x9601, 0x56C0, 0x5780, 0x9741, 0x5500, 0x95C1, 0x9481, 0x5440,
+ 0x9C01, 0x5CC0, 0x5D80, 0x9D41, 0x5F00, 0x9FC1, 0x9E81, 0x5E40,
+ 0x5A00, 0x9AC1, 0x9B81, 0x5B40, 0x9901, 0x59C0, 0x5880, 0x9841,
+ 0x8801, 0x48C0, 0x4980, 0x8941, 0x4B00, 0x8BC1, 0x8A81, 0x4A40,
+ 0x4E00, 0x8EC1, 0x8F81, 0x4F40, 0x8D01, 0x4DC0, 0x4C80, 0x8C41,
+ 0x4400, 0x84C1, 0x8581, 0x4540, 0x8701, 0x47C0, 0x4680, 0x8641,
+ 0x8201, 0x42C0, 0x4380, 0x8341, 0x4100, 0x81C1, 0x8081, 0x4040
+};
+
+/**
+ * crc16 - compute the CRC-16 for the data buffer
+ * @crc: previous CRC value
+ * @buffer: data pointer
+ * @len: number of bytes in the buffer
+ *
+ * Returns the updated CRC value.
+ */
+u16 crc16(u16 crc, u8 const *buffer, size_t len)
+{
+ while (len--)
+ crc = crc16_byte(crc, *buffer++);
+ return crc;
+}
diff --git a/fs/ubifs/crc16.h b/fs/ubifs/crc16.h
new file mode 100644
index 0000000..9443c08
--- /dev/null
+++ b/fs/ubifs/crc16.h
@@ -0,0 +1,30 @@
+/*
+ * crc16.h - CRC-16 routine
+ *
+ * Implements the standard CRC-16:
+ * Width 16
+ * Poly 0x8005 (x^16 + x^15 + x^2 + 1)
+ * Init 0
+ *
+ * Copyright (c) 2005 Ben Gardner <bgardner@wabtec.com>
+ *
+ * This source code is licensed under the GNU General Public License,
+ * Version 2. See the file COPYING for more details.
+ */
+
+#ifndef __CRC16_H
+#define __CRC16_H
+
+#include <linux/types.h>
+
+extern u16 const crc16_table[256];
+
+extern u16 crc16(u16 crc, const u8 *buffer, size_t len);
+
+static inline u16 crc16_byte(u16 crc, const u8 data)
+{
+ return (crc >> 8) ^ crc16_table[(crc ^ data) & 0xff];
+}
+
+#endif /* __CRC16_H */
+
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
new file mode 100644
index 0000000..6afb883
--- /dev/null
+++ b/fs/ubifs/debug.c
@@ -0,0 +1,156 @@
+/*
+ * 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 most of the debugging stuff which is compiled in only
+ * when it is enabled. But some debugging check functions are implemented in
+ * corresponding subsystem, just because they are closely related and utilize
+ * various local functions of those subsystems.
+ */
+
+#define UBIFS_DBG_PRESERVE_UBI
+
+#include "ubifs.h"
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+
+DEFINE_SPINLOCK(dbg_lock);
+
+static char dbg_key_buf0[128];
+static char dbg_key_buf1[128];
+
+unsigned int ubifs_msg_flags = UBIFS_MSG_FLAGS_DEFAULT;
+unsigned int ubifs_chk_flags = UBIFS_CHK_FLAGS_DEFAULT;
+unsigned int ubifs_tst_flags;
+
+module_param_named(debug_msgs, ubifs_msg_flags, uint, S_IRUGO | S_IWUSR);
+module_param_named(debug_chks, ubifs_chk_flags, uint, S_IRUGO | S_IWUSR);
+module_param_named(debug_tsts, ubifs_tst_flags, uint, S_IRUGO | S_IWUSR);
+
+MODULE_PARM_DESC(debug_msgs, "Debug message type flags");
+MODULE_PARM_DESC(debug_chks, "Debug check flags");
+MODULE_PARM_DESC(debug_tsts, "Debug special test flags");
+
+static const char *get_key_type(int type)
+{
+ switch (type) {
+ case UBIFS_INO_KEY:
+ return "inode";
+ case UBIFS_DENT_KEY:
+ return "direntry";
+ case UBIFS_XENT_KEY:
+ return "xentry";
+ case UBIFS_DATA_KEY:
+ return "data";
+ case UBIFS_TRUN_KEY:
+ return "truncate";
+ default:
+ return "unknown/invalid key";
+ }
+}
+
+static void sprintf_key(const struct ubifs_info *c, const union ubifs_key *key,
+ char *buffer)
+{
+ char *p = buffer;
+ int type = key_type(c, key);
+
+ if (c->key_fmt == UBIFS_SIMPLE_KEY_FMT) {
+ switch (type) {
+ case UBIFS_INO_KEY:
+ sprintf(p, "(%lu, %s)", (unsigned long)key_inum(c, key),
+ get_key_type(type));
+ break;
+ case UBIFS_DENT_KEY:
+ case UBIFS_XENT_KEY:
+ sprintf(p, "(%lu, %s, %#08x)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type), key_hash(c, key));
+ break;
+ case UBIFS_DATA_KEY:
+ sprintf(p, "(%lu, %s, %u)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type), key_block(c, key));
+ break;
+ case UBIFS_TRUN_KEY:
+ sprintf(p, "(%lu, %s)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type));
+ break;
+ default:
+ sprintf(p, "(bad key type: %#08x, %#08x)",
+ key->u32[0], key->u32[1]);
+ }
+ } else
+ sprintf(p, "bad key format %d", c->key_fmt);
+}
+
+const char *dbg_key_str0(const struct ubifs_info *c, const union ubifs_key *key)
+{
+ /* dbg_lock must be held */
+ sprintf_key(c, key, dbg_key_buf0);
+ return dbg_key_buf0;
+}
+
+const char *dbg_key_str1(const struct ubifs_info *c, const union ubifs_key *key)
+{
+ /* dbg_lock must be held */
+ sprintf_key(c, key, dbg_key_buf1);
+ return dbg_key_buf1;
+}
+
+/**
+ * ubifs_debugging_init - initialize UBIFS debugging.
+ * @c: UBIFS file-system description object
+ *
+ * This function initializes debugging-related data for the file system.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_debugging_init(struct ubifs_info *c)
+{
+ c->dbg = kzalloc(sizeof(struct ubifs_debug_info), GFP_KERNEL);
+ if (!c->dbg)
+ return -ENOMEM;
+
+ c->dbg->buf = vmalloc(c->leb_size);
+ if (!c->dbg->buf)
+ goto out;
+
+ return 0;
+
+out:
+ kfree(c->dbg);
+ return -ENOMEM;
+}
+
+/**
+ * ubifs_debugging_exit - free debugging data.
+ * @c: UBIFS file-system description object
+ */
+void ubifs_debugging_exit(struct ubifs_info *c)
+{
+ vfree(c->dbg->buf);
+ kfree(c->dbg);
+}
+
+#endif /* CONFIG_UBIFS_FS_DEBUG */
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
new file mode 100644
index 0000000..62617b6
--- /dev/null
+++ b/fs/ubifs/debug.h
@@ -0,0 +1,392 @@
+/*
+ * 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
+ */
+
+#ifndef __UBIFS_DEBUG_H__
+#define __UBIFS_DEBUG_H__
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+
+/**
+ * ubifs_debug_info - per-FS debugging information.
+ * @buf: a buffer of LEB size, used for various purposes
+ * @old_zroot: old index root - used by 'dbg_check_old_index()'
+ * @old_zroot_level: old index root level - used by 'dbg_check_old_index()'
+ * @old_zroot_sqnum: old index root sqnum - used by 'dbg_check_old_index()'
+ * @failure_mode: failure mode for recovery testing
+ * @fail_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls
+ * @fail_timeout: time in jiffies when delay of failure mode expires
+ * @fail_cnt: current number of calls to failure mode I/O functions
+ * @fail_cnt_max: number of calls by which to delay failure mode
+ * @chk_lpt_sz: used by LPT tree size checker
+ * @chk_lpt_sz2: used by LPT tree size checker
+ * @chk_lpt_wastage: used by LPT tree size checker
+ * @chk_lpt_lebs: used by LPT tree size checker
+ * @new_nhead_offs: used by LPT tree size checker
+ * @new_ihead_lnum: used by debugging to check @c->ihead_lnum
+ * @new_ihead_offs: used by debugging to check @c->ihead_offs
+ *
+ * @saved_lst: saved lprops statistics (used by 'dbg_save_space_info()')
+ * @saved_free: saved free space (used by 'dbg_save_space_info()')
+ *
+ * dfs_dir_name: name of debugfs directory containing this file-system's files
+ * dfs_dir: direntry object of the file-system debugfs directory
+ * dfs_dump_lprops: "dump lprops" debugfs knob
+ * dfs_dump_budg: "dump budgeting information" debugfs knob
+ * dfs_dump_tnc: "dump TNC" debugfs knob
+ */
+struct ubifs_debug_info {
+ void *buf;
+ struct ubifs_zbranch old_zroot;
+ int old_zroot_level;
+ unsigned long long old_zroot_sqnum;
+ int failure_mode;
+ int fail_delay;
+ unsigned long fail_timeout;
+ unsigned int fail_cnt;
+ unsigned int fail_cnt_max;
+ long long chk_lpt_sz;
+ long long chk_lpt_sz2;
+ long long chk_lpt_wastage;
+ int chk_lpt_lebs;
+ int new_nhead_offs;
+ int new_ihead_lnum;
+ int new_ihead_offs;
+
+ struct ubifs_lp_stats saved_lst;
+ long long saved_free;
+
+ char dfs_dir_name[100];
+ struct dentry *dfs_dir;
+ struct dentry *dfs_dump_lprops;
+ struct dentry *dfs_dump_budg;
+ struct dentry *dfs_dump_tnc;
+};
+
+#define UBIFS_DBG(op) op
+
+#define ubifs_assert(expr) do { \
+ if (unlikely(!(expr))) { \
+ printk(KERN_CRIT "UBIFS assert failed in %s at %u (pid %d)\n", \
+ __func__, __LINE__, 0); \
+ dbg_dump_stack(); \
+ } \
+} while (0)
+
+#define ubifs_assert_cmt_locked(c) do { \
+ if (unlikely(down_write_trylock(&(c)->commit_sem))) { \
+ up_write(&(c)->commit_sem); \
+ printk(KERN_CRIT "commit lock is not locked!\n"); \
+ ubifs_assert(0); \
+ } \
+} while (0)
+
+#define dbg_dump_stack() do { \
+ if (!dbg_failure_mode) \
+ dump_stack(); \
+} while (0)
+
+/* Generic debugging messages */
+#define dbg_msg(fmt, ...) do { \
+ spin_lock(&dbg_lock); \
+ printk(KERN_DEBUG "UBIFS DBG (pid %d): %s: " fmt "\n", 0, \
+ __func__, ##__VA_ARGS__); \
+ spin_unlock(&dbg_lock); \
+} while (0)
+
+#define dbg_do_msg(typ, fmt, ...) do { \
+ if (ubifs_msg_flags & typ) \
+ dbg_msg(fmt, ##__VA_ARGS__); \
+} while (0)
+
+#define dbg_err(fmt, ...) do { \
+ spin_lock(&dbg_lock); \
+ ubifs_err(fmt, ##__VA_ARGS__); \
+ spin_unlock(&dbg_lock); \
+} while (0)
+
+const char *dbg_key_str0(const struct ubifs_info *c,
+ const union ubifs_key *key);
+const char *dbg_key_str1(const struct ubifs_info *c,
+ const union ubifs_key *key);
+
+/*
+ * DBGKEY macros require @dbg_lock to be held, which it is in the dbg message
+ * macros.
+ */
+#define DBGKEY(key) dbg_key_str0(c, (key))
+#define DBGKEY1(key) dbg_key_str1(c, (key))
+
+/* General messages */
+#define dbg_gen(fmt, ...) dbg_do_msg(UBIFS_MSG_GEN, fmt, ##__VA_ARGS__)
+
+/* Additional journal messages */
+#define dbg_jnl(fmt, ...) dbg_do_msg(UBIFS_MSG_JNL, fmt, ##__VA_ARGS__)
+
+/* Additional TNC messages */
+#define dbg_tnc(fmt, ...) dbg_do_msg(UBIFS_MSG_TNC, fmt, ##__VA_ARGS__)
+
+/* Additional lprops messages */
+#define dbg_lp(fmt, ...) dbg_do_msg(UBIFS_MSG_LP, fmt, ##__VA_ARGS__)
+
+/* Additional LEB find messages */
+#define dbg_find(fmt, ...) dbg_do_msg(UBIFS_MSG_FIND, fmt, ##__VA_ARGS__)
+
+/* Additional mount messages */
+#define dbg_mnt(fmt, ...) dbg_do_msg(UBIFS_MSG_MNT, fmt, ##__VA_ARGS__)
+
+/* Additional I/O messages */
+#define dbg_io(fmt, ...) dbg_do_msg(UBIFS_MSG_IO, fmt, ##__VA_ARGS__)
+
+/* Additional commit messages */
+#define dbg_cmt(fmt, ...) dbg_do_msg(UBIFS_MSG_CMT, fmt, ##__VA_ARGS__)
+
+/* Additional budgeting messages */
+#define dbg_budg(fmt, ...) dbg_do_msg(UBIFS_MSG_BUDG, fmt, ##__VA_ARGS__)
+
+/* Additional log messages */
+#define dbg_log(fmt, ...) dbg_do_msg(UBIFS_MSG_LOG, fmt, ##__VA_ARGS__)
+
+/* Additional gc messages */
+#define dbg_gc(fmt, ...) dbg_do_msg(UBIFS_MSG_GC, fmt, ##__VA_ARGS__)
+
+/* Additional scan messages */
+#define dbg_scan(fmt, ...) dbg_do_msg(UBIFS_MSG_SCAN, fmt, ##__VA_ARGS__)
+
+/* Additional recovery messages */
+#define dbg_rcvry(fmt, ...) dbg_do_msg(UBIFS_MSG_RCVRY, fmt, ##__VA_ARGS__)
+
+/*
+ * Debugging message type flags (must match msg_type_names in debug.c).
+ *
+ * UBIFS_MSG_GEN: general messages
+ * UBIFS_MSG_JNL: journal messages
+ * UBIFS_MSG_MNT: mount messages
+ * UBIFS_MSG_CMT: commit messages
+ * UBIFS_MSG_FIND: LEB find messages
+ * UBIFS_MSG_BUDG: budgeting messages
+ * UBIFS_MSG_GC: garbage collection messages
+ * UBIFS_MSG_TNC: TNC messages
+ * UBIFS_MSG_LP: lprops messages
+ * UBIFS_MSG_IO: I/O messages
+ * UBIFS_MSG_LOG: log messages
+ * UBIFS_MSG_SCAN: scan messages
+ * UBIFS_MSG_RCVRY: recovery messages
+ */
+enum {
+ UBIFS_MSG_GEN = 0x1,
+ UBIFS_MSG_JNL = 0x2,
+ UBIFS_MSG_MNT = 0x4,
+ UBIFS_MSG_CMT = 0x8,
+ UBIFS_MSG_FIND = 0x10,
+ UBIFS_MSG_BUDG = 0x20,
+ UBIFS_MSG_GC = 0x40,
+ UBIFS_MSG_TNC = 0x80,
+ UBIFS_MSG_LP = 0x100,
+ UBIFS_MSG_IO = 0x200,
+ UBIFS_MSG_LOG = 0x400,
+ UBIFS_MSG_SCAN = 0x800,
+ UBIFS_MSG_RCVRY = 0x1000,
+};
+
+/* Debugging message type flags for each default debug message level */
+#define UBIFS_MSG_LVL_0 0
+#define UBIFS_MSG_LVL_1 0x1
+#define UBIFS_MSG_LVL_2 0x7f
+#define UBIFS_MSG_LVL_3 0xffff
+
+/*
+ * Debugging check flags (must match chk_names in debug.c).
+ *
+ * UBIFS_CHK_GEN: general checks
+ * UBIFS_CHK_TNC: check TNC
+ * UBIFS_CHK_IDX_SZ: check index size
+ * UBIFS_CHK_ORPH: check orphans
+ * UBIFS_CHK_OLD_IDX: check the old index
+ * UBIFS_CHK_LPROPS: check lprops
+ * UBIFS_CHK_FS: check the file-system
+ */
+enum {
+ UBIFS_CHK_GEN = 0x1,
+ UBIFS_CHK_TNC = 0x2,
+ UBIFS_CHK_IDX_SZ = 0x4,
+ UBIFS_CHK_ORPH = 0x8,
+ UBIFS_CHK_OLD_IDX = 0x10,
+ UBIFS_CHK_LPROPS = 0x20,
+ UBIFS_CHK_FS = 0x40,
+};
+
+/*
+ * Special testing flags (must match tst_names in debug.c).
+ *
+ * UBIFS_TST_FORCE_IN_THE_GAPS: force the use of in-the-gaps method
+ * UBIFS_TST_RCVRY: failure mode for recovery testing
+ */
+enum {
+ UBIFS_TST_FORCE_IN_THE_GAPS = 0x2,
+ UBIFS_TST_RCVRY = 0x4,
+};
+
+#if CONFIG_UBIFS_FS_DEBUG_MSG_LVL == 1
+#define UBIFS_MSG_FLAGS_DEFAULT UBIFS_MSG_LVL_1
+#elif CONFIG_UBIFS_FS_DEBUG_MSG_LVL == 2
+#define UBIFS_MSG_FLAGS_DEFAULT UBIFS_MSG_LVL_2
+#elif CONFIG_UBIFS_FS_DEBUG_MSG_LVL == 3
+#define UBIFS_MSG_FLAGS_DEFAULT UBIFS_MSG_LVL_3
+#else
+#define UBIFS_MSG_FLAGS_DEFAULT UBIFS_MSG_LVL_0
+#endif
+
+#ifdef CONFIG_UBIFS_FS_DEBUG_CHKS
+#define UBIFS_CHK_FLAGS_DEFAULT 0xffffffff
+#else
+#define UBIFS_CHK_FLAGS_DEFAULT 0
+#endif
+
+#define dbg_ntype(type) ""
+#define dbg_cstate(cmt_state) ""
+#define dbg_get_key_dump(c, key) ({})
+#define dbg_dump_inode(c, inode) ({})
+#define dbg_dump_node(c, node) ({})
+#define dbg_dump_budget_req(req) ({})
+#define dbg_dump_lstats(lst) ({})
+#define dbg_dump_budg(c) ({})
+#define dbg_dump_lprop(c, lp) ({})
+#define dbg_dump_lprops(c) ({})
+#define dbg_dump_lpt_info(c) ({})
+#define dbg_dump_leb(c, lnum) ({})
+#define dbg_dump_znode(c, znode) ({})
+#define dbg_dump_heap(c, heap, cat) ({})
+#define dbg_dump_pnode(c, pnode, parent, iip) ({})
+#define dbg_dump_tnc(c) ({})
+#define dbg_dump_index(c) ({})
+
+#define dbg_walk_index(c, leaf_cb, znode_cb, priv) 0
+#define dbg_old_index_check_init(c, zroot) 0
+#define dbg_check_old_index(c, zroot) 0
+#define dbg_check_cats(c) 0
+#define dbg_check_ltab(c) 0
+#define dbg_chk_lpt_free_spc(c) 0
+#define dbg_chk_lpt_sz(c, action, len) 0
+#define dbg_check_synced_i_size(inode) 0
+#define dbg_check_dir_size(c, dir) 0
+#define dbg_check_tnc(c, x) 0
+#define dbg_check_idx_size(c, idx_size) 0
+#define dbg_check_filesystem(c) 0
+#define dbg_check_heap(c, heap, cat, add_pos) ({})
+#define dbg_check_lprops(c) 0
+#define dbg_check_lpt_nodes(c, cnode, row, col) 0
+#define dbg_force_in_the_gaps_enabled 0
+#define dbg_force_in_the_gaps() 0
+#define dbg_failure_mode 0
+#define dbg_failure_mode_registration(c) ({})
+#define dbg_failure_mode_deregistration(c) ({})
+
+int ubifs_debugging_init(struct ubifs_info *c);
+void ubifs_debugging_exit(struct ubifs_info *c);
+
+#else /* !CONFIG_UBIFS_FS_DEBUG */
+
+#define UBIFS_DBG(op)
+
+/* Use "if (0)" to make compiler check arguments even if debugging is off */
+#define ubifs_assert(expr) do { \
+ if (0 && (expr)) \
+ printk(KERN_CRIT "UBIFS assert failed in %s at %u (pid %d)\n", \
+ __func__, __LINE__, 0); \
+} while (0)
+
+#define dbg_err(fmt, ...) do { \
+ if (0) \
+ ubifs_err(fmt, ##__VA_ARGS__); \
+} while (0)
+
+#define dbg_msg(fmt, ...) do { \
+ if (0) \
+ printk(KERN_DEBUG "UBIFS DBG (pid %d): %s: " fmt "\n", \
+ 0, __func__, ##__VA_ARGS__); \
+} while (0)
+
+#define dbg_dump_stack()
+#define ubifs_assert_cmt_locked(c)
+
+#define dbg_gen(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_jnl(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_tnc(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_lp(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_find(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_mnt(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_io(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_cmt(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_budg(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_log(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_gc(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_scan(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+#define dbg_rcvry(fmt, ...) dbg_msg(fmt, ##__VA_ARGS__)
+
+#define DBGKEY(key) ((char *)(key))
+#define DBGKEY1(key) ((char *)(key))
+
+#define ubifs_debugging_init(c) 0
+#define ubifs_debugging_exit(c) ({})
+
+#define dbg_ntype(type) ""
+#define dbg_cstate(cmt_state) ""
+#define dbg_get_key_dump(c, key) ({})
+#define dbg_dump_inode(c, inode) ({})
+#define dbg_dump_node(c, node) ({})
+#define dbg_dump_budget_req(req) ({})
+#define dbg_dump_lstats(lst) ({})
+#define dbg_dump_budg(c) ({})
+#define dbg_dump_lprop(c, lp) ({})
+#define dbg_dump_lprops(c) ({})
+#define dbg_dump_lpt_info(c) ({})
+#define dbg_dump_leb(c, lnum) ({})
+#define dbg_dump_znode(c, znode) ({})
+#define dbg_dump_heap(c, heap, cat) ({})
+#define dbg_dump_pnode(c, pnode, parent, iip) ({})
+#define dbg_dump_tnc(c) ({})
+#define dbg_dump_index(c) ({})
+
+#define dbg_walk_index(c, leaf_cb, znode_cb, priv) 0
+#define dbg_old_index_check_init(c, zroot) 0
+#define dbg_check_old_index(c, zroot) 0
+#define dbg_check_cats(c) 0
+#define dbg_check_ltab(c) 0
+#define dbg_chk_lpt_free_spc(c) 0
+#define dbg_chk_lpt_sz(c, action, len) 0
+#define dbg_check_synced_i_size(inode) 0
+#define dbg_check_dir_size(c, dir) 0
+#define dbg_check_tnc(c, x) 0
+#define dbg_check_idx_size(c, idx_size) 0
+#define dbg_check_filesystem(c) 0
+#define dbg_check_heap(c, heap, cat, add_pos) ({})
+#define dbg_check_lprops(c) 0
+#define dbg_check_lpt_nodes(c, cnode, row, col) 0
+#define dbg_force_in_the_gaps_enabled 0
+#define dbg_force_in_the_gaps() 0
+#define dbg_failure_mode 0
+#define dbg_failure_mode_registration(c) ({})
+#define dbg_failure_mode_deregistration(c) ({})
+
+#endif /* !CONFIG_UBIFS_FS_DEBUG */
+
+#endif /* !__UBIFS_DEBUG_H__ */
diff --git a/fs/ubifs/io.c b/fs/ubifs/io.c
new file mode 100644
index 0000000..aae5c65
--- /dev/null
+++ b/fs/ubifs/io.c
@@ -0,0 +1,316 @@
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ * Copyright (C) 2006, 2007 University of Szeged, Hungary
+ *
+ * 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
+ * Zoltan Sogor
+ */
+
+/*
+ * This file implements UBIFS I/O subsystem which provides various I/O-related
+ * helper functions (reading/writing/checking/validating nodes) and implements
+ * write-buffering support. Write buffers help to save space which otherwise
+ * would have been wasted for padding to the nearest minimal I/O unit boundary.
+ * Instead, data first goes to the write-buffer and is flushed when the
+ * buffer is full or when it is not used for some time (by timer). This is
+ * similar to the mechanism is used by JFFS2.
+ *
+ * Write-buffers are defined by 'struct ubifs_wbuf' objects and protected by
+ * mutexes defined inside these objects. Since sometimes upper-level code
+ * has to lock the write-buffer (e.g. journal space reservation code), many
+ * functions related to write-buffers have "nolock" suffix which means that the
+ * caller has to lock the write-buffer before calling this function.
+ *
+ * UBIFS stores nodes at 64 bit-aligned addresses. If the node length is not
+ * aligned, UBIFS starts the next node from the aligned address, and the padded
+ * bytes may contain any rubbish. In other words, UBIFS does not put padding
+ * bytes in those small gaps. Common headers of nodes store real node lengths,
+ * not aligned lengths. Indexing nodes also store real lengths in branches.
+ *
+ * UBIFS uses padding when it pads to the next min. I/O unit. In this case it
+ * uses padding nodes or padding bytes, if the padding node does not fit.
+ *
+ * All UBIFS nodes are protected by CRC checksums and UBIFS checks all nodes
+ * every time they are read from the flash media.
+ */
+
+#include "ubifs.h"
+
+/**
+ * ubifs_ro_mode - switch UBIFS to read read-only mode.
+ * @c: UBIFS file-system description object
+ * @err: error code which is the reason of switching to R/O mode
+ */
+void ubifs_ro_mode(struct ubifs_info *c, int err)
+{
+ if (!c->ro_media) {
+ c->ro_media = 1;
+ c->no_chk_data_crc = 0;
+ ubifs_warn("switched to read-only mode, error %d", err);
+ dbg_dump_stack();
+ }
+}
+
+/**
+ * ubifs_check_node - check node.
+ * @c: UBIFS file-system description object
+ * @buf: node to check
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ * @quiet: print no messages
+ * @must_chk_crc: indicates whether to always check the CRC
+ *
+ * This function checks node magic number and CRC checksum. This function also
+ * validates node length to prevent UBIFS from becoming crazy when an attacker
+ * feeds it a file-system image with incorrect nodes. For example, too large
+ * node length in the common header could cause UBIFS to read memory outside of
+ * allocated buffer when checking the CRC checksum.
+ *
+ * This function may skip data nodes CRC checking if @c->no_chk_data_crc is
+ * true, which is controlled by corresponding UBIFS mount option. However, if
+ * @must_chk_crc is true, then @c->no_chk_data_crc is ignored and CRC is
+ * checked. Similarly, if @c->always_chk_crc is true, @c->no_chk_data_crc is
+ * ignored and CRC is checked.
+ *
+ * This function returns zero in case of success and %-EUCLEAN in case of bad
+ * CRC or magic.
+ */
+int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
+ int offs, int quiet, int must_chk_crc)
+{
+ int err = -EINVAL, type, node_len;
+ uint32_t crc, node_crc, magic;
+ const struct ubifs_ch *ch = buf;
+
+ ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(!(offs & 7) && offs < c->leb_size);
+
+ magic = le32_to_cpu(ch->magic);
+ if (magic != UBIFS_NODE_MAGIC) {
+ if (!quiet)
+ ubifs_err("bad magic %#08x, expected %#08x",
+ magic, UBIFS_NODE_MAGIC);
+ err = -EUCLEAN;
+ goto out;
+ }
+
+ type = ch->node_type;
+ if (type < 0 || type >= UBIFS_NODE_TYPES_CNT) {
+ if (!quiet)
+ ubifs_err("bad node type %d", type);
+ goto out;
+ }
+
+ node_len = le32_to_cpu(ch->len);
+ if (node_len + offs > c->leb_size)
+ goto out_len;
+
+ if (c->ranges[type].max_len == 0) {
+ if (node_len != c->ranges[type].len)
+ goto out_len;
+ } else if (node_len < c->ranges[type].min_len ||
+ node_len > c->ranges[type].max_len)
+ goto out_len;
+
+ if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->always_chk_crc &&
+ c->no_chk_data_crc)
+ return 0;
+
+ crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
+ node_crc = le32_to_cpu(ch->crc);
+ if (crc != node_crc) {
+ if (!quiet)
+ ubifs_err("bad CRC: calculated %#08x, read %#08x",
+ crc, node_crc);
+ err = -EUCLEAN;
+ goto out;
+ }
+
+ return 0;
+
+out_len:
+ if (!quiet)
+ ubifs_err("bad node length %d", node_len);
+out:
+ if (!quiet) {
+ ubifs_err("bad node at LEB %d:%d", lnum, offs);
+ dbg_dump_node(c, buf);
+ dbg_dump_stack();
+ }
+ return err;
+}
+
+/**
+ * ubifs_pad - pad flash space.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to put padding to
+ * @pad: how many bytes to pad
+ *
+ * The flash media obliges us to write only in chunks of %c->min_io_size and
+ * when we have to write less data we add padding node to the write-buffer and
+ * pad it to the next minimal I/O unit's boundary. Padding nodes help when the
+ * media is being scanned. If the amount of wasted space is not enough to fit a
+ * padding node which takes %UBIFS_PAD_NODE_SZ bytes, we write padding bytes
+ * pattern (%UBIFS_PADDING_BYTE).
+ *
+ * Padding nodes are also used to fill gaps when the "commit-in-gaps" method is
+ * used.
+ */
+void ubifs_pad(const struct ubifs_info *c, void *buf, int pad)
+{
+ uint32_t crc;
+
+ ubifs_assert(pad >= 0 && !(pad & 7));
+
+ if (pad >= UBIFS_PAD_NODE_SZ) {
+ struct ubifs_ch *ch = buf;
+ struct ubifs_pad_node *pad_node = buf;
+
+ ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
+ ch->node_type = UBIFS_PAD_NODE;
+ ch->group_type = UBIFS_NO_NODE_GROUP;
+ ch->padding[0] = ch->padding[1] = 0;
+ ch->sqnum = 0;
+ ch->len = cpu_to_le32(UBIFS_PAD_NODE_SZ);
+ pad -= UBIFS_PAD_NODE_SZ;
+ pad_node->pad_len = cpu_to_le32(pad);
+ crc = crc32(UBIFS_CRC32_INIT, buf + 8, UBIFS_PAD_NODE_SZ - 8);
+ ch->crc = cpu_to_le32(crc);
+ memset(buf + UBIFS_PAD_NODE_SZ, 0, pad);
+ } else if (pad > 0)
+ /* Too little space, padding node won't fit */
+ memset(buf, UBIFS_PADDING_BYTE, pad);
+}
+
+/**
+ * next_sqnum - get next sequence number.
+ * @c: UBIFS file-system description object
+ */
+static unsigned long long next_sqnum(struct ubifs_info *c)
+{
+ unsigned long long sqnum;
+
+ spin_lock(&c->cnt_lock);
+ sqnum = ++c->max_sqnum;
+ spin_unlock(&c->cnt_lock);
+
+ if (unlikely(sqnum >= SQNUM_WARN_WATERMARK)) {
+ if (sqnum >= SQNUM_WATERMARK) {
+ ubifs_err("sequence number overflow %llu, end of life",
+ sqnum);
+ ubifs_ro_mode(c, -EINVAL);
+ }
+ ubifs_warn("running out of sequence numbers, end of life soon");
+ }
+
+ return sqnum;
+}
+
+/**
+ * ubifs_prepare_node - prepare node to be written to flash.
+ * @c: UBIFS file-system description object
+ * @node: the node to pad
+ * @len: node length
+ * @pad: if the buffer has to be padded
+ *
+ * This function prepares node at @node to be written to the media - it
+ * calculates node CRC, fills the common header, and adds proper padding up to
+ * the next minimum I/O unit if @pad is not zero.
+ */
+void ubifs_prepare_node(struct ubifs_info *c, void *node, int len, int pad)
+{
+ uint32_t crc;
+ struct ubifs_ch *ch = node;
+ unsigned long long sqnum = next_sqnum(c);
+
+ ubifs_assert(len >= UBIFS_CH_SZ);
+
+ ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
+ ch->len = cpu_to_le32(len);
+ ch->group_type = UBIFS_NO_NODE_GROUP;
+ ch->sqnum = cpu_to_le64(sqnum);
+ ch->padding[0] = ch->padding[1] = 0;
+ crc = crc32(UBIFS_CRC32_INIT, node + 8, len - 8);
+ ch->crc = cpu_to_le32(crc);
+
+ if (pad) {
+ len = ALIGN(len, 8);
+ pad = ALIGN(len, c->min_io_size) - len;
+ ubifs_pad(c, node + len, pad);
+ }
+}
+
+/**
+ * ubifs_read_node - read node.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to read to
+ * @type: node type
+ * @len: node length (not aligned)
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ *
+ * This function reads a node of known type and and length, checks it and
+ * stores in @buf. Returns zero in case of success, %-EUCLEAN if CRC mismatched
+ * and a negative error code in case of failure.
+ */
+int ubifs_read_node(const struct ubifs_info *c, void *buf, int type, int len,
+ int lnum, int offs)
+{
+ int err, l;
+ struct ubifs_ch *ch = buf;
+
+ dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len);
+ ubifs_assert(lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(len >= UBIFS_CH_SZ && offs + len <= c->leb_size);
+ ubifs_assert(!(offs & 7) && offs < c->leb_size);
+ ubifs_assert(type >= 0 && type < UBIFS_NODE_TYPES_CNT);
+
+ err = ubi_read(c->ubi, lnum, buf, offs, len);
+ if (err && err != -EBADMSG) {
+ ubifs_err("cannot read node %d from LEB %d:%d, error %d",
+ type, lnum, offs, err);
+ return err;
+ }
+
+ if (type != ch->node_type) {
+ ubifs_err("bad node type (%d but expected %d)",
+ ch->node_type, type);
+ goto out;
+ }
+
+ err = ubifs_check_node(c, buf, lnum, offs, 0, 0);
+ if (err) {
+ ubifs_err("expected node type %d", type);
+ return err;
+ }
+
+ l = le32_to_cpu(ch->len);
+ if (l != len) {
+ ubifs_err("bad node length %d, expected %d", l, len);
+ goto out;
+ }
+
+ return 0;
+
+out:
+ ubifs_err("bad node at LEB %d:%d", lnum, offs);
+ dbg_dump_node(c, buf);
+ dbg_dump_stack();
+ return -EINVAL;
+}
diff --git a/fs/ubifs/key.h b/fs/ubifs/key.h
new file mode 100644
index 0000000..efb3430
--- /dev/null
+++ b/fs/ubifs/key.h
@@ -0,0 +1,557 @@
+/*
+ * 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 header contains various key-related definitions and helper function.
+ * UBIFS allows several key schemes, so we access key fields only via these
+ * helpers. At the moment only one key scheme is supported.
+ *
+ * Simple key scheme
+ * ~~~~~~~~~~~~~~~~~
+ *
+ * Keys are 64-bits long. First 32-bits are inode number (parent inode number
+ * in case of direntry key). Next 3 bits are node type. The last 29 bits are
+ * 4KiB offset in case of inode node, and direntry hash in case of a direntry
+ * node. We use "r5" hash borrowed from reiserfs.
+ */
+
+#ifndef __UBIFS_KEY_H__
+#define __UBIFS_KEY_H__
+
+/**
+ * key_mask_hash - mask a valid hash value.
+ * @val: value to be masked
+ *
+ * We use hash values as offset in directories, so values %0 and %1 are
+ * reserved for "." and "..". %2 is reserved for "end of readdir" marker. This
+ * function makes sure the reserved values are not used.
+ */
+static inline uint32_t key_mask_hash(uint32_t hash)
+{
+ hash &= UBIFS_S_KEY_HASH_MASK;
+ if (unlikely(hash <= 2))
+ hash += 3;
+ return hash;
+}
+
+/**
+ * key_r5_hash - R5 hash function (borrowed from reiserfs).
+ * @s: direntry name
+ * @len: name length
+ */
+static inline uint32_t key_r5_hash(const char *s, int len)
+{
+ uint32_t a = 0;
+ const signed char *str = (const signed char *)s;
+
+ while (*str) {
+ a += *str << 4;
+ a += *str >> 4;
+ a *= 11;
+ str++;
+ }
+
+ return key_mask_hash(a);
+}
+
+/**
+ * key_test_hash - testing hash function.
+ * @str: direntry name
+ * @len: name length
+ */
+static inline uint32_t key_test_hash(const char *str, int len)
+{
+ uint32_t a = 0;
+
+ len = min_t(uint32_t, len, 4);
+ memcpy(&a, str, len);
+ return key_mask_hash(a);
+}
+
+/**
+ * ino_key_init - initialize inode key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ */
+static inline void ino_key_init(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = UBIFS_INO_KEY << UBIFS_S_KEY_BLOCK_BITS;
+}
+
+/**
+ * ino_key_init_flash - initialize on-flash inode key.
+ * @c: UBIFS file-system description object
+ * @k: key to initialize
+ * @inum: inode number
+ */
+static inline void ino_key_init_flash(const struct ubifs_info *c, void *k,
+ ino_t inum)
+{
+ union ubifs_key *key = k;
+
+ key->j32[0] = cpu_to_le32(inum);
+ key->j32[1] = cpu_to_le32(UBIFS_INO_KEY << UBIFS_S_KEY_BLOCK_BITS);
+ memset(k + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+}
+
+/**
+ * lowest_ino_key - get the lowest possible inode key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ */
+static inline void lowest_ino_key(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = 0;
+}
+
+/**
+ * highest_ino_key - get the highest possible inode key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ */
+static inline void highest_ino_key(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = 0xffffffff;
+}
+
+/**
+ * dent_key_init - initialize directory entry key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: parent inode number
+ * @nm: direntry name and length
+ */
+static inline void dent_key_init(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum,
+ const struct qstr *nm)
+{
+ uint32_t hash = c->key_hash(nm->name, nm->len);
+
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->u32[0] = inum;
+ key->u32[1] = hash | (UBIFS_DENT_KEY << UBIFS_S_KEY_HASH_BITS);
+}
+
+/**
+ * dent_key_init_hash - initialize directory entry key without re-calculating
+ * hash function.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: parent inode number
+ * @hash: direntry name hash
+ */
+static inline void dent_key_init_hash(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum,
+ uint32_t hash)
+{
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->u32[0] = inum;
+ key->u32[1] = hash | (UBIFS_DENT_KEY << UBIFS_S_KEY_HASH_BITS);
+}
+
+/**
+ * dent_key_init_flash - initialize on-flash directory entry key.
+ * @c: UBIFS file-system description object
+ * @k: key to initialize
+ * @inum: parent inode number
+ * @nm: direntry name and length
+ */
+static inline void dent_key_init_flash(const struct ubifs_info *c, void *k,
+ ino_t inum, const struct qstr *nm)
+{
+ union ubifs_key *key = k;
+ uint32_t hash = c->key_hash(nm->name, nm->len);
+
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->j32[0] = cpu_to_le32(inum);
+ key->j32[1] = cpu_to_le32(hash |
+ (UBIFS_DENT_KEY << UBIFS_S_KEY_HASH_BITS));
+ memset(k + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+}
+
+/**
+ * lowest_dent_key - get the lowest possible directory entry key.
+ * @c: UBIFS file-system description object
+ * @key: where to store the lowest key
+ * @inum: parent inode number
+ */
+static inline void lowest_dent_key(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = UBIFS_DENT_KEY << UBIFS_S_KEY_HASH_BITS;
+}
+
+/**
+ * xent_key_init - initialize extended attribute entry key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: host inode number
+ * @nm: extended attribute entry name and length
+ */
+static inline void xent_key_init(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum,
+ const struct qstr *nm)
+{
+ uint32_t hash = c->key_hash(nm->name, nm->len);
+
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->u32[0] = inum;
+ key->u32[1] = hash | (UBIFS_XENT_KEY << UBIFS_S_KEY_HASH_BITS);
+}
+
+/**
+ * xent_key_init_hash - initialize extended attribute entry key without
+ * re-calculating hash function.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: host inode number
+ * @hash: extended attribute entry name hash
+ */
+static inline void xent_key_init_hash(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum,
+ uint32_t hash)
+{
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->u32[0] = inum;
+ key->u32[1] = hash | (UBIFS_XENT_KEY << UBIFS_S_KEY_HASH_BITS);
+}
+
+/**
+ * xent_key_init_flash - initialize on-flash extended attribute entry key.
+ * @c: UBIFS file-system description object
+ * @k: key to initialize
+ * @inum: host inode number
+ * @nm: extended attribute entry name and length
+ */
+static inline void xent_key_init_flash(const struct ubifs_info *c, void *k,
+ ino_t inum, const struct qstr *nm)
+{
+ union ubifs_key *key = k;
+ uint32_t hash = c->key_hash(nm->name, nm->len);
+
+ ubifs_assert(!(hash & ~UBIFS_S_KEY_HASH_MASK));
+ key->j32[0] = cpu_to_le32(inum);
+ key->j32[1] = cpu_to_le32(hash |
+ (UBIFS_XENT_KEY << UBIFS_S_KEY_HASH_BITS));
+ memset(k + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+}
+
+/**
+ * lowest_xent_key - get the lowest possible extended attribute entry key.
+ * @c: UBIFS file-system description object
+ * @key: where to store the lowest key
+ * @inum: host inode number
+ */
+static inline void lowest_xent_key(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = UBIFS_XENT_KEY << UBIFS_S_KEY_HASH_BITS;
+}
+
+/**
+ * data_key_init - initialize data key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ * @block: block number
+ */
+static inline void data_key_init(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum,
+ unsigned int block)
+{
+ ubifs_assert(!(block & ~UBIFS_S_KEY_BLOCK_MASK));
+ key->u32[0] = inum;
+ key->u32[1] = block | (UBIFS_DATA_KEY << UBIFS_S_KEY_BLOCK_BITS);
+}
+
+/**
+ * data_key_init_flash - initialize on-flash data key.
+ * @c: UBIFS file-system description object
+ * @k: key to initialize
+ * @inum: inode number
+ * @block: block number
+ */
+static inline void data_key_init_flash(const struct ubifs_info *c, void *k,
+ ino_t inum, unsigned int block)
+{
+ union ubifs_key *key = k;
+
+ ubifs_assert(!(block & ~UBIFS_S_KEY_BLOCK_MASK));
+ key->j32[0] = cpu_to_le32(inum);
+ key->j32[1] = cpu_to_le32(block |
+ (UBIFS_DATA_KEY << UBIFS_S_KEY_BLOCK_BITS));
+ memset(k + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+}
+
+/**
+ * trun_key_init - initialize truncation node key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ *
+ * Note, UBIFS does not have truncation keys on the media and this function is
+ * only used for purposes of replay.
+ */
+static inline void trun_key_init(const struct ubifs_info *c,
+ union ubifs_key *key, ino_t inum)
+{
+ key->u32[0] = inum;
+ key->u32[1] = UBIFS_TRUN_KEY << UBIFS_S_KEY_BLOCK_BITS;
+}
+
+/**
+ * key_type - get key type.
+ * @c: UBIFS file-system description object
+ * @key: key to get type of
+ */
+static inline int key_type(const struct ubifs_info *c,
+ const union ubifs_key *key)
+{
+ return key->u32[1] >> UBIFS_S_KEY_BLOCK_BITS;
+}
+
+/**
+ * key_type_flash - get type of a on-flash formatted key.
+ * @c: UBIFS file-system description object
+ * @k: key to get type of
+ */
+static inline int key_type_flash(const struct ubifs_info *c, const void *k)
+{
+ const union ubifs_key *key = k;
+
+ return le32_to_cpu(key->j32[1]) >> UBIFS_S_KEY_BLOCK_BITS;
+}
+
+/**
+ * key_inum - fetch inode number from key.
+ * @c: UBIFS file-system description object
+ * @k: key to fetch inode number from
+ */
+static inline ino_t key_inum(const struct ubifs_info *c, const void *k)
+{
+ const union ubifs_key *key = k;
+
+ return key->u32[0];
+}
+
+/**
+ * key_inum_flash - fetch inode number from an on-flash formatted key.
+ * @c: UBIFS file-system description object
+ * @k: key to fetch inode number from
+ */
+static inline ino_t key_inum_flash(const struct ubifs_info *c, const void *k)
+{
+ const union ubifs_key *key = k;
+
+ return le32_to_cpu(key->j32[0]);
+}
+
+/**
+ * key_hash - get directory entry hash.
+ * @c: UBIFS file-system description object
+ * @key: the key to get hash from
+ */
+static inline int key_hash(const struct ubifs_info *c,
+ const union ubifs_key *key)
+{
+ return key->u32[1] & UBIFS_S_KEY_HASH_MASK;
+}
+
+/**
+ * key_hash_flash - get directory entry hash from an on-flash formatted key.
+ * @c: UBIFS file-system description object
+ * @k: the key to get hash from
+ */
+static inline int key_hash_flash(const struct ubifs_info *c, const void *k)
+{
+ const union ubifs_key *key = k;
+
+ return le32_to_cpu(key->j32[1]) & UBIFS_S_KEY_HASH_MASK;
+}
+
+/**
+ * key_block - get data block number.
+ * @c: UBIFS file-system description object
+ * @key: the key to get the block number from
+ */
+static inline unsigned int key_block(const struct ubifs_info *c,
+ const union ubifs_key *key)
+{
+ return key->u32[1] & UBIFS_S_KEY_BLOCK_MASK;
+}
+
+/**
+ * key_block_flash - get data block number from an on-flash formatted key.
+ * @c: UBIFS file-system description object
+ * @k: the key to get the block number from
+ */
+static inline unsigned int key_block_flash(const struct ubifs_info *c,
+ const void *k)
+{
+ const union ubifs_key *key = k;
+
+ return le32_to_cpu(key->j32[1]) & UBIFS_S_KEY_BLOCK_MASK;
+}
+
+/**
+ * key_read - transform a key to in-memory format.
+ * @c: UBIFS file-system description object
+ * @from: the key to transform
+ * @to: the key to store the result
+ */
+static inline void key_read(const struct ubifs_info *c, const void *from,
+ union ubifs_key *to)
+{
+ const union ubifs_key *f = from;
+
+ to->u32[0] = le32_to_cpu(f->j32[0]);
+ to->u32[1] = le32_to_cpu(f->j32[1]);
+}
+
+/**
+ * key_write - transform a key from in-memory format.
+ * @c: UBIFS file-system description object
+ * @from: the key to transform
+ * @to: the key to store the result
+ */
+static inline void key_write(const struct ubifs_info *c,
+ const union ubifs_key *from, void *to)
+{
+ union ubifs_key *t = to;
+
+ t->j32[0] = cpu_to_le32(from->u32[0]);
+ t->j32[1] = cpu_to_le32(from->u32[1]);
+ memset(to + 8, 0, UBIFS_MAX_KEY_LEN - 8);
+}
+
+/**
+ * key_write_idx - transform a key from in-memory format for the index.
+ * @c: UBIFS file-system description object
+ * @from: the key to transform
+ * @to: the key to store the result
+ */
+static inline void key_write_idx(const struct ubifs_info *c,
+ const union ubifs_key *from, void *to)
+{
+ union ubifs_key *t = to;
+
+ t->j32[0] = cpu_to_le32(from->u32[0]);
+ t->j32[1] = cpu_to_le32(from->u32[1]);
+}
+
+/**
+ * key_copy - copy a key.
+ * @c: UBIFS file-system description object
+ * @from: the key to copy from
+ * @to: the key to copy to
+ */
+static inline void key_copy(const struct ubifs_info *c,
+ const union ubifs_key *from, union ubifs_key *to)
+{
+ to->u64[0] = from->u64[0];
+}
+
+/**
+ * keys_cmp - compare keys.
+ * @c: UBIFS file-system description object
+ * @key1: the first key to compare
+ * @key2: the second key to compare
+ *
+ * This function compares 2 keys and returns %-1 if @key1 is less than
+ * @key2, %0 if the keys are equivalent and %1 if @key1 is greater than @key2.
+ */
+static inline int keys_cmp(const struct ubifs_info *c,
+ const union ubifs_key *key1,
+ const union ubifs_key *key2)
+{
+ if (key1->u32[0] < key2->u32[0])
+ return -1;
+ if (key1->u32[0] > key2->u32[0])
+ return 1;
+ if (key1->u32[1] < key2->u32[1])
+ return -1;
+ if (key1->u32[1] > key2->u32[1])
+ return 1;
+
+ return 0;
+}
+
+/**
+ * keys_eq - determine if keys are equivalent.
+ * @c: UBIFS file-system description object
+ * @key1: the first key to compare
+ * @key2: the second key to compare
+ *
+ * This function compares 2 keys and returns %1 if @key1 is equal to @key2 and
+ * %0 if not.
+ */
+static inline int keys_eq(const struct ubifs_info *c,
+ const union ubifs_key *key1,
+ const union ubifs_key *key2)
+{
+ if (key1->u32[0] != key2->u32[0])
+ return 0;
+ if (key1->u32[1] != key2->u32[1])
+ return 0;
+ return 1;
+}
+
+/**
+ * is_hash_key - is a key vulnerable to hash collisions.
+ * @c: UBIFS file-system description object
+ * @key: key
+ *
+ * This function returns %1 if @key is a hashed key or %0 otherwise.
+ */
+static inline int is_hash_key(const struct ubifs_info *c,
+ const union ubifs_key *key)
+{
+ int type = key_type(c, key);
+
+ return type == UBIFS_DENT_KEY || type == UBIFS_XENT_KEY;
+}
+
+/**
+ * key_max_inode_size - get maximum file size allowed by current key format.
+ * @c: UBIFS file-system description object
+ */
+static inline unsigned long long key_max_inode_size(const struct ubifs_info *c)
+{
+ switch (c->key_fmt) {
+ case UBIFS_SIMPLE_KEY_FMT:
+ return (1ULL << UBIFS_S_KEY_BLOCK_BITS) * UBIFS_BLOCK_SIZE;
+ default:
+ return 0;
+ }
+}
+#endif /* !__UBIFS_KEY_H__ */
diff --git a/fs/ubifs/log.c b/fs/ubifs/log.c
new file mode 100644
index 0000000..68a9bd9
--- /dev/null
+++ b/fs/ubifs/log.c
@@ -0,0 +1,104 @@
+/*
+ * 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 is a part of UBIFS journal implementation and contains various
+ * functions which manipulate the log. The log is a fixed area on the flash
+ * which does not contain any data but refers to buds. The log is a part of the
+ * journal.
+ */
+
+#include "ubifs.h"
+
+/**
+ * ubifs_search_bud - search bud LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: logical eraseblock number to search
+ *
+ * This function searches bud LEB @lnum. Returns bud description object in case
+ * of success and %NULL if there is no bud with this LEB number.
+ */
+struct ubifs_bud *ubifs_search_bud(struct ubifs_info *c, int lnum)
+{
+ struct rb_node *p;
+ struct ubifs_bud *bud;
+
+ spin_lock(&c->buds_lock);
+ p = c->buds.rb_node;
+ while (p) {
+ bud = rb_entry(p, struct ubifs_bud, rb);
+ if (lnum < bud->lnum)
+ p = p->rb_left;
+ else if (lnum > bud->lnum)
+ p = p->rb_right;
+ else {
+ spin_unlock(&c->buds_lock);
+ return bud;
+ }
+ }
+ spin_unlock(&c->buds_lock);
+ return NULL;
+}
+
+/**
+ * ubifs_add_bud - add bud LEB to the tree of buds and its journal head list.
+ * @c: UBIFS file-system description object
+ * @bud: the bud to add
+ */
+void ubifs_add_bud(struct ubifs_info *c, struct ubifs_bud *bud)
+{
+ struct rb_node **p, *parent = NULL;
+ struct ubifs_bud *b;
+ struct ubifs_jhead *jhead;
+
+ spin_lock(&c->buds_lock);
+ p = &c->buds.rb_node;
+ while (*p) {
+ parent = *p;
+ b = rb_entry(parent, struct ubifs_bud, rb);
+ ubifs_assert(bud->lnum != b->lnum);
+ if (bud->lnum < b->lnum)
+ p = &(*p)->rb_left;
+ else
+ p = &(*p)->rb_right;
+ }
+
+ rb_link_node(&bud->rb, parent, p);
+ rb_insert_color(&bud->rb, &c->buds);
+ if (c->jheads) {
+ jhead = &c->jheads[bud->jhead];
+ list_add_tail(&bud->list, &jhead->buds_list);
+ } else
+ ubifs_assert(c->replaying && (c->vfs_sb->s_flags & MS_RDONLY));
+
+ /*
+ * Note, although this is a new bud, we anyway account this space now,
+ * before any data has been written to it, because this is about to
+ * guarantee fixed mount time, and this bud will anyway be read and
+ * scanned.
+ */
+ c->bud_bytes += c->leb_size - bud->start;
+
+ dbg_log("LEB %d:%d, jhead %d, bud_bytes %lld", bud->lnum,
+ bud->start, bud->jhead, c->bud_bytes);
+ spin_unlock(&c->buds_lock);
+}
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
new file mode 100644
index 0000000..8ce4949
--- /dev/null
+++ b/fs/ubifs/lprops.c
@@ -0,0 +1,842 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements the functions that access LEB properties and their
+ * categories. LEBs are categorized based on the needs of UBIFS, and the
+ * categories are stored as either heaps or lists to provide a fast way of
+ * finding a LEB in a particular category. For example, UBIFS may need to find
+ * an empty LEB for the journal, or a very dirty LEB for garbage collection.
+ */
+
+#include "ubifs.h"
+
+/**
+ * get_heap_comp_val - get the LEB properties value for heap comparisons.
+ * @lprops: LEB properties
+ * @cat: LEB category
+ */
+static int get_heap_comp_val(struct ubifs_lprops *lprops, int cat)
+{
+ switch (cat) {
+ case LPROPS_FREE:
+ return lprops->free;
+ case LPROPS_DIRTY_IDX:
+ return lprops->free + lprops->dirty;
+ default:
+ return lprops->dirty;
+ }
+}
+
+/**
+ * move_up_lpt_heap - move a new heap entry up as far as possible.
+ * @c: UBIFS file-system description object
+ * @heap: LEB category heap
+ * @lprops: LEB properties to move
+ * @cat: LEB category
+ *
+ * New entries to a heap are added at the bottom and then moved up until the
+ * parent's value is greater. In the case of LPT's category heaps, the value
+ * is either the amount of free space or the amount of dirty space, depending
+ * on the category.
+ */
+static void move_up_lpt_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
+ struct ubifs_lprops *lprops, int cat)
+{
+ int val1, val2, hpos;
+
+ hpos = lprops->hpos;
+ if (!hpos)
+ return; /* Already top of the heap */
+ val1 = get_heap_comp_val(lprops, cat);
+ /* Compare to parent and, if greater, move up the heap */
+ do {
+ int ppos = (hpos - 1) / 2;
+
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val2 >= val1)
+ return;
+ /* Greater than parent so move up */
+ heap->arr[ppos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[ppos];
+ heap->arr[ppos] = lprops;
+ lprops->hpos = ppos;
+ hpos = ppos;
+ } while (hpos);
+}
+
+/**
+ * adjust_lpt_heap - move a changed heap entry up or down the heap.
+ * @c: UBIFS file-system description object
+ * @heap: LEB category heap
+ * @lprops: LEB properties to move
+ * @hpos: heap position of @lprops
+ * @cat: LEB category
+ *
+ * Changed entries in a heap are moved up or down until the parent's value is
+ * greater. In the case of LPT's category heaps, the value is either the amount
+ * of free space or the amount of dirty space, depending on the category.
+ */
+static void adjust_lpt_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
+ struct ubifs_lprops *lprops, int hpos, int cat)
+{
+ int val1, val2, val3, cpos;
+
+ val1 = get_heap_comp_val(lprops, cat);
+ /* Compare to parent and, if greater than parent, move up the heap */
+ if (hpos) {
+ int ppos = (hpos - 1) / 2;
+
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val1 > val2) {
+ /* Greater than parent so move up */
+ while (1) {
+ heap->arr[ppos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[ppos];
+ heap->arr[ppos] = lprops;
+ lprops->hpos = ppos;
+ hpos = ppos;
+ if (!hpos)
+ return;
+ ppos = (hpos - 1) / 2;
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val1 <= val2)
+ return;
+ /* Still greater than parent so keep going */
+ }
+ }
+ }
+
+ /* Not greater than parent, so compare to children */
+ while (1) {
+ /* Compare to left child */
+ cpos = hpos * 2 + 1;
+ if (cpos >= heap->cnt)
+ return;
+ val2 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 < val2) {
+ /* Less than left child, so promote biggest child */
+ if (cpos + 1 < heap->cnt) {
+ val3 = get_heap_comp_val(heap->arr[cpos + 1],
+ cat);
+ if (val3 > val2)
+ cpos += 1; /* Right child is bigger */
+ }
+ heap->arr[cpos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[cpos];
+ heap->arr[cpos] = lprops;
+ lprops->hpos = cpos;
+ hpos = cpos;
+ continue;
+ }
+ /* Compare to right child */
+ cpos += 1;
+ if (cpos >= heap->cnt)
+ return;
+ val3 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 < val3) {
+ /* Less than right child, so promote right child */
+ heap->arr[cpos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[cpos];
+ heap->arr[cpos] = lprops;
+ lprops->hpos = cpos;
+ hpos = cpos;
+ continue;
+ }
+ return;
+ }
+}
+
+/**
+ * add_to_lpt_heap - add LEB properties to a LEB category heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to add
+ * @cat: LEB category
+ *
+ * This function returns %1 if @lprops is added to the heap for LEB category
+ * @cat, otherwise %0 is returned because the heap is full.
+ */
+static int add_to_lpt_heap(struct ubifs_info *c, struct ubifs_lprops *lprops,
+ int cat)
+{
+ struct ubifs_lpt_heap *heap = &c->lpt_heap[cat - 1];
+
+ if (heap->cnt >= heap->max_cnt) {
+ const int b = LPT_HEAP_SZ / 2 - 1;
+ int cpos, val1, val2;
+
+ /* Compare to some other LEB on the bottom of heap */
+ /* Pick a position kind of randomly */
+ cpos = (((size_t)lprops >> 4) & b) + b;
+ ubifs_assert(cpos >= b);
+ ubifs_assert(cpos < LPT_HEAP_SZ);
+ ubifs_assert(cpos < heap->cnt);
+
+ val1 = get_heap_comp_val(lprops, cat);
+ val2 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 > val2) {
+ struct ubifs_lprops *lp;
+
+ lp = heap->arr[cpos];
+ lp->flags &= ~LPROPS_CAT_MASK;
+ lp->flags |= LPROPS_UNCAT;
+ list_add(&lp->list, &c->uncat_list);
+ lprops->hpos = cpos;
+ heap->arr[cpos] = lprops;
+ move_up_lpt_heap(c, heap, lprops, cat);
+ dbg_check_heap(c, heap, cat, lprops->hpos);
+ return 1; /* Added to heap */
+ }
+ dbg_check_heap(c, heap, cat, -1);
+ return 0; /* Not added to heap */
+ } else {
+ lprops->hpos = heap->cnt++;
+ heap->arr[lprops->hpos] = lprops;
+ move_up_lpt_heap(c, heap, lprops, cat);
+ dbg_check_heap(c, heap, cat, lprops->hpos);
+ return 1; /* Added to heap */
+ }
+}
+
+/**
+ * remove_from_lpt_heap - remove LEB properties from a LEB category heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to remove
+ * @cat: LEB category
+ */
+static void remove_from_lpt_heap(struct ubifs_info *c,
+ struct ubifs_lprops *lprops, int cat)
+{
+ struct ubifs_lpt_heap *heap;
+ int hpos = lprops->hpos;
+
+ heap = &c->lpt_heap[cat - 1];
+ ubifs_assert(hpos >= 0 && hpos < heap->cnt);
+ ubifs_assert(heap->arr[hpos] == lprops);
+ heap->cnt -= 1;
+ if (hpos < heap->cnt) {
+ heap->arr[hpos] = heap->arr[heap->cnt];
+ heap->arr[hpos]->hpos = hpos;
+ adjust_lpt_heap(c, heap, heap->arr[hpos], hpos, cat);
+ }
+ dbg_check_heap(c, heap, cat, -1);
+}
+
+/**
+ * lpt_heap_replace - replace lprops in a category heap.
+ * @c: UBIFS file-system description object
+ * @old_lprops: LEB properties to replace
+ * @new_lprops: LEB properties with which to replace
+ * @cat: LEB category
+ *
+ * During commit it is sometimes necessary to copy a pnode (see dirty_cow_pnode)
+ * and the lprops that the pnode contains. When that happens, references in
+ * the category heaps to those lprops must be updated to point to the new
+ * lprops. This function does that.
+ */
+static void lpt_heap_replace(struct ubifs_info *c,
+ struct ubifs_lprops *old_lprops,
+ struct ubifs_lprops *new_lprops, int cat)
+{
+ struct ubifs_lpt_heap *heap;
+ int hpos = new_lprops->hpos;
+
+ heap = &c->lpt_heap[cat - 1];
+ heap->arr[hpos] = new_lprops;
+}
+
+/**
+ * ubifs_add_to_cat - add LEB properties to a category list or heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to add
+ * @cat: LEB category to which to add
+ *
+ * LEB properties are categorized to enable fast find operations.
+ */
+void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops,
+ int cat)
+{
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ if (add_to_lpt_heap(c, lprops, cat))
+ break;
+ /* No more room on heap so make it uncategorized */
+ cat = LPROPS_UNCAT;
+ /* Fall through */
+ case LPROPS_UNCAT:
+ list_add(&lprops->list, &c->uncat_list);
+ break;
+ case LPROPS_EMPTY:
+ list_add(&lprops->list, &c->empty_list);
+ break;
+ case LPROPS_FREEABLE:
+ list_add(&lprops->list, &c->freeable_list);
+ c->freeable_cnt += 1;
+ break;
+ case LPROPS_FRDI_IDX:
+ list_add(&lprops->list, &c->frdi_idx_list);
+ break;
+ default:
+ ubifs_assert(0);
+ }
+ lprops->flags &= ~LPROPS_CAT_MASK;
+ lprops->flags |= cat;
+}
+
+/**
+ * ubifs_remove_from_cat - remove LEB properties from a category list or heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to remove
+ * @cat: LEB category from which to remove
+ *
+ * LEB properties are categorized to enable fast find operations.
+ */
+static void ubifs_remove_from_cat(struct ubifs_info *c,
+ struct ubifs_lprops *lprops, int cat)
+{
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ remove_from_lpt_heap(c, lprops, cat);
+ break;
+ case LPROPS_FREEABLE:
+ c->freeable_cnt -= 1;
+ ubifs_assert(c->freeable_cnt >= 0);
+ /* Fall through */
+ case LPROPS_UNCAT:
+ case LPROPS_EMPTY:
+ case LPROPS_FRDI_IDX:
+ ubifs_assert(!list_empty(&lprops->list));
+ list_del(&lprops->list);
+ break;
+ default:
+ ubifs_assert(0);
+ }
+}
+
+/**
+ * ubifs_replace_cat - replace lprops in a category list or heap.
+ * @c: UBIFS file-system description object
+ * @old_lprops: LEB properties to replace
+ * @new_lprops: LEB properties with which to replace
+ *
+ * During commit it is sometimes necessary to copy a pnode (see dirty_cow_pnode)
+ * and the lprops that the pnode contains. When that happens, references in
+ * category lists and heaps must be replaced. This function does that.
+ */
+void ubifs_replace_cat(struct ubifs_info *c, struct ubifs_lprops *old_lprops,
+ struct ubifs_lprops *new_lprops)
+{
+ int cat;
+
+ cat = new_lprops->flags & LPROPS_CAT_MASK;
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ lpt_heap_replace(c, old_lprops, new_lprops, cat);
+ break;
+ case LPROPS_UNCAT:
+ case LPROPS_EMPTY:
+ case LPROPS_FREEABLE:
+ case LPROPS_FRDI_IDX:
+ list_replace(&old_lprops->list, &new_lprops->list);
+ break;
+ default:
+ ubifs_assert(0);
+ }
+}
+
+/**
+ * ubifs_ensure_cat - ensure LEB properties are categorized.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties
+ *
+ * A LEB may have fallen off of the bottom of a heap, and ended up as
+ * uncategorized even though it has enough space for us now. If that is the case
+ * this function will put the LEB back onto a heap.
+ */
+void ubifs_ensure_cat(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ int cat = lprops->flags & LPROPS_CAT_MASK;
+
+ if (cat != LPROPS_UNCAT)
+ return;
+ cat = ubifs_categorize_lprops(c, lprops);
+ if (cat == LPROPS_UNCAT)
+ return;
+ ubifs_remove_from_cat(c, lprops, LPROPS_UNCAT);
+ ubifs_add_to_cat(c, lprops, cat);
+}
+
+/**
+ * ubifs_categorize_lprops - categorize LEB properties.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to categorize
+ *
+ * LEB properties are categorized to enable fast find operations. This function
+ * returns the LEB category to which the LEB properties belong. Note however
+ * that if the LEB category is stored as a heap and the heap is full, the
+ * LEB properties may have their category changed to %LPROPS_UNCAT.
+ */
+int ubifs_categorize_lprops(const struct ubifs_info *c,
+ const struct ubifs_lprops *lprops)
+{
+ if (lprops->flags & LPROPS_TAKEN)
+ return LPROPS_UNCAT;
+
+ if (lprops->free == c->leb_size) {
+ ubifs_assert(!(lprops->flags & LPROPS_INDEX));
+ return LPROPS_EMPTY;
+ }
+
+ if (lprops->free + lprops->dirty == c->leb_size) {
+ if (lprops->flags & LPROPS_INDEX)
+ return LPROPS_FRDI_IDX;
+ else
+ return LPROPS_FREEABLE;
+ }
+
+ if (lprops->flags & LPROPS_INDEX) {
+ if (lprops->dirty + lprops->free >= c->min_idx_node_sz)
+ return LPROPS_DIRTY_IDX;
+ } else {
+ if (lprops->dirty >= c->dead_wm &&
+ lprops->dirty > lprops->free)
+ return LPROPS_DIRTY;
+ if (lprops->free > 0)
+ return LPROPS_FREE;
+ }
+
+ return LPROPS_UNCAT;
+}
+
+/**
+ * change_category - change LEB properties category.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to recategorize
+ *
+ * LEB properties are categorized to enable fast find operations. When the LEB
+ * properties change they must be recategorized.
+ */
+static void change_category(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ int old_cat = lprops->flags & LPROPS_CAT_MASK;
+ int new_cat = ubifs_categorize_lprops(c, lprops);
+
+ if (old_cat == new_cat) {
+ struct ubifs_lpt_heap *heap = &c->lpt_heap[new_cat - 1];
+
+ /* lprops on a heap now must be moved up or down */
+ if (new_cat < 1 || new_cat > LPROPS_HEAP_CNT)
+ return; /* Not on a heap */
+ heap = &c->lpt_heap[new_cat - 1];
+ adjust_lpt_heap(c, heap, lprops, lprops->hpos, new_cat);
+ } else {
+ ubifs_remove_from_cat(c, lprops, old_cat);
+ ubifs_add_to_cat(c, lprops, new_cat);
+ }
+}
+
+/**
+ * calc_dark - calculate LEB dark space size.
+ * @c: the UBIFS file-system description object
+ * @spc: amount of free and dirty space in the LEB
+ *
+ * This function calculates amount of dark space in an LEB which has @spc bytes
+ * of free and dirty space. Returns the calculations result.
+ *
+ * Dark space is the space which is not always usable - it depends on which
+ * nodes are written in which order. E.g., if an LEB has only 512 free bytes,
+ * it is dark space, because it cannot fit a large data node. So UBIFS cannot
+ * count on this LEB and treat these 512 bytes as usable because it is not true
+ * if, for example, only big chunks of uncompressible data will be written to
+ * the FS.
+ */
+static int calc_dark(struct ubifs_info *c, int spc)
+{
+ ubifs_assert(!(spc & 7));
+
+ if (spc < c->dark_wm)
+ return spc;
+
+ /*
+ * If we have slightly more space then the dark space watermark, we can
+ * anyway safely assume it we'll be able to write a node of the
+ * smallest size there.
+ */
+ if (spc - c->dark_wm < MIN_WRITE_SZ)
+ return spc - MIN_WRITE_SZ;
+
+ return c->dark_wm;
+}
+
+/**
+ * is_lprops_dirty - determine if LEB properties are dirty.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to test
+ */
+static int is_lprops_dirty(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ struct ubifs_pnode *pnode;
+ int pos;
+
+ pos = (lprops->lnum - c->main_first) & (UBIFS_LPT_FANOUT - 1);
+ pnode = (struct ubifs_pnode *)container_of(lprops - pos,
+ struct ubifs_pnode,
+ lprops[0]);
+ return !test_bit(COW_ZNODE, &pnode->flags) &&
+ test_bit(DIRTY_CNODE, &pnode->flags);
+}
+
+/**
+ * ubifs_change_lp - change LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lp: LEB properties to change
+ * @free: new free space amount
+ * @dirty: new dirty space amount
+ * @flags: new flags
+ * @idx_gc_cnt: change to the count of idx_gc list
+ *
+ * This function changes LEB properties (@free, @dirty or @flag). However, the
+ * property which has the %LPROPS_NC value is not changed. Returns a pointer to
+ * the updated LEB properties on success and a negative error code on failure.
+ *
+ * Note, the LEB properties may have had to be copied (due to COW) and
+ * consequently the pointer returned may not be the same as the pointer
+ * passed.
+ */
+const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
+ const struct ubifs_lprops *lp,
+ int free, int dirty, int flags,
+ int idx_gc_cnt)
+{
+ /*
+ * This is the only function that is allowed to change lprops, so we
+ * discard the const qualifier.
+ */
+ struct ubifs_lprops *lprops = (struct ubifs_lprops *)lp;
+
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d",
+ lprops->lnum, free, dirty, flags);
+
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+ ubifs_assert(c->lst.empty_lebs >= 0 &&
+ c->lst.empty_lebs <= c->main_lebs);
+ ubifs_assert(c->freeable_cnt >= 0);
+ ubifs_assert(c->freeable_cnt <= c->main_lebs);
+ ubifs_assert(c->lst.taken_empty_lebs >= 0);
+ ubifs_assert(c->lst.taken_empty_lebs <= c->lst.empty_lebs);
+ ubifs_assert(!(c->lst.total_free & 7) && !(c->lst.total_dirty & 7));
+ ubifs_assert(!(c->lst.total_dead & 7) && !(c->lst.total_dark & 7));
+ ubifs_assert(!(c->lst.total_used & 7));
+ ubifs_assert(free == LPROPS_NC || free >= 0);
+ ubifs_assert(dirty == LPROPS_NC || dirty >= 0);
+
+ if (!is_lprops_dirty(c, lprops)) {
+ lprops = ubifs_lpt_lookup_dirty(c, lprops->lnum);
+ if (IS_ERR(lprops))
+ return lprops;
+ } else
+ ubifs_assert(lprops == ubifs_lpt_lookup_dirty(c, lprops->lnum));
+
+ ubifs_assert(!(lprops->free & 7) && !(lprops->dirty & 7));
+
+ spin_lock(&c->space_lock);
+ if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size)
+ c->lst.taken_empty_lebs -= 1;
+
+ if (!(lprops->flags & LPROPS_INDEX)) {
+ int old_spc;
+
+ old_spc = lprops->free + lprops->dirty;
+ if (old_spc < c->dead_wm)
+ c->lst.total_dead -= old_spc;
+ else
+ c->lst.total_dark -= calc_dark(c, old_spc);
+
+ c->lst.total_used -= c->leb_size - old_spc;
+ }
+
+ if (free != LPROPS_NC) {
+ free = ALIGN(free, 8);
+ c->lst.total_free += free - lprops->free;
+
+ /* Increase or decrease empty LEBs counter if needed */
+ if (free == c->leb_size) {
+ if (lprops->free != c->leb_size)
+ c->lst.empty_lebs += 1;
+ } else if (lprops->free == c->leb_size)
+ c->lst.empty_lebs -= 1;
+ lprops->free = free;
+ }
+
+ if (dirty != LPROPS_NC) {
+ dirty = ALIGN(dirty, 8);
+ c->lst.total_dirty += dirty - lprops->dirty;
+ lprops->dirty = dirty;
+ }
+
+ if (flags != LPROPS_NC) {
+ /* Take care about indexing LEBs counter if needed */
+ if ((lprops->flags & LPROPS_INDEX)) {
+ if (!(flags & LPROPS_INDEX))
+ c->lst.idx_lebs -= 1;
+ } else if (flags & LPROPS_INDEX)
+ c->lst.idx_lebs += 1;
+ lprops->flags = flags;
+ }
+
+ if (!(lprops->flags & LPROPS_INDEX)) {
+ int new_spc;
+
+ new_spc = lprops->free + lprops->dirty;
+ if (new_spc < c->dead_wm)
+ c->lst.total_dead += new_spc;
+ else
+ c->lst.total_dark += calc_dark(c, new_spc);
+
+ c->lst.total_used += c->leb_size - new_spc;
+ }
+
+ if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size)
+ c->lst.taken_empty_lebs += 1;
+
+ change_category(c, lprops);
+ c->idx_gc_cnt += idx_gc_cnt;
+ spin_unlock(&c->space_lock);
+ return lprops;
+}
+
+/**
+ * ubifs_get_lp_stats - get lprops statistics.
+ * @c: UBIFS file-system description object
+ * @st: return statistics
+ */
+void ubifs_get_lp_stats(struct ubifs_info *c, struct ubifs_lp_stats *lst)
+{
+ spin_lock(&c->space_lock);
+ memcpy(lst, &c->lst, sizeof(struct ubifs_lp_stats));
+ spin_unlock(&c->space_lock);
+}
+
+/**
+ * ubifs_change_one_lp - change LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to change properties for
+ * @free: amount of free space
+ * @dirty: amount of dirty space
+ * @flags_set: flags to set
+ * @flags_clean: flags to clean
+ * @idx_gc_cnt: change to the count of idx_gc list
+ *
+ * This function changes properties of LEB @lnum. It is a helper wrapper over
+ * 'ubifs_change_lp()' which hides lprops get/release. The arguments are the
+ * same as in case of 'ubifs_change_lp()'. Returns zero in case of success and
+ * a negative error code in case of failure.
+ */
+int ubifs_change_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean, int idx_gc_cnt)
+{
+ int err = 0, flags;
+ const struct ubifs_lprops *lp;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ flags = (lp->flags | flags_set) & ~flags_clean;
+ lp = ubifs_change_lp(c, lp, free, dirty, flags, idx_gc_cnt);
+ if (IS_ERR(lp))
+ err = PTR_ERR(lp);
+
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_update_one_lp - update LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to change properties for
+ * @free: amount of free space
+ * @dirty: amount of dirty space to add
+ * @flags_set: flags to set
+ * @flags_clean: flags to clean
+ *
+ * This function is the same as 'ubifs_change_one_lp()' but @dirty is added to
+ * current dirty space, not substitutes it.
+ */
+int ubifs_update_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean)
+{
+ int err = 0, flags;
+ const struct ubifs_lprops *lp;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ flags = (lp->flags | flags_set) & ~flags_clean;
+ lp = ubifs_change_lp(c, lp, free, lp->dirty + dirty, flags, 0);
+ if (IS_ERR(lp))
+ err = PTR_ERR(lp);
+
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_read_one_lp - read LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to read properties for
+ * @lp: where to store read properties
+ *
+ * This helper function reads properties of a LEB @lnum and stores them in @lp.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_read_one_lp(struct ubifs_info *c, int lnum, struct ubifs_lprops *lp)
+{
+ int err = 0;
+ const struct ubifs_lprops *lpp;
+
+ ubifs_get_lprops(c);
+
+ lpp = ubifs_lpt_lookup(c, lnum);
+ if (IS_ERR(lpp)) {
+ err = PTR_ERR(lpp);
+ goto out;
+ }
+
+ memcpy(lp, lpp, sizeof(struct ubifs_lprops));
+
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_fast_find_free - try to find a LEB with free space quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a LEB with free space or %NULL if
+ * the function is unable to find a LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_free(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+ struct ubifs_lpt_heap *heap;
+
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+
+ heap = &c->lpt_heap[LPROPS_FREE - 1];
+ if (heap->cnt == 0)
+ return NULL;
+
+ lprops = heap->arr[0];
+ ubifs_assert(!(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(!(lprops->flags & LPROPS_INDEX));
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_empty - try to find an empty LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for an empty LEB or %NULL if the
+ * function is unable to find an empty LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_empty(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->empty_list))
+ return NULL;
+
+ lprops = list_entry(c->empty_list.next, struct ubifs_lprops, list);
+ ubifs_assert(!(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(!(lprops->flags & LPROPS_INDEX));
+ ubifs_assert(lprops->free == c->leb_size);
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_freeable - try to find a freeable LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a freeable LEB or %NULL if the
+ * function is unable to find a freeable LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_freeable(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->freeable_list))
+ return NULL;
+
+ lprops = list_entry(c->freeable_list.next, struct ubifs_lprops, list);
+ ubifs_assert(!(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(!(lprops->flags & LPROPS_INDEX));
+ ubifs_assert(lprops->free + lprops->dirty == c->leb_size);
+ ubifs_assert(c->freeable_cnt > 0);
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_frdi_idx - try to find a freeable index LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a freeable index LEB or %NULL if the
+ * function is unable to find a freeable index LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->frdi_idx_list))
+ return NULL;
+
+ lprops = list_entry(c->frdi_idx_list.next, struct ubifs_lprops, list);
+ ubifs_assert(!(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert((lprops->flags & LPROPS_INDEX));
+ ubifs_assert(lprops->free + lprops->dirty == c->leb_size);
+ return lprops;
+}
diff --git a/fs/ubifs/lpt.c b/fs/ubifs/lpt.c
new file mode 100644
index 0000000..1a50d4c
--- /dev/null
+++ b/fs/ubifs/lpt.c
@@ -0,0 +1,1105 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements the LEB properties tree (LPT) area. The LPT area
+ * contains the LEB properties tree, a table of LPT area eraseblocks (ltab), and
+ * (for the "big" model) a table of saved LEB numbers (lsave). The LPT area sits
+ * between the log and the orphan area.
+ *
+ * The LPT area is like a miniature self-contained file system. It is required
+ * that it never runs out of space, is fast to access and update, and scales
+ * logarithmically. The LEB properties tree is implemented as a wandering tree
+ * much like the TNC, and the LPT area has its own garbage collection.
+ *
+ * The LPT has two slightly different forms called the "small model" and the
+ * "big model". The small model is used when the entire LEB properties table
+ * can be written into a single eraseblock. In that case, garbage collection
+ * consists of just writing the whole table, which therefore makes all other
+ * eraseblocks reusable. In the case of the big model, dirty eraseblocks are
+ * selected for garbage collection, which consists of marking the clean nodes in
+ * that LEB as dirty, and then only the dirty nodes are written out. Also, in
+ * the case of the big model, a table of LEB numbers is saved so that the entire
+ * LPT does not to be scanned looking for empty eraseblocks when UBIFS is first
+ * mounted.
+ */
+
+#include "ubifs.h"
+#include "crc16.h"
+#include <linux/math64.h>
+
+/**
+ * do_calc_lpt_geom - calculate sizes for the LPT area.
+ * @c: the UBIFS file-system description object
+ *
+ * Calculate the sizes of LPT bit fields, nodes, and tree, based on the
+ * properties of the flash and whether LPT is "big" (c->big_lpt).
+ */
+static void do_calc_lpt_geom(struct ubifs_info *c)
+{
+ int i, n, bits, per_leb_wastage, max_pnode_cnt;
+ long long sz, tot_wastage;
+
+ n = c->main_lebs + c->max_leb_cnt - c->leb_cnt;
+ max_pnode_cnt = DIV_ROUND_UP(n, UBIFS_LPT_FANOUT);
+
+ c->lpt_hght = 1;
+ n = UBIFS_LPT_FANOUT;
+ while (n < max_pnode_cnt) {
+ c->lpt_hght += 1;
+ n <<= UBIFS_LPT_FANOUT_SHIFT;
+ }
+
+ c->pnode_cnt = DIV_ROUND_UP(c->main_lebs, UBIFS_LPT_FANOUT);
+
+ n = DIV_ROUND_UP(c->pnode_cnt, UBIFS_LPT_FANOUT);
+ c->nnode_cnt = n;
+ for (i = 1; i < c->lpt_hght; i++) {
+ n = DIV_ROUND_UP(n, UBIFS_LPT_FANOUT);
+ c->nnode_cnt += n;
+ }
+
+ c->space_bits = fls(c->leb_size) - 3;
+ c->lpt_lnum_bits = fls(c->lpt_lebs);
+ c->lpt_offs_bits = fls(c->leb_size - 1);
+ c->lpt_spc_bits = fls(c->leb_size);
+
+ n = DIV_ROUND_UP(c->max_leb_cnt, UBIFS_LPT_FANOUT);
+ c->pcnt_bits = fls(n - 1);
+
+ c->lnum_bits = fls(c->max_leb_cnt - 1);
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ (c->big_lpt ? c->pcnt_bits : 0) +
+ (c->space_bits * 2 + 1) * UBIFS_LPT_FANOUT;
+ c->pnode_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ (c->big_lpt ? c->pcnt_bits : 0) +
+ (c->lpt_lnum_bits + c->lpt_offs_bits) * UBIFS_LPT_FANOUT;
+ c->nnode_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ c->lpt_lebs * c->lpt_spc_bits * 2;
+ c->ltab_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ c->lnum_bits * c->lsave_cnt;
+ c->lsave_sz = (bits + 7) / 8;
+
+ /* Calculate the minimum LPT size */
+ c->lpt_sz = (long long)c->pnode_cnt * c->pnode_sz;
+ c->lpt_sz += (long long)c->nnode_cnt * c->nnode_sz;
+ c->lpt_sz += c->ltab_sz;
+ if (c->big_lpt)
+ c->lpt_sz += c->lsave_sz;
+
+ /* Add wastage */
+ sz = c->lpt_sz;
+ per_leb_wastage = max_t(int, c->pnode_sz, c->nnode_sz);
+ sz += per_leb_wastage;
+ tot_wastage = per_leb_wastage;
+ while (sz > c->leb_size) {
+ sz += per_leb_wastage;
+ sz -= c->leb_size;
+ tot_wastage += per_leb_wastage;
+ }
+ tot_wastage += ALIGN(sz, c->min_io_size) - sz;
+ c->lpt_sz += tot_wastage;
+}
+
+/**
+ * ubifs_calc_lpt_geom - calculate and check sizes for the LPT area.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_calc_lpt_geom(struct ubifs_info *c)
+{
+ int lebs_needed;
+ long long sz;
+
+ do_calc_lpt_geom(c);
+
+ /* Verify that lpt_lebs is big enough */
+ sz = c->lpt_sz * 2; /* Must have at least 2 times the size */
+ lebs_needed = div_u64(sz + c->leb_size - 1, c->leb_size);
+ if (lebs_needed > c->lpt_lebs) {
+ ubifs_err("too few LPT LEBs");
+ return -EINVAL;
+ }
+
+ /* Verify that ltab fits in a single LEB (since ltab is a single node */
+ if (c->ltab_sz > c->leb_size) {
+ ubifs_err("LPT ltab too big");
+ return -EINVAL;
+ }
+
+ c->check_lpt_free = c->big_lpt;
+ return 0;
+}
+
+/**
+ * ubifs_unpack_bits - unpack bit fields.
+ * @addr: address at which to unpack (passed and next address returned)
+ * @pos: bit position at which to unpack (passed and next position returned)
+ * @nrbits: number of bits of value to unpack (1-32)
+ *
+ * This functions returns the value unpacked.
+ */
+uint32_t ubifs_unpack_bits(uint8_t **addr, int *pos, int nrbits)
+{
+ const int k = 32 - nrbits;
+ uint8_t *p = *addr;
+ int b = *pos;
+ uint32_t uninitialized_var(val);
+ const int bytes = (nrbits + b + 7) >> 3;
+
+ ubifs_assert(nrbits > 0);
+ ubifs_assert(nrbits <= 32);
+ ubifs_assert(*pos >= 0);
+ ubifs_assert(*pos < 8);
+ if (b) {
+ switch (bytes) {
+ case 2:
+ val = p[1];
+ break;
+ case 3:
+ val = p[1] | ((uint32_t)p[2] << 8);
+ break;
+ case 4:
+ val = p[1] | ((uint32_t)p[2] << 8) |
+ ((uint32_t)p[3] << 16);
+ break;
+ case 5:
+ val = p[1] | ((uint32_t)p[2] << 8) |
+ ((uint32_t)p[3] << 16) |
+ ((uint32_t)p[4] << 24);
+ }
+ val <<= (8 - b);
+ val |= *p >> b;
+ nrbits += b;
+ } else {
+ switch (bytes) {
+ case 1:
+ val = p[0];
+ break;
+ case 2:
+ val = p[0] | ((uint32_t)p[1] << 8);
+ break;
+ case 3:
+ val = p[0] | ((uint32_t)p[1] << 8) |
+ ((uint32_t)p[2] << 16);
+ break;
+ case 4:
+ val = p[0] | ((uint32_t)p[1] << 8) |
+ ((uint32_t)p[2] << 16) |
+ ((uint32_t)p[3] << 24);
+ break;
+ }
+ }
+ val <<= k;
+ val >>= k;
+ b = nrbits & 7;
+ p += nrbits >> 3;
+ *addr = p;
+ *pos = b;
+ ubifs_assert((val >> nrbits) == 0 || nrbits - b == 32);
+ return val;
+}
+
+/**
+ * ubifs_add_lpt_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to which to add dirty space
+ * @dirty: amount of dirty space to add
+ */
+void ubifs_add_lpt_dirt(struct ubifs_info *c, int lnum, int dirty)
+{
+ if (!dirty || !lnum)
+ return;
+ dbg_lp("LEB %d add %d to %d",
+ lnum, dirty, c->ltab[lnum - c->lpt_first].dirty);
+ ubifs_assert(lnum >= c->lpt_first && lnum <= c->lpt_last);
+ c->ltab[lnum - c->lpt_first].dirty += dirty;
+}
+
+/**
+ * ubifs_add_nnode_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode for which to add dirt
+ */
+void ubifs_add_nnode_dirt(struct ubifs_info *c, struct ubifs_nnode *nnode)
+{
+ struct ubifs_nnode *np = nnode->parent;
+
+ if (np)
+ ubifs_add_lpt_dirt(c, np->nbranch[nnode->iip].lnum,
+ c->nnode_sz);
+ else {
+ ubifs_add_lpt_dirt(c, c->lpt_lnum, c->nnode_sz);
+ if (!(c->lpt_drty_flgs & LTAB_DIRTY)) {
+ c->lpt_drty_flgs |= LTAB_DIRTY;
+ ubifs_add_lpt_dirt(c, c->ltab_lnum, c->ltab_sz);
+ }
+ }
+}
+
+/**
+ * add_pnode_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode for which to add dirt
+ */
+static void add_pnode_dirt(struct ubifs_info *c, struct ubifs_pnode *pnode)
+{
+ ubifs_add_lpt_dirt(c, pnode->parent->nbranch[pnode->iip].lnum,
+ c->pnode_sz);
+}
+
+/**
+ * calc_nnode_num_from_parent - calculate nnode number.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * The nnode number is a number that uniquely identifies a nnode and can be used
+ * easily to traverse the tree from the root to that nnode.
+ *
+ * This function calculates and returns the nnode number based on the parent's
+ * nnode number and the index in parent.
+ */
+static int calc_nnode_num_from_parent(const struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ int num, shft;
+
+ if (!parent)
+ return 1;
+ shft = (c->lpt_hght - parent->level) * UBIFS_LPT_FANOUT_SHIFT;
+ num = parent->num ^ (1 << shft);
+ num |= (UBIFS_LPT_FANOUT + iip) << shft;
+ return num;
+}
+
+/**
+ * calc_pnode_num_from_parent - calculate pnode number.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * The pnode number is a number that uniquely identifies a pnode and can be used
+ * easily to traverse the tree from the root to that pnode.
+ *
+ * This function calculates and returns the pnode number based on the parent's
+ * nnode number and the index in parent.
+ */
+static int calc_pnode_num_from_parent(const struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i, n = c->lpt_hght - 1, pnum = parent->num, num = 0;
+
+ for (i = 0; i < n; i++) {
+ num <<= UBIFS_LPT_FANOUT_SHIFT;
+ num |= pnum & (UBIFS_LPT_FANOUT - 1);
+ pnum >>= UBIFS_LPT_FANOUT_SHIFT;
+ }
+ num <<= UBIFS_LPT_FANOUT_SHIFT;
+ num |= iip;
+ return num;
+}
+
+/**
+ * update_cats - add LEB properties of a pnode to LEB category lists and heaps.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode
+ *
+ * When a pnode is loaded into memory, the LEB properties it contains are added,
+ * by this function, to the LEB category lists and heaps.
+ */
+static void update_cats(struct ubifs_info *c, struct ubifs_pnode *pnode)
+{
+ int i;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int cat = pnode->lprops[i].flags & LPROPS_CAT_MASK;
+ int lnum = pnode->lprops[i].lnum;
+
+ if (!lnum)
+ return;
+ ubifs_add_to_cat(c, &pnode->lprops[i], cat);
+ }
+}
+
+/**
+ * replace_cats - add LEB properties of a pnode to LEB category lists and heaps.
+ * @c: UBIFS file-system description object
+ * @old_pnode: pnode copied
+ * @new_pnode: pnode copy
+ *
+ * During commit it is sometimes necessary to copy a pnode
+ * (see dirty_cow_pnode). When that happens, references in
+ * category lists and heaps must be replaced. This function does that.
+ */
+static void replace_cats(struct ubifs_info *c, struct ubifs_pnode *old_pnode,
+ struct ubifs_pnode *new_pnode)
+{
+ int i;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (!new_pnode->lprops[i].lnum)
+ return;
+ ubifs_replace_cat(c, &old_pnode->lprops[i],
+ &new_pnode->lprops[i]);
+ }
+}
+
+/**
+ * check_lpt_crc - check LPT node crc is correct.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing node
+ * @len: length of node
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int check_lpt_crc(void *buf, int len)
+{
+ int pos = 0;
+ uint8_t *addr = buf;
+ uint16_t crc, calc_crc;
+
+ crc = ubifs_unpack_bits(&addr, &pos, UBIFS_LPT_CRC_BITS);
+ calc_crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ len - UBIFS_LPT_CRC_BYTES);
+ if (crc != calc_crc) {
+ ubifs_err("invalid crc in LPT node: crc %hx calc %hx", crc,
+ calc_crc);
+ dbg_dump_stack();
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * check_lpt_type - check LPT node type is correct.
+ * @c: UBIFS file-system description object
+ * @addr: address of type bit field is passed and returned updated here
+ * @pos: position of type bit field is passed and returned updated here
+ * @type: expected type
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int check_lpt_type(uint8_t **addr, int *pos, int type)
+{
+ int node_type;
+
+ node_type = ubifs_unpack_bits(addr, pos, UBIFS_LPT_TYPE_BITS);
+ if (node_type != type) {
+ ubifs_err("invalid type (%d) in LPT node type %d", node_type,
+ type);
+ dbg_dump_stack();
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * unpack_pnode - unpack a pnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing packed pnode to unpack
+ * @pnode: pnode structure to fill
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int unpack_pnode(const struct ubifs_info *c, void *buf,
+ struct ubifs_pnode *pnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(&addr, &pos, UBIFS_LPT_PNODE);
+ if (err)
+ return err;
+ if (c->big_lpt)
+ pnode->num = ubifs_unpack_bits(&addr, &pos, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops * const lprops = &pnode->lprops[i];
+
+ lprops->free = ubifs_unpack_bits(&addr, &pos, c->space_bits);
+ lprops->free <<= 3;
+ lprops->dirty = ubifs_unpack_bits(&addr, &pos, c->space_bits);
+ lprops->dirty <<= 3;
+
+ if (ubifs_unpack_bits(&addr, &pos, 1))
+ lprops->flags = LPROPS_INDEX;
+ else
+ lprops->flags = 0;
+ lprops->flags |= ubifs_categorize_lprops(c, lprops);
+ }
+ err = check_lpt_crc(buf, c->pnode_sz);
+ return err;
+}
+
+/**
+ * ubifs_unpack_nnode - unpack a nnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing packed nnode to unpack
+ * @nnode: nnode structure to fill
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_unpack_nnode(const struct ubifs_info *c, void *buf,
+ struct ubifs_nnode *nnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(&addr, &pos, UBIFS_LPT_NNODE);
+ if (err)
+ return err;
+ if (c->big_lpt)
+ nnode->num = ubifs_unpack_bits(&addr, &pos, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int lnum;
+
+ lnum = ubifs_unpack_bits(&addr, &pos, c->lpt_lnum_bits) +
+ c->lpt_first;
+ if (lnum == c->lpt_last + 1)
+ lnum = 0;
+ nnode->nbranch[i].lnum = lnum;
+ nnode->nbranch[i].offs = ubifs_unpack_bits(&addr, &pos,
+ c->lpt_offs_bits);
+ }
+ err = check_lpt_crc(buf, c->nnode_sz);
+ return err;
+}
+
+/**
+ * unpack_ltab - unpack the LPT's own lprops table.
+ * @c: UBIFS file-system description object
+ * @buf: buffer from which to unpack
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int unpack_ltab(const struct ubifs_info *c, void *buf)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(&addr, &pos, UBIFS_LPT_LTAB);
+ if (err)
+ return err;
+ for (i = 0; i < c->lpt_lebs; i++) {
+ int free = ubifs_unpack_bits(&addr, &pos, c->lpt_spc_bits);
+ int dirty = ubifs_unpack_bits(&addr, &pos, c->lpt_spc_bits);
+
+ if (free < 0 || free > c->leb_size || dirty < 0 ||
+ dirty > c->leb_size || free + dirty > c->leb_size)
+ return -EINVAL;
+
+ c->ltab[i].free = free;
+ c->ltab[i].dirty = dirty;
+ c->ltab[i].tgc = 0;
+ c->ltab[i].cmt = 0;
+ }
+ err = check_lpt_crc(buf, c->ltab_sz);
+ return err;
+}
+
+/**
+ * validate_nnode - validate a nnode.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode to validate
+ * @parent: parent nnode (or NULL for the root nnode)
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int validate_nnode(const struct ubifs_info *c, struct ubifs_nnode *nnode,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i, lvl, max_offs;
+
+ if (c->big_lpt) {
+ int num = calc_nnode_num_from_parent(c, parent, iip);
+
+ if (nnode->num != num)
+ return -EINVAL;
+ }
+ lvl = parent ? parent->level - 1 : c->lpt_hght;
+ if (lvl < 1)
+ return -EINVAL;
+ if (lvl == 1)
+ max_offs = c->leb_size - c->pnode_sz;
+ else
+ max_offs = c->leb_size - c->nnode_sz;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int lnum = nnode->nbranch[i].lnum;
+ int offs = nnode->nbranch[i].offs;
+
+ if (lnum == 0) {
+ if (offs != 0)
+ return -EINVAL;
+ continue;
+ }
+ if (lnum < c->lpt_first || lnum > c->lpt_last)
+ return -EINVAL;
+ if (offs < 0 || offs > max_offs)
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * validate_pnode - validate a pnode.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to validate
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int validate_pnode(const struct ubifs_info *c, struct ubifs_pnode *pnode,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i;
+
+ if (c->big_lpt) {
+ int num = calc_pnode_num_from_parent(c, parent, iip);
+
+ if (pnode->num != num)
+ return -EINVAL;
+ }
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int free = pnode->lprops[i].free;
+ int dirty = pnode->lprops[i].dirty;
+
+ if (free < 0 || free > c->leb_size || free % c->min_io_size ||
+ (free & 7))
+ return -EINVAL;
+ if (dirty < 0 || dirty > c->leb_size || (dirty & 7))
+ return -EINVAL;
+ if (dirty + free > c->leb_size)
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * set_pnode_lnum - set LEB numbers on a pnode.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to update
+ *
+ * This function calculates the LEB numbers for the LEB properties it contains
+ * based on the pnode number.
+ */
+static void set_pnode_lnum(const struct ubifs_info *c,
+ struct ubifs_pnode *pnode)
+{
+ int i, lnum;
+
+ lnum = (pnode->num << UBIFS_LPT_FANOUT_SHIFT) + c->main_first;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (lnum >= c->leb_cnt)
+ return;
+ pnode->lprops[i].lnum = lnum++;
+ }
+}
+
+/**
+ * ubifs_read_nnode - read a nnode from flash and link it to the tree in memory.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode (or NULL for the root)
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch = NULL;
+ struct ubifs_nnode *nnode = NULL;
+ void *buf = c->lpt_nod_buf;
+ int err, lnum, offs;
+
+ if (parent) {
+ branch = &parent->nbranch[iip];
+ lnum = branch->lnum;
+ offs = branch->offs;
+ } else {
+ lnum = c->lpt_lnum;
+ offs = c->lpt_offs;
+ }
+ nnode = kzalloc(sizeof(struct ubifs_nnode), GFP_NOFS);
+ if (!nnode) {
+ err = -ENOMEM;
+ goto out;
+ }
+ if (lnum == 0) {
+ /*
+ * This nnode was not written which just means that the LEB
+ * properties in the subtree below it describe empty LEBs. We
+ * make the nnode as though we had read it, which in fact means
+ * doing almost nothing.
+ */
+ if (c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ } else {
+ err = ubi_read(c->ubi, lnum, buf, offs, c->nnode_sz);
+ if (err)
+ goto out;
+ err = ubifs_unpack_nnode(c, buf, nnode);
+ if (err)
+ goto out;
+ }
+ err = validate_nnode(c, nnode, parent, iip);
+ if (err)
+ goto out;
+ if (!c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ if (parent) {
+ branch->nnode = nnode;
+ nnode->level = parent->level - 1;
+ } else {
+ c->nroot = nnode;
+ nnode->level = c->lpt_hght;
+ }
+ nnode->parent = parent;
+ nnode->iip = iip;
+ return 0;
+
+out:
+ ubifs_err("error %d reading nnode at %d:%d", err, lnum, offs);
+ kfree(nnode);
+ return err;
+}
+
+/**
+ * read_pnode - read a pnode from flash and link it to the tree in memory.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int read_pnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_pnode *pnode = NULL;
+ void *buf = c->lpt_nod_buf;
+ int err, lnum, offs;
+
+ branch = &parent->nbranch[iip];
+ lnum = branch->lnum;
+ offs = branch->offs;
+ pnode = kzalloc(sizeof(struct ubifs_pnode), GFP_NOFS);
+ if (!pnode) {
+ err = -ENOMEM;
+ goto out;
+ }
+ if (lnum == 0) {
+ /*
+ * This pnode was not written which just means that the LEB
+ * properties in it describe empty LEBs. We make the pnode as
+ * though we had read it.
+ */
+ int i;
+
+ if (c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops * const lprops = &pnode->lprops[i];
+
+ lprops->free = c->leb_size;
+ lprops->flags = ubifs_categorize_lprops(c, lprops);
+ }
+ } else {
+ err = ubi_read(c->ubi, lnum, buf, offs, c->pnode_sz);
+ if (err)
+ goto out;
+ err = unpack_pnode(c, buf, pnode);
+ if (err)
+ goto out;
+ }
+ err = validate_pnode(c, pnode, parent, iip);
+ if (err)
+ goto out;
+ if (!c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ branch->pnode = pnode;
+ pnode->parent = parent;
+ pnode->iip = iip;
+ set_pnode_lnum(c, pnode);
+ c->pnodes_have += 1;
+ return 0;
+
+out:
+ ubifs_err("error %d reading pnode at %d:%d", err, lnum, offs);
+ dbg_dump_pnode(c, pnode, parent, iip);
+ dbg_msg("calc num: %d", calc_pnode_num_from_parent(c, parent, iip));
+ kfree(pnode);
+ return err;
+}
+
+/**
+ * read_ltab - read LPT's own lprops table.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int read_ltab(struct ubifs_info *c)
+{
+ int err;
+ void *buf;
+
+ buf = vmalloc(c->ltab_sz);
+ if (!buf)
+ return -ENOMEM;
+ err = ubi_read(c->ubi, c->ltab_lnum, buf, c->ltab_offs, c->ltab_sz);
+ if (err)
+ goto out;
+ err = unpack_ltab(c, buf);
+out:
+ vfree(buf);
+ return err;
+}
+
+/**
+ * ubifs_get_nnode - get a nnode.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode (or NULL for the root)
+ * @iip: index in parent
+ *
+ * This function returns a pointer to the nnode on success or a negative error
+ * code on failure.
+ */
+struct ubifs_nnode *ubifs_get_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_nnode *nnode;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ nnode = branch->nnode;
+ if (nnode)
+ return nnode;
+ err = ubifs_read_nnode(c, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ return branch->nnode;
+}
+
+/**
+ * ubifs_get_pnode - get a pnode.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns a pointer to the pnode on success or a negative error
+ * code on failure.
+ */
+struct ubifs_pnode *ubifs_get_pnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_pnode *pnode;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ pnode = branch->pnode;
+ if (pnode)
+ return pnode;
+ err = read_pnode(c, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ update_cats(c, branch->pnode);
+ return branch->pnode;
+}
+
+/**
+ * ubifs_lpt_lookup - lookup LEB properties in the LPT.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to lookup
+ *
+ * This function returns a pointer to the LEB properties on success or a
+ * negative error code on failure.
+ */
+struct ubifs_lprops *ubifs_lpt_lookup(struct ubifs_info *c, int lnum)
+{
+ int err, i, h, iip, shft;
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return ERR_PTR(err);
+ }
+ nnode = c->nroot;
+ i = lnum - c->main_first;
+ shft = c->lpt_hght * UBIFS_LPT_FANOUT_SHIFT;
+ for (h = 1; h < c->lpt_hght; h++) {
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = ubifs_get_nnode(c, nnode, iip);
+ if (IS_ERR(nnode))
+ return ERR_PTR(PTR_ERR(nnode));
+ }
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ pnode = ubifs_get_pnode(c, nnode, iip);
+ if (IS_ERR(pnode))
+ return ERR_PTR(PTR_ERR(pnode));
+ iip = (i & (UBIFS_LPT_FANOUT - 1));
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d", lnum,
+ pnode->lprops[iip].free, pnode->lprops[iip].dirty,
+ pnode->lprops[iip].flags);
+ return &pnode->lprops[iip];
+}
+
+/**
+ * dirty_cow_nnode - ensure a nnode is not being committed.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode to check
+ *
+ * Returns dirtied nnode on success or negative error code on failure.
+ */
+static struct ubifs_nnode *dirty_cow_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *nnode)
+{
+ struct ubifs_nnode *n;
+ int i;
+
+ if (!test_bit(COW_CNODE, &nnode->flags)) {
+ /* nnode is not being committed */
+ if (!test_and_set_bit(DIRTY_CNODE, &nnode->flags)) {
+ c->dirty_nn_cnt += 1;
+ ubifs_add_nnode_dirt(c, nnode);
+ }
+ return nnode;
+ }
+
+ /* nnode is being committed, so copy it */
+ n = kmalloc(sizeof(struct ubifs_nnode), GFP_NOFS);
+ if (unlikely(!n))
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(n, nnode, sizeof(struct ubifs_nnode));
+ n->cnext = NULL;
+ __set_bit(DIRTY_CNODE, &n->flags);
+ __clear_bit(COW_CNODE, &n->flags);
+
+ /* The children now have new parent */
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_nbranch *branch = &n->nbranch[i];
+
+ if (branch->cnode)
+ branch->cnode->parent = n;
+ }
+
+ ubifs_assert(!test_bit(OBSOLETE_CNODE, &nnode->flags));
+ __set_bit(OBSOLETE_CNODE, &nnode->flags);
+
+ c->dirty_nn_cnt += 1;
+ ubifs_add_nnode_dirt(c, nnode);
+ if (nnode->parent)
+ nnode->parent->nbranch[n->iip].nnode = n;
+ else
+ c->nroot = n;
+ return n;
+}
+
+/**
+ * dirty_cow_pnode - ensure a pnode is not being committed.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to check
+ *
+ * Returns dirtied pnode on success or negative error code on failure.
+ */
+static struct ubifs_pnode *dirty_cow_pnode(struct ubifs_info *c,
+ struct ubifs_pnode *pnode)
+{
+ struct ubifs_pnode *p;
+
+ if (!test_bit(COW_CNODE, &pnode->flags)) {
+ /* pnode is not being committed */
+ if (!test_and_set_bit(DIRTY_CNODE, &pnode->flags)) {
+ c->dirty_pn_cnt += 1;
+ add_pnode_dirt(c, pnode);
+ }
+ return pnode;
+ }
+
+ /* pnode is being committed, so copy it */
+ p = kmalloc(sizeof(struct ubifs_pnode), GFP_NOFS);
+ if (unlikely(!p))
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(p, pnode, sizeof(struct ubifs_pnode));
+ p->cnext = NULL;
+ __set_bit(DIRTY_CNODE, &p->flags);
+ __clear_bit(COW_CNODE, &p->flags);
+ replace_cats(c, pnode, p);
+
+ ubifs_assert(!test_bit(OBSOLETE_CNODE, &pnode->flags));
+ __set_bit(OBSOLETE_CNODE, &pnode->flags);
+
+ c->dirty_pn_cnt += 1;
+ add_pnode_dirt(c, pnode);
+ pnode->parent->nbranch[p->iip].pnode = p;
+ return p;
+}
+
+/**
+ * ubifs_lpt_lookup_dirty - lookup LEB properties in the LPT.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to lookup
+ *
+ * This function returns a pointer to the LEB properties on success or a
+ * negative error code on failure.
+ */
+struct ubifs_lprops *ubifs_lpt_lookup_dirty(struct ubifs_info *c, int lnum)
+{
+ int err, i, h, iip, shft;
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return ERR_PTR(err);
+ }
+ nnode = c->nroot;
+ nnode = dirty_cow_nnode(c, nnode);
+ if (IS_ERR(nnode))
+ return ERR_PTR(PTR_ERR(nnode));
+ i = lnum - c->main_first;
+ shft = c->lpt_hght * UBIFS_LPT_FANOUT_SHIFT;
+ for (h = 1; h < c->lpt_hght; h++) {
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = ubifs_get_nnode(c, nnode, iip);
+ if (IS_ERR(nnode))
+ return ERR_PTR(PTR_ERR(nnode));
+ nnode = dirty_cow_nnode(c, nnode);
+ if (IS_ERR(nnode))
+ return ERR_PTR(PTR_ERR(nnode));
+ }
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ pnode = ubifs_get_pnode(c, nnode, iip);
+ if (IS_ERR(pnode))
+ return ERR_PTR(PTR_ERR(pnode));
+ pnode = dirty_cow_pnode(c, pnode);
+ if (IS_ERR(pnode))
+ return ERR_PTR(PTR_ERR(pnode));
+ iip = (i & (UBIFS_LPT_FANOUT - 1));
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d", lnum,
+ pnode->lprops[iip].free, pnode->lprops[iip].dirty,
+ pnode->lprops[iip].flags);
+ ubifs_assert(test_bit(DIRTY_CNODE, &pnode->flags));
+ return &pnode->lprops[iip];
+}
+
+/**
+ * lpt_init_rd - initialize the LPT for reading.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int lpt_init_rd(struct ubifs_info *c)
+{
+ int err, i;
+
+ c->ltab = vmalloc(sizeof(struct ubifs_lpt_lprops) * c->lpt_lebs);
+ if (!c->ltab)
+ return -ENOMEM;
+
+ i = max_t(int, c->nnode_sz, c->pnode_sz);
+ c->lpt_nod_buf = kmalloc(i, GFP_KERNEL);
+ if (!c->lpt_nod_buf)
+ return -ENOMEM;
+
+ for (i = 0; i < LPROPS_HEAP_CNT; i++) {
+ c->lpt_heap[i].arr = kmalloc(sizeof(void *) * LPT_HEAP_SZ,
+ GFP_KERNEL);
+ if (!c->lpt_heap[i].arr)
+ return -ENOMEM;
+ c->lpt_heap[i].cnt = 0;
+ c->lpt_heap[i].max_cnt = LPT_HEAP_SZ;
+ }
+
+ c->dirty_idx.arr = kmalloc(sizeof(void *) * LPT_HEAP_SZ, GFP_KERNEL);
+ if (!c->dirty_idx.arr)
+ return -ENOMEM;
+ c->dirty_idx.cnt = 0;
+ c->dirty_idx.max_cnt = LPT_HEAP_SZ;
+
+ err = read_ltab(c);
+ if (err)
+ return err;
+
+ dbg_lp("space_bits %d", c->space_bits);
+ dbg_lp("lpt_lnum_bits %d", c->lpt_lnum_bits);
+ dbg_lp("lpt_offs_bits %d", c->lpt_offs_bits);
+ dbg_lp("lpt_spc_bits %d", c->lpt_spc_bits);
+ dbg_lp("pcnt_bits %d", c->pcnt_bits);
+ dbg_lp("lnum_bits %d", c->lnum_bits);
+ dbg_lp("pnode_sz %d", c->pnode_sz);
+ dbg_lp("nnode_sz %d", c->nnode_sz);
+ dbg_lp("ltab_sz %d", c->ltab_sz);
+ dbg_lp("lsave_sz %d", c->lsave_sz);
+ dbg_lp("lsave_cnt %d", c->lsave_cnt);
+ dbg_lp("lpt_hght %d", c->lpt_hght);
+ dbg_lp("big_lpt %d", c->big_lpt);
+ dbg_lp("LPT root is at %d:%d", c->lpt_lnum, c->lpt_offs);
+ dbg_lp("LPT head is at %d:%d", c->nhead_lnum, c->nhead_offs);
+ dbg_lp("LPT ltab is at %d:%d", c->ltab_lnum, c->ltab_offs);
+ if (c->big_lpt)
+ dbg_lp("LPT lsave is at %d:%d", c->lsave_lnum, c->lsave_offs);
+
+ return 0;
+}
+
+/**
+ * ubifs_lpt_init - initialize the LPT.
+ * @c: UBIFS file-system description object
+ * @rd: whether to initialize lpt for reading
+ * @wr: whether to initialize lpt for writing
+ *
+ * For mounting 'rw', @rd and @wr are both true. For mounting 'ro', @rd is true
+ * and @wr is false. For mounting from 'ro' to 'rw', @rd is false and @wr is
+ * true.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_lpt_init(struct ubifs_info *c, int rd, int wr)
+{
+ int err;
+
+ if (rd) {
+ err = lpt_init_rd(c);
+ if (err)
+ return err;
+ }
+
+ return 0;
+}
diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
new file mode 100644
index 0000000..c0af818
--- /dev/null
+++ b/fs/ubifs/lpt_commit.c
@@ -0,0 +1,171 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements commit-related functionality of the LEB properties
+ * subsystem.
+ */
+
+#include "crc16.h"
+#include "ubifs.h"
+
+/**
+ * free_obsolete_cnodes - free obsolete cnodes for commit end.
+ * @c: UBIFS file-system description object
+ */
+static void free_obsolete_cnodes(struct ubifs_info *c)
+{
+ struct ubifs_cnode *cnode, *cnext;
+
+ cnext = c->lpt_cnext;
+ if (!cnext)
+ return;
+ do {
+ cnode = cnext;
+ cnext = cnode->cnext;
+ if (test_bit(OBSOLETE_CNODE, &cnode->flags))
+ kfree(cnode);
+ else
+ cnode->cnext = NULL;
+ } while (cnext != c->lpt_cnext);
+ c->lpt_cnext = NULL;
+}
+
+/**
+ * first_nnode - find the first nnode in memory.
+ * @c: UBIFS file-system description object
+ * @hght: height of tree where nnode found is returned here
+ *
+ * This function returns a pointer to the nnode found or %NULL if no nnode is
+ * found. This function is a helper to 'ubifs_lpt_free()'.
+ */
+static struct ubifs_nnode *first_nnode(struct ubifs_info *c, int *hght)
+{
+ struct ubifs_nnode *nnode;
+ int h, i, found;
+
+ nnode = c->nroot;
+ *hght = 0;
+ if (!nnode)
+ return NULL;
+ for (h = 1; h < c->lpt_hght; h++) {
+ found = 0;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (nnode->nbranch[i].nnode) {
+ found = 1;
+ nnode = nnode->nbranch[i].nnode;
+ *hght = h;
+ break;
+ }
+ }
+ if (!found)
+ break;
+ }
+ return nnode;
+}
+
+/**
+ * next_nnode - find the next nnode in memory.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode from which to start.
+ * @hght: height of tree where nnode is, is passed and returned here
+ *
+ * This function returns a pointer to the nnode found or %NULL if no nnode is
+ * found. This function is a helper to 'ubifs_lpt_free()'.
+ */
+static struct ubifs_nnode *next_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *nnode, int *hght)
+{
+ struct ubifs_nnode *parent;
+ int iip, h, i, found;
+
+ parent = nnode->parent;
+ if (!parent)
+ return NULL;
+ if (nnode->iip == UBIFS_LPT_FANOUT - 1) {
+ *hght -= 1;
+ return parent;
+ }
+ for (iip = nnode->iip + 1; iip < UBIFS_LPT_FANOUT; iip++) {
+ nnode = parent->nbranch[iip].nnode;
+ if (nnode)
+ break;
+ }
+ if (!nnode) {
+ *hght -= 1;
+ return parent;
+ }
+ for (h = *hght + 1; h < c->lpt_hght; h++) {
+ found = 0;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (nnode->nbranch[i].nnode) {
+ found = 1;
+ nnode = nnode->nbranch[i].nnode;
+ *hght = h;
+ break;
+ }
+ }
+ if (!found)
+ break;
+ }
+ return nnode;
+}
+
+/**
+ * ubifs_lpt_free - free resources owned by the LPT.
+ * @c: UBIFS file-system description object
+ * @wr_only: free only resources used for writing
+ */
+void ubifs_lpt_free(struct ubifs_info *c, int wr_only)
+{
+ struct ubifs_nnode *nnode;
+ int i, hght;
+
+ /* Free write-only things first */
+
+ free_obsolete_cnodes(c); /* Leftover from a failed commit */
+
+ vfree(c->ltab_cmt);
+ c->ltab_cmt = NULL;
+ vfree(c->lpt_buf);
+ c->lpt_buf = NULL;
+ kfree(c->lsave);
+ c->lsave = NULL;
+
+ if (wr_only)
+ return;
+
+ /* Now free the rest */
+
+ nnode = first_nnode(c, &hght);
+ while (nnode) {
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++)
+ kfree(nnode->nbranch[i].nnode);
+ nnode = next_nnode(c, nnode, &hght);
+ }
+ for (i = 0; i < LPROPS_HEAP_CNT; i++)
+ kfree(c->lpt_heap[i].arr);
+ kfree(c->dirty_idx.arr);
+ kfree(c->nroot);
+ vfree(c->ltab);
+ kfree(c->lpt_nod_buf);
+}
diff --git a/fs/ubifs/master.c b/fs/ubifs/master.c
new file mode 100644
index 0000000..3f2926e
--- /dev/null
+++ b/fs/ubifs/master.c
@@ -0,0 +1,341 @@
+/*
+ * 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 reading and writing the master node */
+
+#include "ubifs.h"
+
+/**
+ * scan_for_master - search the valid master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function scans the master node LEBs and search for the latest master
+ * node. Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int scan_for_master(struct ubifs_info *c)
+{
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ int lnum, offs = 0, nodes_cnt;
+
+ lnum = UBIFS_MST_LNUM;
+
+ sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+ nodes_cnt = sleb->nodes_cnt;
+ if (nodes_cnt > 0) {
+ snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node,
+ list);
+ if (snod->type != UBIFS_MST_NODE)
+ goto out;
+ memcpy(c->mst_node, snod->node, snod->len);
+ offs = snod->offs;
+ }
+ ubifs_scan_destroy(sleb);
+
+ lnum += 1;
+
+ sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+ if (sleb->nodes_cnt != nodes_cnt)
+ goto out;
+ if (!sleb->nodes_cnt)
+ goto out;
+ snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node, list);
+ if (snod->type != UBIFS_MST_NODE)
+ goto out;
+ if (snod->offs != offs)
+ goto out;
+ if (memcmp((void *)c->mst_node + UBIFS_CH_SZ,
+ (void *)snod->node + UBIFS_CH_SZ,
+ UBIFS_MST_NODE_SZ - UBIFS_CH_SZ))
+ goto out;
+ c->mst_offs = offs;
+ ubifs_scan_destroy(sleb);
+ return 0;
+
+out:
+ ubifs_scan_destroy(sleb);
+ return -EINVAL;
+}
+
+/**
+ * validate_master - validate master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function validates data which was read from master node. Returns zero
+ * if the data is all right and %-EINVAL if not.
+ */
+static int validate_master(const struct ubifs_info *c)
+{
+ long long main_sz;
+ int err;
+
+ if (c->max_sqnum >= SQNUM_WATERMARK) {
+ err = 1;
+ goto out;
+ }
+
+ if (c->cmt_no >= c->max_sqnum) {
+ err = 2;
+ goto out;
+ }
+
+ if (c->highest_inum >= INUM_WATERMARK) {
+ err = 3;
+ goto out;
+ }
+
+ if (c->lhead_lnum < UBIFS_LOG_LNUM ||
+ c->lhead_lnum >= UBIFS_LOG_LNUM + c->log_lebs ||
+ c->lhead_offs < 0 || c->lhead_offs >= c->leb_size ||
+ c->lhead_offs & (c->min_io_size - 1)) {
+ err = 4;
+ goto out;
+ }
+
+ if (c->zroot.lnum >= c->leb_cnt || c->zroot.lnum < c->main_first ||
+ c->zroot.offs >= c->leb_size || c->zroot.offs & 7) {
+ err = 5;
+ goto out;
+ }
+
+ if (c->zroot.len < c->ranges[UBIFS_IDX_NODE].min_len ||
+ c->zroot.len > c->ranges[UBIFS_IDX_NODE].max_len) {
+ err = 6;
+ goto out;
+ }
+
+ if (c->gc_lnum >= c->leb_cnt || c->gc_lnum < c->main_first) {
+ err = 7;
+ goto out;
+ }
+
+ if (c->ihead_lnum >= c->leb_cnt || c->ihead_lnum < c->main_first ||
+ c->ihead_offs % c->min_io_size || c->ihead_offs < 0 ||
+ c->ihead_offs > c->leb_size || c->ihead_offs & 7) {
+ err = 8;
+ goto out;
+ }
+
+ main_sz = (long long)c->main_lebs * c->leb_size;
+ if (c->old_idx_sz & 7 || c->old_idx_sz >= main_sz) {
+ err = 9;
+ goto out;
+ }
+
+ if (c->lpt_lnum < c->lpt_first || c->lpt_lnum > c->lpt_last ||
+ c->lpt_offs < 0 || c->lpt_offs + c->nnode_sz > c->leb_size) {
+ err = 10;
+ goto out;
+ }
+
+ if (c->nhead_lnum < c->lpt_first || c->nhead_lnum > c->lpt_last ||
+ c->nhead_offs < 0 || c->nhead_offs % c->min_io_size ||
+ c->nhead_offs > c->leb_size) {
+ err = 11;
+ goto out;
+ }
+
+ if (c->ltab_lnum < c->lpt_first || c->ltab_lnum > c->lpt_last ||
+ c->ltab_offs < 0 ||
+ c->ltab_offs + c->ltab_sz > c->leb_size) {
+ err = 12;
+ goto out;
+ }
+
+ if (c->big_lpt && (c->lsave_lnum < c->lpt_first ||
+ c->lsave_lnum > c->lpt_last || c->lsave_offs < 0 ||
+ c->lsave_offs + c->lsave_sz > c->leb_size)) {
+ err = 13;
+ goto out;
+ }
+
+ if (c->lscan_lnum < c->main_first || c->lscan_lnum >= c->leb_cnt) {
+ err = 14;
+ goto out;
+ }
+
+ if (c->lst.empty_lebs < 0 || c->lst.empty_lebs > c->main_lebs - 2) {
+ err = 15;
+ goto out;
+ }
+
+ if (c->lst.idx_lebs < 0 || c->lst.idx_lebs > c->main_lebs - 1) {
+ err = 16;
+ goto out;
+ }
+
+ if (c->lst.total_free < 0 || c->lst.total_free > main_sz ||
+ c->lst.total_free & 7) {
+ err = 17;
+ goto out;
+ }
+
+ if (c->lst.total_dirty < 0 || (c->lst.total_dirty & 7)) {
+ err = 18;
+ goto out;
+ }
+
+ if (c->lst.total_used < 0 || (c->lst.total_used & 7)) {
+ err = 19;
+ goto out;
+ }
+
+ if (c->lst.total_free + c->lst.total_dirty +
+ c->lst.total_used > main_sz) {
+ err = 20;
+ goto out;
+ }
+
+ if (c->lst.total_dead + c->lst.total_dark +
+ c->lst.total_used + c->old_idx_sz > main_sz) {
+ err = 21;
+ goto out;
+ }
+
+ if (c->lst.total_dead < 0 ||
+ c->lst.total_dead > c->lst.total_free + c->lst.total_dirty ||
+ c->lst.total_dead & 7) {
+ err = 22;
+ goto out;
+ }
+
+ if (c->lst.total_dark < 0 ||
+ c->lst.total_dark > c->lst.total_free + c->lst.total_dirty ||
+ c->lst.total_dark & 7) {
+ err = 23;
+ goto out;
+ }
+
+ return 0;
+
+out:
+ ubifs_err("bad master node at offset %d error %d", c->mst_offs, err);
+ dbg_dump_node(c, c->mst_node);
+ return -EINVAL;
+}
+
+/**
+ * ubifs_read_master - read master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function finds and reads the master node during file-system mount. If
+ * the flash is empty, it creates default master node as well. Returns zero in
+ * case of success and a negative error code in case of failure.
+ */
+int ubifs_read_master(struct ubifs_info *c)
+{
+ int err, old_leb_cnt;
+
+ c->mst_node = kzalloc(c->mst_node_alsz, GFP_KERNEL);
+ if (!c->mst_node)
+ return -ENOMEM;
+
+ err = scan_for_master(c);
+ if (err) {
+ err = ubifs_recover_master_node(c);
+ if (err)
+ /*
+ * Note, we do not free 'c->mst_node' here because the
+ * unmount routine will take care of this.
+ */
+ return err;
+ }
+
+ /* Make sure that the recovery flag is clear */
+ c->mst_node->flags &= cpu_to_le32(~UBIFS_MST_RCVRY);
+
+ c->max_sqnum = le64_to_cpu(c->mst_node->ch.sqnum);
+ c->highest_inum = le64_to_cpu(c->mst_node->highest_inum);
+ c->cmt_no = le64_to_cpu(c->mst_node->cmt_no);
+ c->zroot.lnum = le32_to_cpu(c->mst_node->root_lnum);
+ c->zroot.offs = le32_to_cpu(c->mst_node->root_offs);
+ c->zroot.len = le32_to_cpu(c->mst_node->root_len);
+ c->lhead_lnum = le32_to_cpu(c->mst_node->log_lnum);
+ c->gc_lnum = le32_to_cpu(c->mst_node->gc_lnum);
+ c->ihead_lnum = le32_to_cpu(c->mst_node->ihead_lnum);
+ c->ihead_offs = le32_to_cpu(c->mst_node->ihead_offs);
+ c->old_idx_sz = le64_to_cpu(c->mst_node->index_size);
+ c->lpt_lnum = le32_to_cpu(c->mst_node->lpt_lnum);
+ c->lpt_offs = le32_to_cpu(c->mst_node->lpt_offs);
+ c->nhead_lnum = le32_to_cpu(c->mst_node->nhead_lnum);
+ c->nhead_offs = le32_to_cpu(c->mst_node->nhead_offs);
+ c->ltab_lnum = le32_to_cpu(c->mst_node->ltab_lnum);
+ c->ltab_offs = le32_to_cpu(c->mst_node->ltab_offs);
+ c->lsave_lnum = le32_to_cpu(c->mst_node->lsave_lnum);
+ c->lsave_offs = le32_to_cpu(c->mst_node->lsave_offs);
+ c->lscan_lnum = le32_to_cpu(c->mst_node->lscan_lnum);
+ c->lst.empty_lebs = le32_to_cpu(c->mst_node->empty_lebs);
+ c->lst.idx_lebs = le32_to_cpu(c->mst_node->idx_lebs);
+ old_leb_cnt = le32_to_cpu(c->mst_node->leb_cnt);
+ c->lst.total_free = le64_to_cpu(c->mst_node->total_free);
+ c->lst.total_dirty = le64_to_cpu(c->mst_node->total_dirty);
+ c->lst.total_used = le64_to_cpu(c->mst_node->total_used);
+ c->lst.total_dead = le64_to_cpu(c->mst_node->total_dead);
+ c->lst.total_dark = le64_to_cpu(c->mst_node->total_dark);
+
+ c->calc_idx_sz = c->old_idx_sz;
+
+ if (c->mst_node->flags & cpu_to_le32(UBIFS_MST_NO_ORPHS))
+ c->no_orphs = 1;
+
+ if (old_leb_cnt != c->leb_cnt) {
+ /* The file system has been resized */
+ int growth = c->leb_cnt - old_leb_cnt;
+
+ if (c->leb_cnt < old_leb_cnt ||
+ c->leb_cnt < UBIFS_MIN_LEB_CNT) {
+ ubifs_err("bad leb_cnt on master node");
+ dbg_dump_node(c, c->mst_node);
+ return -EINVAL;
+ }
+
+ dbg_mnt("Auto resizing (master) from %d LEBs to %d LEBs",
+ old_leb_cnt, c->leb_cnt);
+ c->lst.empty_lebs += growth;
+ c->lst.total_free += growth * (long long)c->leb_size;
+ c->lst.total_dark += growth * (long long)c->dark_wm;
+
+ /*
+ * Reflect changes back onto the master node. N.B. the master
+ * node gets written immediately whenever mounting (or
+ * remounting) in read-write mode, so we do not need to write it
+ * here.
+ */
+ c->mst_node->leb_cnt = cpu_to_le32(c->leb_cnt);
+ c->mst_node->empty_lebs = cpu_to_le32(c->lst.empty_lebs);
+ c->mst_node->total_free = cpu_to_le64(c->lst.total_free);
+ c->mst_node->total_dark = cpu_to_le64(c->lst.total_dark);
+ }
+
+ err = validate_master(c);
+ if (err)
+ return err;
+
+ err = dbg_old_index_check_init(c, &c->zroot);
+
+ return err;
+}
diff --git a/fs/ubifs/misc.h b/fs/ubifs/misc.h
new file mode 100644
index 0000000..b745d86
--- /dev/null
+++ b/fs/ubifs/misc.h
@@ -0,0 +1,310 @@
+/*
+ * 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 contains miscellaneous helper functions.
+ */
+
+#ifndef __UBIFS_MISC_H__
+#define __UBIFS_MISC_H__
+
+/**
+ * ubifs_zn_dirty - check if znode is dirty.
+ * @znode: znode to check
+ *
+ * This helper function returns %1 if @znode is dirty and %0 otherwise.
+ */
+static inline int ubifs_zn_dirty(const struct ubifs_znode *znode)
+{
+ return !!test_bit(DIRTY_ZNODE, &znode->flags);
+}
+
+/**
+ * ubifs_wake_up_bgt - wake up background thread.
+ * @c: UBIFS file-system description object
+ */
+static inline void ubifs_wake_up_bgt(struct ubifs_info *c)
+{
+ if (c->bgt && !c->need_bgt) {
+ c->need_bgt = 1;
+ wake_up_process(c->bgt);
+ }
+}
+
+/**
+ * ubifs_tnc_find_child - find next child in znode.
+ * @znode: znode to search at
+ * @start: the zbranch index to start at
+ *
+ * This helper function looks for znode child starting at index @start. Returns
+ * the child or %NULL if no children were found.
+ */
+static inline struct ubifs_znode *
+ubifs_tnc_find_child(struct ubifs_znode *znode, int start)
+{
+ while (start < znode->child_cnt) {
+ if (znode->zbranch[start].znode)
+ return znode->zbranch[start].znode;
+ start += 1;
+ }
+
+ return NULL;
+}
+
+/**
+ * ubifs_inode - get UBIFS inode information by VFS 'struct inode' object.
+ * @inode: the VFS 'struct inode' pointer
+ */
+static inline struct ubifs_inode *ubifs_inode(const struct inode *inode)
+{
+ return container_of(inode, struct ubifs_inode, vfs_inode);
+}
+
+/**
+ * ubifs_compr_present - check if compressor was compiled in.
+ * @compr_type: compressor type to check
+ *
+ * This function returns %1 of compressor of type @compr_type is present, and
+ * %0 if not.
+ */
+static inline int ubifs_compr_present(int compr_type)
+{
+ ubifs_assert(compr_type >= 0 && compr_type < UBIFS_COMPR_TYPES_CNT);
+ return !!ubifs_compressors[compr_type]->capi_name;
+}
+
+/**
+ * ubifs_compr_name - get compressor name string by its type.
+ * @compr_type: compressor type
+ *
+ * This function returns compressor type string.
+ */
+static inline const char *ubifs_compr_name(int compr_type)
+{
+ ubifs_assert(compr_type >= 0 && compr_type < UBIFS_COMPR_TYPES_CNT);
+ return ubifs_compressors[compr_type]->name;
+}
+
+/**
+ * ubifs_wbuf_sync - synchronize write-buffer.
+ * @wbuf: write-buffer to synchronize
+ *
+ * This is the same as as 'ubifs_wbuf_sync_nolock()' but it does not assume
+ * that the write-buffer is already locked.
+ */
+static inline int ubifs_wbuf_sync(struct ubifs_wbuf *wbuf)
+{
+ int err;
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ mutex_unlock(&wbuf->io_mutex);
+ return err;
+}
+
+/**
+ * ubifs_leb_unmap - unmap an LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to unmap
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static inline int ubifs_leb_unmap(const struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ if (c->ro_media)
+ return -EROFS;
+ err = ubi_leb_unmap(c->ubi, lnum);
+ if (err) {
+ ubifs_err("unmap LEB %d failed, error %d", lnum, err);
+ return err;
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_leb_write - write to a LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to write
+ * @buf: buffer to write from
+ * @offs: offset within LEB to write to
+ * @len: length to write
+ * @dtype: data type
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static inline int ubifs_leb_write(const struct ubifs_info *c, int lnum,
+ const void *buf, int offs, int len, int dtype)
+{
+ int err;
+
+ if (c->ro_media)
+ return -EROFS;
+ err = ubi_leb_write(c->ubi, lnum, buf, offs, len, dtype);
+ if (err) {
+ ubifs_err("writing %d bytes at %d:%d, error %d",
+ len, lnum, offs, err);
+ return err;
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_leb_change - atomic LEB change.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to write
+ * @buf: buffer to write from
+ * @len: length to write
+ * @dtype: data type
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static inline int ubifs_leb_change(const struct ubifs_info *c, int lnum,
+ const void *buf, int len, int dtype)
+{
+ int err;
+
+ if (c->ro_media)
+ return -EROFS;
+ err = ubi_leb_change(c->ubi, lnum, buf, len, dtype);
+ if (err) {
+ ubifs_err("changing %d bytes in LEB %d, error %d",
+ len, lnum, err);
+ return err;
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_add_dirt - add dirty space to LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to add dirty space for
+ * @dirty: dirty space to add
+ *
+ * This is a helper function which increased amount of dirty LEB space. Returns
+ * zero in case of success and a negative error code in case of failure.
+ */
+static inline int ubifs_add_dirt(struct ubifs_info *c, int lnum, int dirty)
+{
+ return ubifs_update_one_lp(c, lnum, LPROPS_NC, dirty, 0, 0);
+}
+
+/**
+ * ubifs_return_leb - return LEB to lprops.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to return
+ *
+ * This helper function cleans the "taken" flag of a logical eraseblock in the
+ * lprops. Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static inline int ubifs_return_leb(struct ubifs_info *c, int lnum)
+{
+ return ubifs_change_one_lp(c, lnum, LPROPS_NC, LPROPS_NC, 0,
+ LPROPS_TAKEN, 0);
+}
+
+/**
+ * ubifs_idx_node_sz - return index node size.
+ * @c: the UBIFS file-system description object
+ * @child_cnt: number of children of this index node
+ */
+static inline int ubifs_idx_node_sz(const struct ubifs_info *c, int child_cnt)
+{
+ return UBIFS_IDX_NODE_SZ + (UBIFS_BRANCH_SZ + c->key_len) * child_cnt;
+}
+
+/**
+ * ubifs_idx_branch - return pointer to an index branch.
+ * @c: the UBIFS file-system description object
+ * @idx: index node
+ * @bnum: branch number
+ */
+static inline
+struct ubifs_branch *ubifs_idx_branch(const struct ubifs_info *c,
+ const struct ubifs_idx_node *idx,
+ int bnum)
+{
+ return (struct ubifs_branch *)((void *)idx->branches +
+ (UBIFS_BRANCH_SZ + c->key_len) * bnum);
+}
+
+/**
+ * ubifs_idx_key - return pointer to an index key.
+ * @c: the UBIFS file-system description object
+ * @idx: index node
+ */
+static inline void *ubifs_idx_key(const struct ubifs_info *c,
+ const struct ubifs_idx_node *idx)
+{
+ return (void *)((struct ubifs_branch *)idx->branches)->key;
+}
+
+/**
+ * ubifs_tnc_lookup - look up a file-system node.
+ * @c: UBIFS file-system description object
+ * @key: node key to lookup
+ * @node: the node is returned here
+ *
+ * This function look up and reads node with key @key. The caller has to make
+ * sure the @node buffer is large enough to fit the node. Returns zero in case
+ * of success, %-ENOENT if the node was not found, and a negative error code in
+ * case of failure.
+ */
+static inline int ubifs_tnc_lookup(struct ubifs_info *c,
+ const union ubifs_key *key, void *node)
+{
+ return ubifs_tnc_locate(c, key, node, NULL, NULL);
+}
+
+/**
+ * ubifs_get_lprops - get reference to LEB properties.
+ * @c: the UBIFS file-system description object
+ *
+ * This function locks lprops. Lprops have to be unlocked by
+ * 'ubifs_release_lprops()'.
+ */
+static inline void ubifs_get_lprops(struct ubifs_info *c)
+{
+ mutex_lock(&c->lp_mutex);
+}
+
+/**
+ * ubifs_release_lprops - release lprops lock.
+ * @c: the UBIFS file-system description object
+ *
+ * This function has to be called after each 'ubifs_get_lprops()' call to
+ * unlock lprops.
+ */
+static inline void ubifs_release_lprops(struct ubifs_info *c)
+{
+ ubifs_assert(mutex_is_locked(&c->lp_mutex));
+ ubifs_assert(c->lst.empty_lebs >= 0 &&
+ c->lst.empty_lebs <= c->main_lebs);
+ mutex_unlock(&c->lp_mutex);
+}
+
+#endif /* __UBIFS_MISC_H__ */
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
new file mode 100644
index 0000000..d091031
--- /dev/null
+++ b/fs/ubifs/orphan.c
@@ -0,0 +1,316 @@
+/*
+ * 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
+ *
+ * Author: Adrian Hunter
+ */
+
+#include "ubifs.h"
+
+/*
+ * An orphan is an inode number whose inode node has been committed to the index
+ * with a link count of zero. That happens when an open file is deleted
+ * (unlinked) and then a commit is run. In the normal course of events the inode
+ * would be deleted when the file is closed. However in the case of an unclean
+ * unmount, orphans need to be accounted for. After an unclean unmount, the
+ * orphans' inodes must be deleted which means either scanning the entire index
+ * looking for them, or keeping a list on flash somewhere. This unit implements
+ * the latter approach.
+ *
+ * The orphan area is a fixed number of LEBs situated between the LPT area and
+ * the main area. The number of orphan area LEBs is specified when the file
+ * system is created. The minimum number is 1. The size of the orphan area
+ * should be so that it can hold the maximum number of orphans that are expected
+ * to ever exist at one time.
+ *
+ * The number of orphans that can fit in a LEB is:
+ *
+ * (c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64)
+ *
+ * For example: a 15872 byte LEB can fit 1980 orphans so 1 LEB may be enough.
+ *
+ * Orphans are accumulated in a rb-tree. When an inode's link count drops to
+ * zero, the inode number is added to the rb-tree. It is removed from the tree
+ * when the inode is deleted. Any new orphans that are in the orphan tree when
+ * the commit is run, are written to the orphan area in 1 or more orphan nodes.
+ * If the orphan area is full, it is consolidated to make space. There is
+ * always enough space because validation prevents the user from creating more
+ * than the maximum number of orphans allowed.
+ */
+
+/**
+ * tot_avail_orphs - calculate total space.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns the number of orphans that can be written in half
+ * the total space. That leaves half the space for adding new orphans.
+ */
+static int tot_avail_orphs(struct ubifs_info *c)
+{
+ int avail_lebs, avail;
+
+ avail_lebs = c->orph_lebs;
+ avail = avail_lebs *
+ ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
+ return avail / 2;
+}
+
+/**
+ * ubifs_clear_orphans - erase all LEBs used for orphans.
+ * @c: UBIFS file-system description object
+ *
+ * If recovery is not required, then the orphans from the previous session
+ * are not needed. This function locates the LEBs used to record
+ * orphans, and un-maps them.
+ */
+int ubifs_clear_orphans(struct ubifs_info *c)
+{
+ int lnum, err;
+
+ for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
+ err = ubifs_leb_unmap(c, lnum);
+ if (err)
+ return err;
+ }
+ c->ohead_lnum = c->orph_first;
+ c->ohead_offs = 0;
+ return 0;
+}
+
+/**
+ * insert_dead_orphan - insert an orphan.
+ * @c: UBIFS file-system description object
+ * @inum: orphan inode number
+ *
+ * This function is a helper to the 'do_kill_orphans()' function. The orphan
+ * must be kept until the next commit, so it is added to the rb-tree and the
+ * deletion list.
+ */
+static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
+{
+ struct ubifs_orphan *orphan, *o;
+ struct rb_node **p, *parent = NULL;
+
+ orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
+ if (!orphan)
+ return -ENOMEM;
+ orphan->inum = inum;
+
+ p = &c->orph_tree.rb_node;
+ while (*p) {
+ parent = *p;
+ o = rb_entry(parent, struct ubifs_orphan, rb);
+ if (inum < o->inum)
+ p = &(*p)->rb_left;
+ else if (inum > o->inum)
+ p = &(*p)->rb_right;
+ else {
+ /* Already added - no problem */
+ kfree(orphan);
+ return 0;
+ }
+ }
+ c->tot_orphans += 1;
+ rb_link_node(&orphan->rb, parent, p);
+ rb_insert_color(&orphan->rb, &c->orph_tree);
+ list_add_tail(&orphan->list, &c->orph_list);
+ orphan->dnext = c->orph_dnext;
+ c->orph_dnext = orphan;
+ dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
+ c->new_orphans, c->tot_orphans);
+ return 0;
+}
+
+/**
+ * do_kill_orphans - remove orphan inodes from the index.
+ * @c: UBIFS file-system description object
+ * @sleb: scanned LEB
+ * @last_cmt_no: cmt_no of last orphan node read is passed and returned here
+ * @outofdate: whether the LEB is out of date is returned here
+ * @last_flagged: whether the end orphan node is encountered
+ *
+ * This function is a helper to the 'kill_orphans()' function. It goes through
+ * every orphan node in a LEB and for every inode number recorded, removes
+ * all keys for that inode from the TNC.
+ */
+static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ unsigned long long *last_cmt_no, int *outofdate,
+ int *last_flagged)
+{
+ struct ubifs_scan_node *snod;
+ struct ubifs_orph_node *orph;
+ unsigned long long cmt_no;
+ ino_t inum;
+ int i, n, err, first = 1;
+
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ if (snod->type != UBIFS_ORPH_NODE) {
+ ubifs_err("invalid node type %d in orphan area at "
+ "%d:%d", snod->type, sleb->lnum, snod->offs);
+ dbg_dump_node(c, snod->node);
+ return -EINVAL;
+ }
+
+ orph = snod->node;
+
+ /* Check commit number */
+ cmt_no = le64_to_cpu(orph->cmt_no) & LLONG_MAX;
+ /*
+ * The commit number on the master node may be less, because
+ * of a failed commit. If there are several failed commits in a
+ * row, the commit number written on orphan nodes will continue
+ * to increase (because the commit number is adjusted here) even
+ * though the commit number on the master node stays the same
+ * because the master node has not been re-written.
+ */
+ if (cmt_no > c->cmt_no)
+ c->cmt_no = cmt_no;
+ if (cmt_no < *last_cmt_no && *last_flagged) {
+ /*
+ * The last orphan node had a higher commit number and
+ * was flagged as the last written for that commit
+ * number. That makes this orphan node, out of date.
+ */
+ if (!first) {
+ ubifs_err("out of order commit number %llu in "
+ "orphan node at %d:%d",
+ cmt_no, sleb->lnum, snod->offs);
+ dbg_dump_node(c, snod->node);
+ return -EINVAL;
+ }
+ dbg_rcvry("out of date LEB %d", sleb->lnum);
+ *outofdate = 1;
+ return 0;
+ }
+
+ if (first)
+ first = 0;
+
+ n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
+ for (i = 0; i < n; i++) {
+ inum = le64_to_cpu(orph->inos[i]);
+ dbg_rcvry("deleting orphaned inode %lu",
+ (unsigned long)inum);
+ err = ubifs_tnc_remove_ino(c, inum);
+ if (err)
+ return err;
+ err = insert_dead_orphan(c, inum);
+ if (err)
+ return err;
+ }
+
+ *last_cmt_no = cmt_no;
+ if (le64_to_cpu(orph->cmt_no) & (1ULL << 63)) {
+ dbg_rcvry("last orph node for commit %llu at %d:%d",
+ cmt_no, sleb->lnum, snod->offs);
+ *last_flagged = 1;
+ } else
+ *last_flagged = 0;
+ }
+
+ return 0;
+}
+
+/**
+ * kill_orphans - remove all orphan inodes from the index.
+ * @c: UBIFS file-system description object
+ *
+ * If recovery is required, then orphan inodes recorded during the previous
+ * session (which ended with an unclean unmount) must be deleted from the index.
+ * This is done by updating the TNC, but since the index is not updated until
+ * the next commit, the LEBs where the orphan information is recorded are not
+ * erased until the next commit.
+ */
+static int kill_orphans(struct ubifs_info *c)
+{
+ unsigned long long last_cmt_no = 0;
+ int lnum, err = 0, outofdate = 0, last_flagged = 0;
+
+ c->ohead_lnum = c->orph_first;
+ c->ohead_offs = 0;
+ /* Check no-orphans flag and skip this if no orphans */
+ if (c->no_orphs) {
+ dbg_rcvry("no orphans");
+ return 0;
+ }
+ /*
+ * Orph nodes always start at c->orph_first and are written to each
+ * successive LEB in turn. Generally unused LEBs will have been unmapped
+ * but may contain out of date orphan nodes if the unmap didn't go
+ * through. In addition, the last orphan node written for each commit is
+ * marked (top bit of orph->cmt_no is set to 1). It is possible that
+ * there are orphan nodes from the next commit (i.e. the commit did not
+ * complete successfully). In that case, no orphans will have been lost
+ * due to the way that orphans are written, and any orphans added will
+ * be valid orphans anyway and so can be deleted.
+ */
+ for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
+ struct ubifs_scan_leb *sleb;
+
+ dbg_rcvry("LEB %d", lnum);
+ sleb = ubifs_scan(c, lnum, 0, c->sbuf);
+ if (IS_ERR(sleb)) {
+ sleb = ubifs_recover_leb(c, lnum, 0, c->sbuf, 0);
+ if (IS_ERR(sleb)) {
+ err = PTR_ERR(sleb);
+ break;
+ }
+ }
+ err = do_kill_orphans(c, sleb, &last_cmt_no, &outofdate,
+ &last_flagged);
+ if (err || outofdate) {
+ ubifs_scan_destroy(sleb);
+ break;
+ }
+ if (sleb->endpt) {
+ c->ohead_lnum = lnum;
+ c->ohead_offs = sleb->endpt;
+ }
+ ubifs_scan_destroy(sleb);
+ }
+ return err;
+}
+
+/**
+ * ubifs_mount_orphans - delete orphan inodes and erase LEBs that recorded them.
+ * @c: UBIFS file-system description object
+ * @unclean: indicates recovery from unclean unmount
+ * @read_only: indicates read only mount
+ *
+ * This function is called when mounting to erase orphans from the previous
+ * session. If UBIFS was not unmounted cleanly, then the inodes recorded as
+ * orphans are deleted.
+ */
+int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only)
+{
+ int err = 0;
+
+ c->max_orphans = tot_avail_orphs(c);
+
+ if (!read_only) {
+ c->orph_buf = vmalloc(c->leb_size);
+ if (!c->orph_buf)
+ return -ENOMEM;
+ }
+
+ if (unclean)
+ err = kill_orphans(c);
+ else if (!read_only)
+ err = ubifs_clear_orphans(c);
+
+ return err;
+}
diff --git a/fs/ubifs/recovery.c b/fs/ubifs/recovery.c
new file mode 100644
index 0000000..fe3b364
--- /dev/null
+++ b/fs/ubifs/recovery.c
@@ -0,0 +1,1249 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements functions needed to recover from unclean un-mounts.
+ * When UBIFS is mounted, it checks a flag on the master node to determine if
+ * an un-mount was completed sucessfully. If not, the process of mounting
+ * incorparates additional checking and fixing of on-flash data structures.
+ * UBIFS always cleans away all remnants of an unclean un-mount, so that
+ * errors do not accumulate. However UBIFS defers recovery if it is mounted
+ * read-only, and the flash is not modified in that case.
+ */
+
+#include "ubifs.h"
+
+/**
+ * is_empty - determine whether a buffer is empty (contains all 0xff).
+ * @buf: buffer to clean
+ * @len: length of buffer
+ *
+ * This function returns %1 if the buffer is empty (contains all 0xff) otherwise
+ * %0 is returned.
+ */
+static int is_empty(void *buf, int len)
+{
+ uint8_t *p = buf;
+ int i;
+
+ for (i = 0; i < len; i++)
+ if (*p++ != 0xff)
+ return 0;
+ return 1;
+}
+
+/**
+ * get_master_node - get the last valid master node allowing for corruption.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number
+ * @pbuf: buffer containing the LEB read, is returned here
+ * @mst: master node, if found, is returned here
+ * @cor: corruption, if found, is returned here
+ *
+ * This function allocates a buffer, reads the LEB into it, and finds and
+ * returns the last valid master node allowing for one area of corruption.
+ * The corrupt area, if there is one, must be consistent with the assumption
+ * that it is the result of an unclean unmount while the master node was being
+ * written. Under those circumstances, it is valid to use the previously written
+ * master node.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int get_master_node(const struct ubifs_info *c, int lnum, void **pbuf,
+ struct ubifs_mst_node **mst, void **cor)
+{
+ const int sz = c->mst_node_alsz;
+ int err, offs, len;
+ void *sbuf, *buf;
+
+ sbuf = vmalloc(c->leb_size);
+ if (!sbuf)
+ return -ENOMEM;
+
+ err = ubi_read(c->ubi, lnum, sbuf, 0, c->leb_size);
+ if (err && err != -EBADMSG)
+ goto out_free;
+
+ /* Find the first position that is definitely not a node */
+ offs = 0;
+ buf = sbuf;
+ len = c->leb_size;
+ while (offs + UBIFS_MST_NODE_SZ <= c->leb_size) {
+ struct ubifs_ch *ch = buf;
+
+ if (le32_to_cpu(ch->magic) != UBIFS_NODE_MAGIC)
+ break;
+ offs += sz;
+ buf += sz;
+ len -= sz;
+ }
+ /* See if there was a valid master node before that */
+ if (offs) {
+ int ret;
+
+ offs -= sz;
+ buf -= sz;
+ len += sz;
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1);
+ if (ret != SCANNED_A_NODE && offs) {
+ /* Could have been corruption so check one place back */
+ offs -= sz;
+ buf -= sz;
+ len += sz;
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1);
+ if (ret != SCANNED_A_NODE)
+ /*
+ * We accept only one area of corruption because
+ * we are assuming that it was caused while
+ * trying to write a master node.
+ */
+ goto out_err;
+ }
+ if (ret == SCANNED_A_NODE) {
+ struct ubifs_ch *ch = buf;
+
+ if (ch->node_type != UBIFS_MST_NODE)
+ goto out_err;
+ dbg_rcvry("found a master node at %d:%d", lnum, offs);
+ *mst = buf;
+ offs += sz;
+ buf += sz;
+ len -= sz;
+ }
+ }
+ /* Check for corruption */
+ if (offs < c->leb_size) {
+ if (!is_empty(buf, min_t(int, len, sz))) {
+ *cor = buf;
+ dbg_rcvry("found corruption at %d:%d", lnum, offs);
+ }
+ offs += sz;
+ buf += sz;
+ len -= sz;
+ }
+ /* Check remaining empty space */
+ if (offs < c->leb_size)
+ if (!is_empty(buf, len))
+ goto out_err;
+ *pbuf = sbuf;
+ return 0;
+
+out_err:
+ err = -EINVAL;
+out_free:
+ vfree(sbuf);
+ *mst = NULL;
+ *cor = NULL;
+ return err;
+}
+
+/**
+ * write_rcvrd_mst_node - write recovered master node.
+ * @c: UBIFS file-system description object
+ * @mst: master node
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int write_rcvrd_mst_node(struct ubifs_info *c,
+ struct ubifs_mst_node *mst)
+{
+ int err = 0, lnum = UBIFS_MST_LNUM, sz = c->mst_node_alsz;
+ __le32 save_flags;
+
+ dbg_rcvry("recovery");
+
+ save_flags = mst->flags;
+ mst->flags |= cpu_to_le32(UBIFS_MST_RCVRY);
+
+ ubifs_prepare_node(c, mst, UBIFS_MST_NODE_SZ, 1);
+ err = ubi_leb_change(c->ubi, lnum, mst, sz, UBI_SHORTTERM);
+ if (err)
+ goto out;
+ err = ubi_leb_change(c->ubi, lnum + 1, mst, sz, UBI_SHORTTERM);
+ if (err)
+ goto out;
+out:
+ mst->flags = save_flags;
+ return err;
+}
+
+/**
+ * ubifs_recover_master_node - recover the master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function recovers the master node from corruption that may occur due to
+ * an unclean unmount.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_recover_master_node(struct ubifs_info *c)
+{
+ void *buf1 = NULL, *buf2 = NULL, *cor1 = NULL, *cor2 = NULL;
+ struct ubifs_mst_node *mst1 = NULL, *mst2 = NULL, *mst;
+ const int sz = c->mst_node_alsz;
+ int err, offs1, offs2;
+
+ dbg_rcvry("recovery");
+
+ err = get_master_node(c, UBIFS_MST_LNUM, &buf1, &mst1, &cor1);
+ if (err)
+ goto out_free;
+
+ err = get_master_node(c, UBIFS_MST_LNUM + 1, &buf2, &mst2, &cor2);
+ if (err)
+ goto out_free;
+
+ if (mst1) {
+ offs1 = (void *)mst1 - buf1;
+ if ((le32_to_cpu(mst1->flags) & UBIFS_MST_RCVRY) &&
+ (offs1 == 0 && !cor1)) {
+ /*
+ * mst1 was written by recovery at offset 0 with no
+ * corruption.
+ */
+ dbg_rcvry("recovery recovery");
+ mst = mst1;
+ } else if (mst2) {
+ offs2 = (void *)mst2 - buf2;
+ if (offs1 == offs2) {
+ /* Same offset, so must be the same */
+ if (memcmp((void *)mst1 + UBIFS_CH_SZ,
+ (void *)mst2 + UBIFS_CH_SZ,
+ UBIFS_MST_NODE_SZ - UBIFS_CH_SZ))
+ goto out_err;
+ mst = mst1;
+ } else if (offs2 + sz == offs1) {
+ /* 1st LEB was written, 2nd was not */
+ if (cor1)
+ goto out_err;
+ mst = mst1;
+ } else if (offs1 == 0 && offs2 + sz >= c->leb_size) {
+ /* 1st LEB was unmapped and written, 2nd not */
+ if (cor1)
+ goto out_err;
+ mst = mst1;
+ } else
+ goto out_err;
+ } else {
+ /*
+ * 2nd LEB was unmapped and about to be written, so
+ * there must be only one master node in the first LEB
+ * and no corruption.
+ */
+ if (offs1 != 0 || cor1)
+ goto out_err;
+ mst = mst1;
+ }
+ } else {
+ if (!mst2)
+ goto out_err;
+ /*
+ * 1st LEB was unmapped and about to be written, so there must
+ * be no room left in 2nd LEB.
+ */
+ offs2 = (void *)mst2 - buf2;
+ if (offs2 + sz + sz <= c->leb_size)
+ goto out_err;
+ mst = mst2;
+ }
+
+ dbg_rcvry("recovered master node from LEB %d",
+ (mst == mst1 ? UBIFS_MST_LNUM : UBIFS_MST_LNUM + 1));
+
+ memcpy(c->mst_node, mst, UBIFS_MST_NODE_SZ);
+
+ if ((c->vfs_sb->s_flags & MS_RDONLY)) {
+ /* Read-only mode. Keep a copy for switching to rw mode */
+ c->rcvrd_mst_node = kmalloc(sz, GFP_KERNEL);
+ if (!c->rcvrd_mst_node) {
+ err = -ENOMEM;
+ goto out_free;
+ }
+ memcpy(c->rcvrd_mst_node, c->mst_node, UBIFS_MST_NODE_SZ);
+ }
+
+ vfree(buf2);
+ vfree(buf1);
+
+ return 0;
+
+out_err:
+ err = -EINVAL;
+out_free:
+ ubifs_err("failed to recover master node");
+ if (mst1) {
+ dbg_err("dumping first master node");
+ dbg_dump_node(c, mst1);
+ }
+ if (mst2) {
+ dbg_err("dumping second master node");
+ dbg_dump_node(c, mst2);
+ }
+ vfree(buf2);
+ vfree(buf1);
+ return err;
+}
+
+/**
+ * ubifs_write_rcvrd_mst_node - write the recovered master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function writes the master node that was recovered during mounting in
+ * read-only mode and must now be written because we are remounting rw.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_write_rcvrd_mst_node(struct ubifs_info *c)
+{
+ int err;
+
+ if (!c->rcvrd_mst_node)
+ return 0;
+ c->rcvrd_mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY);
+ c->mst_node->flags |= cpu_to_le32(UBIFS_MST_DIRTY);
+ err = write_rcvrd_mst_node(c, c->rcvrd_mst_node);
+ if (err)
+ return err;
+ kfree(c->rcvrd_mst_node);
+ c->rcvrd_mst_node = NULL;
+ return 0;
+}
+
+/**
+ * is_last_write - determine if an offset was in the last write to a LEB.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to check
+ * @offs: offset to check
+ *
+ * This function returns %1 if @offs was in the last write to the LEB whose data
+ * is in @buf, otherwise %0 is returned. The determination is made by checking
+ * for subsequent empty space starting from the next min_io_size boundary (or a
+ * bit less than the common header size if min_io_size is one).
+ */
+static int is_last_write(const struct ubifs_info *c, void *buf, int offs)
+{
+ int empty_offs;
+ int check_len;
+ uint8_t *p;
+
+ if (c->min_io_size == 1) {
+ check_len = c->leb_size - offs;
+ p = buf + check_len;
+ for (; check_len > 0; check_len--)
+ if (*--p != 0xff)
+ break;
+ /*
+ * 'check_len' is the size of the corruption which cannot be
+ * more than the size of 1 node if it was caused by an unclean
+ * unmount.
+ */
+ if (check_len > UBIFS_MAX_NODE_SZ)
+ return 0;
+ return 1;
+ }
+
+ /*
+ * Round up to the next c->min_io_size boundary i.e. 'offs' is in the
+ * last wbuf written. After that should be empty space.
+ */
+ empty_offs = ALIGN(offs + 1, c->min_io_size);
+ check_len = c->leb_size - empty_offs;
+ p = buf + empty_offs - offs;
+
+ for (; check_len > 0; check_len--)
+ if (*p++ != 0xff)
+ return 0;
+ return 1;
+}
+
+/**
+ * clean_buf - clean the data from an LEB sitting in a buffer.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to clean
+ * @lnum: LEB number to clean
+ * @offs: offset from which to clean
+ * @len: length of buffer
+ *
+ * This function pads up to the next min_io_size boundary (if there is one) and
+ * sets empty space to all 0xff. @buf, @offs and @len are updated to the next
+ * min_io_size boundary (if there is one).
+ */
+static void clean_buf(const struct ubifs_info *c, void **buf, int lnum,
+ int *offs, int *len)
+{
+ int empty_offs, pad_len;
+
+ lnum = lnum;
+ dbg_rcvry("cleaning corruption at %d:%d", lnum, *offs);
+
+ if (c->min_io_size == 1) {
+ memset(*buf, 0xff, c->leb_size - *offs);
+ return;
+ }
+
+ ubifs_assert(!(*offs & 7));
+ empty_offs = ALIGN(*offs, c->min_io_size);
+ pad_len = empty_offs - *offs;
+ ubifs_pad(c, *buf, pad_len);
+ *offs += pad_len;
+ *buf += pad_len;
+ *len -= pad_len;
+ memset(*buf, 0xff, c->leb_size - empty_offs);
+}
+
+/**
+ * no_more_nodes - determine if there are no more nodes in a buffer.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to check
+ * @len: length of buffer
+ * @lnum: LEB number of the LEB from which @buf was read
+ * @offs: offset from which @buf was read
+ *
+ * This function scans @buf for more nodes and returns %0 is a node is found and
+ * %1 if no more nodes are found.
+ */
+static int no_more_nodes(const struct ubifs_info *c, void *buf, int len,
+ int lnum, int offs)
+{
+ int skip, next_offs = 0;
+
+ if (len > UBIFS_DATA_NODE_SZ) {
+ struct ubifs_ch *ch = buf;
+ int dlen = le32_to_cpu(ch->len);
+
+ if (ch->node_type == UBIFS_DATA_NODE && dlen >= UBIFS_CH_SZ &&
+ dlen <= UBIFS_MAX_DATA_NODE_SZ)
+ /* The corrupt node looks like a data node */
+ next_offs = ALIGN(offs + dlen, 8);
+ }
+
+ if (c->min_io_size == 1)
+ skip = 8;
+ else
+ skip = ALIGN(offs + 1, c->min_io_size) - offs;
+
+ offs += skip;
+ buf += skip;
+ len -= skip;
+ while (len > 8) {
+ struct ubifs_ch *ch = buf;
+ uint32_t magic = le32_to_cpu(ch->magic);
+ int ret;
+
+ if (magic == UBIFS_NODE_MAGIC) {
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 1);
+ if (ret == SCANNED_A_NODE || ret > 0) {
+ /*
+ * There is a small chance this is just data in
+ * a data node, so check that possibility. e.g.
+ * this is part of a file that itself contains
+ * a UBIFS image.
+ */
+ if (next_offs && offs + le32_to_cpu(ch->len) <=
+ next_offs)
+ continue;
+ dbg_rcvry("unexpected node at %d:%d", lnum,
+ offs);
+ return 0;
+ }
+ }
+ offs += 8;
+ buf += 8;
+ len -= 8;
+ }
+ return 1;
+}
+
+/**
+ * fix_unclean_leb - fix an unclean LEB.
+ * @c: UBIFS file-system description object
+ * @sleb: scanned LEB information
+ * @start: offset where scan started
+ */
+static int fix_unclean_leb(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ int start)
+{
+ int lnum = sleb->lnum, endpt = start;
+
+ /* Get the end offset of the last node we are keeping */
+ if (!list_empty(&sleb->nodes)) {
+ struct ubifs_scan_node *snod;
+
+ snod = list_entry(sleb->nodes.prev,
+ struct ubifs_scan_node, list);
+ endpt = snod->offs + snod->len;
+ }
+
+ if ((c->vfs_sb->s_flags & MS_RDONLY) && !c->remounting_rw) {
+ /* Add to recovery list */
+ struct ubifs_unclean_leb *ucleb;
+
+ dbg_rcvry("need to fix LEB %d start %d endpt %d",
+ lnum, start, sleb->endpt);
+ ucleb = kzalloc(sizeof(struct ubifs_unclean_leb), GFP_NOFS);
+ if (!ucleb)
+ return -ENOMEM;
+ ucleb->lnum = lnum;
+ ucleb->endpt = endpt;
+ list_add_tail(&ucleb->list, &c->unclean_leb_list);
+ }
+ return 0;
+}
+
+/**
+ * drop_incomplete_group - drop nodes from an incomplete group.
+ * @sleb: scanned LEB information
+ * @offs: offset of dropped nodes is returned here
+ *
+ * This function returns %1 if nodes are dropped and %0 otherwise.
+ */
+static int drop_incomplete_group(struct ubifs_scan_leb *sleb, int *offs)
+{
+ int dropped = 0;
+
+ while (!list_empty(&sleb->nodes)) {
+ struct ubifs_scan_node *snod;
+ struct ubifs_ch *ch;
+
+ snod = list_entry(sleb->nodes.prev, struct ubifs_scan_node,
+ list);
+ ch = snod->node;
+ if (ch->group_type != UBIFS_IN_NODE_GROUP)
+ return dropped;
+ dbg_rcvry("dropping node at %d:%d", sleb->lnum, snod->offs);
+ *offs = snod->offs;
+ list_del(&snod->list);
+ kfree(snod);
+ sleb->nodes_cnt -= 1;
+ dropped = 1;
+ }
+ return dropped;
+}
+
+/**
+ * ubifs_recover_leb - scan and recover a LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number
+ * @offs: offset
+ * @sbuf: LEB-sized buffer to use
+ * @grouped: nodes may be grouped for recovery
+ *
+ * This function does a scan of a LEB, but caters for errors that might have
+ * been caused by the unclean unmount from which we are attempting to recover.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
+ int offs, void *sbuf, int grouped)
+{
+ int err, len = c->leb_size - offs, need_clean = 0, quiet = 1;
+ int empty_chkd = 0, start = offs;
+ struct ubifs_scan_leb *sleb;
+ void *buf = sbuf + offs;
+
+ dbg_rcvry("%d:%d", lnum, offs);
+
+ sleb = ubifs_start_scan(c, lnum, offs, sbuf);
+ if (IS_ERR(sleb))
+ return sleb;
+
+ if (sleb->ecc)
+ need_clean = 1;
+
+ while (len >= 8) {
+ int ret;
+
+ dbg_scan("look at LEB %d:%d (%d bytes left)",
+ lnum, offs, len);
+
+ cond_resched();
+
+ /*
+ * Scan quietly until there is an error from which we cannot
+ * recover
+ */
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, quiet);
+
+ if (ret == SCANNED_A_NODE) {
+ /* A valid node, and not a padding node */
+ struct ubifs_ch *ch = buf;
+ int node_len;
+
+ err = ubifs_add_snod(c, sleb, buf, offs);
+ if (err)
+ goto error;
+ node_len = ALIGN(le32_to_cpu(ch->len), 8);
+ offs += node_len;
+ buf += node_len;
+ len -= node_len;
+ continue;
+ }
+
+ if (ret > 0) {
+ /* Padding bytes or a valid padding node */
+ offs += ret;
+ buf += ret;
+ len -= ret;
+ continue;
+ }
+
+ if (ret == SCANNED_EMPTY_SPACE) {
+ if (!is_empty(buf, len)) {
+ if (!is_last_write(c, buf, offs))
+ break;
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ }
+ empty_chkd = 1;
+ break;
+ }
+
+ if (ret == SCANNED_GARBAGE || ret == SCANNED_A_BAD_PAD_NODE)
+ if (is_last_write(c, buf, offs)) {
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ empty_chkd = 1;
+ break;
+ }
+
+ if (ret == SCANNED_A_CORRUPT_NODE)
+ if (no_more_nodes(c, buf, len, lnum, offs)) {
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ empty_chkd = 1;
+ break;
+ }
+
+ if (quiet) {
+ /* Redo the last scan but noisily */
+ quiet = 0;
+ continue;
+ }
+
+ switch (ret) {
+ case SCANNED_GARBAGE:
+ dbg_err("garbage");
+ goto corrupted;
+ case SCANNED_A_CORRUPT_NODE:
+ case SCANNED_A_BAD_PAD_NODE:
+ dbg_err("bad node");
+ goto corrupted;
+ default:
+ dbg_err("unknown");
+ goto corrupted;
+ }
+ }
+
+ if (!empty_chkd && !is_empty(buf, len)) {
+ if (is_last_write(c, buf, offs)) {
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ } else {
+ ubifs_err("corrupt empty space at LEB %d:%d",
+ lnum, offs);
+ goto corrupted;
+ }
+ }
+
+ /* Drop nodes from incomplete group */
+ if (grouped && drop_incomplete_group(sleb, &offs)) {
+ buf = sbuf + offs;
+ len = c->leb_size - offs;
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ }
+
+ if (offs % c->min_io_size) {
+ clean_buf(c, &buf, lnum, &offs, &len);
+ need_clean = 1;
+ }
+
+ ubifs_end_scan(c, sleb, lnum, offs);
+
+ if (need_clean) {
+ err = fix_unclean_leb(c, sleb, start);
+ if (err)
+ goto error;
+ }
+
+ return sleb;
+
+corrupted:
+ ubifs_scanned_corruption(c, lnum, offs, buf);
+ err = -EUCLEAN;
+error:
+ ubifs_err("LEB %d scanning failed", lnum);
+ ubifs_scan_destroy(sleb);
+ return ERR_PTR(err);
+}
+
+/**
+ * get_cs_sqnum - get commit start sequence number.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of commit start node
+ * @offs: offset of commit start node
+ * @cs_sqnum: commit start sequence number is returned here
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int get_cs_sqnum(struct ubifs_info *c, int lnum, int offs,
+ unsigned long long *cs_sqnum)
+{
+ struct ubifs_cs_node *cs_node = NULL;
+ int err, ret;
+
+ dbg_rcvry("at %d:%d", lnum, offs);
+ cs_node = kmalloc(UBIFS_CS_NODE_SZ, GFP_KERNEL);
+ if (!cs_node)
+ return -ENOMEM;
+ if (c->leb_size - offs < UBIFS_CS_NODE_SZ)
+ goto out_err;
+ err = ubi_read(c->ubi, lnum, (void *)cs_node, offs, UBIFS_CS_NODE_SZ);
+ if (err && err != -EBADMSG)
+ goto out_free;
+ ret = ubifs_scan_a_node(c, cs_node, UBIFS_CS_NODE_SZ, lnum, offs, 0);
+ if (ret != SCANNED_A_NODE) {
+ dbg_err("Not a valid node");
+ goto out_err;
+ }
+ if (cs_node->ch.node_type != UBIFS_CS_NODE) {
+ dbg_err("Node a CS node, type is %d", cs_node->ch.node_type);
+ goto out_err;
+ }
+ if (le64_to_cpu(cs_node->cmt_no) != c->cmt_no) {
+ dbg_err("CS node cmt_no %llu != current cmt_no %llu",
+ (unsigned long long)le64_to_cpu(cs_node->cmt_no),
+ c->cmt_no);
+ goto out_err;
+ }
+ *cs_sqnum = le64_to_cpu(cs_node->ch.sqnum);
+ dbg_rcvry("commit start sqnum %llu", *cs_sqnum);
+ kfree(cs_node);
+ return 0;
+
+out_err:
+ err = -EINVAL;
+out_free:
+ ubifs_err("failed to get CS sqnum");
+ kfree(cs_node);
+ return err;
+}
+
+/**
+ * ubifs_recover_log_leb - scan and recover a log LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number
+ * @offs: offset
+ * @sbuf: LEB-sized buffer to use
+ *
+ * This function does a scan of a LEB, but caters for errors that might have
+ * been caused by the unclean unmount from which we are attempting to recover.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
+ int offs, void *sbuf)
+{
+ struct ubifs_scan_leb *sleb;
+ int next_lnum;
+
+ dbg_rcvry("LEB %d", lnum);
+ next_lnum = lnum + 1;
+ if (next_lnum >= UBIFS_LOG_LNUM + c->log_lebs)
+ next_lnum = UBIFS_LOG_LNUM;
+ if (next_lnum != c->ltail_lnum) {
+ /*
+ * We can only recover at the end of the log, so check that the
+ * next log LEB is empty or out of date.
+ */
+ sleb = ubifs_scan(c, next_lnum, 0, sbuf);
+ if (IS_ERR(sleb))
+ return sleb;
+ if (sleb->nodes_cnt) {
+ struct ubifs_scan_node *snod;
+ unsigned long long cs_sqnum = c->cs_sqnum;
+
+ snod = list_entry(sleb->nodes.next,
+ struct ubifs_scan_node, list);
+ if (cs_sqnum == 0) {
+ int err;
+
+ err = get_cs_sqnum(c, lnum, offs, &cs_sqnum);
+ if (err) {
+ ubifs_scan_destroy(sleb);
+ return ERR_PTR(err);
+ }
+ }
+ if (snod->sqnum > cs_sqnum) {
+ ubifs_err("unrecoverable log corruption "
+ "in LEB %d", lnum);
+ ubifs_scan_destroy(sleb);
+ return ERR_PTR(-EUCLEAN);
+ }
+ }
+ ubifs_scan_destroy(sleb);
+ }
+ return ubifs_recover_leb(c, lnum, offs, sbuf, 0);
+}
+
+/**
+ * recover_head - recover a head.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of head to recover
+ * @offs: offset of head to recover
+ * @sbuf: LEB-sized buffer to use
+ *
+ * This function ensures that there is no data on the flash at a head location.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int recover_head(const struct ubifs_info *c, int lnum, int offs,
+ void *sbuf)
+{
+ int len, err, need_clean = 0;
+
+ if (c->min_io_size > 1)
+ len = c->min_io_size;
+ else
+ len = 512;
+ if (offs + len > c->leb_size)
+ len = c->leb_size - offs;
+
+ if (!len)
+ return 0;
+
+ /* Read at the head location and check it is empty flash */
+ err = ubi_read(c->ubi, lnum, sbuf, offs, len);
+ if (err)
+ need_clean = 1;
+ else {
+ uint8_t *p = sbuf;
+
+ while (len--)
+ if (*p++ != 0xff) {
+ need_clean = 1;
+ break;
+ }
+ }
+
+ if (need_clean) {
+ dbg_rcvry("cleaning head at %d:%d", lnum, offs);
+ if (offs == 0)
+ return ubifs_leb_unmap(c, lnum);
+ err = ubi_read(c->ubi, lnum, sbuf, 0, offs);
+ if (err)
+ return err;
+ return ubi_leb_change(c->ubi, lnum, sbuf, offs, UBI_UNKNOWN);
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_recover_inl_heads - recover index and LPT heads.
+ * @c: UBIFS file-system description object
+ * @sbuf: LEB-sized buffer to use
+ *
+ * This function ensures that there is no data on the flash at the index and
+ * LPT head locations.
+ *
+ * This deals with the recovery of a half-completed journal commit. UBIFS is
+ * careful never to overwrite the last version of the index or the LPT. Because
+ * the index and LPT are wandering trees, data from a half-completed commit will
+ * not be referenced anywhere in UBIFS. The data will be either in LEBs that are
+ * assumed to be empty and will be unmapped anyway before use, or in the index
+ * and LPT heads.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf)
+{
+ int err;
+
+ ubifs_assert(!(c->vfs_sb->s_flags & MS_RDONLY) || c->remounting_rw);
+
+ dbg_rcvry("checking index head at %d:%d", c->ihead_lnum, c->ihead_offs);
+ err = recover_head(c, c->ihead_lnum, c->ihead_offs, sbuf);
+ if (err)
+ return err;
+
+ dbg_rcvry("checking LPT head at %d:%d", c->nhead_lnum, c->nhead_offs);
+ err = recover_head(c, c->nhead_lnum, c->nhead_offs, sbuf);
+ if (err)
+ return err;
+
+ return 0;
+}
+
+/**
+ * clean_an_unclean_leb - read and write a LEB to remove corruption.
+ * @c: UBIFS file-system description object
+ * @ucleb: unclean LEB information
+ * @sbuf: LEB-sized buffer to use
+ *
+ * This function reads a LEB up to a point pre-determined by the mount recovery,
+ * checks the nodes, and writes the result back to the flash, thereby cleaning
+ * off any following corruption, or non-fatal ECC errors.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int clean_an_unclean_leb(const struct ubifs_info *c,
+ struct ubifs_unclean_leb *ucleb, void *sbuf)
+{
+ int err, lnum = ucleb->lnum, offs = 0, len = ucleb->endpt, quiet = 1;
+ void *buf = sbuf;
+
+ dbg_rcvry("LEB %d len %d", lnum, len);
+
+ if (len == 0) {
+ /* Nothing to read, just unmap it */
+ err = ubifs_leb_unmap(c, lnum);
+ if (err)
+ return err;
+ return 0;
+ }
+
+ err = ubi_read(c->ubi, lnum, buf, offs, len);
+ if (err && err != -EBADMSG)
+ return err;
+
+ while (len >= 8) {
+ int ret;
+
+ cond_resched();
+
+ /* Scan quietly until there is an error */
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, quiet);
+
+ if (ret == SCANNED_A_NODE) {
+ /* A valid node, and not a padding node */
+ struct ubifs_ch *ch = buf;
+ int node_len;
+
+ node_len = ALIGN(le32_to_cpu(ch->len), 8);
+ offs += node_len;
+ buf += node_len;
+ len -= node_len;
+ continue;
+ }
+
+ if (ret > 0) {
+ /* Padding bytes or a valid padding node */
+ offs += ret;
+ buf += ret;
+ len -= ret;
+ continue;
+ }
+
+ if (ret == SCANNED_EMPTY_SPACE) {
+ ubifs_err("unexpected empty space at %d:%d",
+ lnum, offs);
+ return -EUCLEAN;
+ }
+
+ if (quiet) {
+ /* Redo the last scan but noisily */
+ quiet = 0;
+ continue;
+ }
+
+ ubifs_scanned_corruption(c, lnum, offs, buf);
+ return -EUCLEAN;
+ }
+
+ /* Pad to min_io_size */
+ len = ALIGN(ucleb->endpt, c->min_io_size);
+ if (len > ucleb->endpt) {
+ int pad_len = len - ALIGN(ucleb->endpt, 8);
+
+ if (pad_len > 0) {
+ buf = c->sbuf + len - pad_len;
+ ubifs_pad(c, buf, pad_len);
+ }
+ }
+
+ /* Write back the LEB atomically */
+ err = ubi_leb_change(c->ubi, lnum, sbuf, len, UBI_UNKNOWN);
+ if (err)
+ return err;
+
+ dbg_rcvry("cleaned LEB %d", lnum);
+
+ return 0;
+}
+
+/**
+ * ubifs_clean_lebs - clean LEBs recovered during read-only mount.
+ * @c: UBIFS file-system description object
+ * @sbuf: LEB-sized buffer to use
+ *
+ * This function cleans a LEB identified during recovery that needs to be
+ * written but was not because UBIFS was mounted read-only. This happens when
+ * remounting to read-write mode.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_clean_lebs(const struct ubifs_info *c, void *sbuf)
+{
+ dbg_rcvry("recovery");
+ while (!list_empty(&c->unclean_leb_list)) {
+ struct ubifs_unclean_leb *ucleb;
+ int err;
+
+ ucleb = list_entry(c->unclean_leb_list.next,
+ struct ubifs_unclean_leb, list);
+ err = clean_an_unclean_leb(c, ucleb, sbuf);
+ if (err)
+ return err;
+ list_del(&ucleb->list);
+ kfree(ucleb);
+ }
+ return 0;
+}
+
+/**
+ * struct size_entry - inode size information for recovery.
+ * @rb: link in the RB-tree of sizes
+ * @inum: inode number
+ * @i_size: size on inode
+ * @d_size: maximum size based on data nodes
+ * @exists: indicates whether the inode exists
+ * @inode: inode if pinned in memory awaiting rw mode to fix it
+ */
+struct size_entry {
+ struct rb_node rb;
+ ino_t inum;
+ loff_t i_size;
+ loff_t d_size;
+ int exists;
+ struct inode *inode;
+};
+
+/**
+ * add_ino - add an entry to the size tree.
+ * @c: UBIFS file-system description object
+ * @inum: inode number
+ * @i_size: size on inode
+ * @d_size: maximum size based on data nodes
+ * @exists: indicates whether the inode exists
+ */
+static int add_ino(struct ubifs_info *c, ino_t inum, loff_t i_size,
+ loff_t d_size, int exists)
+{
+ struct rb_node **p = &c->size_tree.rb_node, *parent = NULL;
+ struct size_entry *e;
+
+ while (*p) {
+ parent = *p;
+ e = rb_entry(parent, struct size_entry, rb);
+ if (inum < e->inum)
+ p = &(*p)->rb_left;
+ else
+ p = &(*p)->rb_right;
+ }
+
+ e = kzalloc(sizeof(struct size_entry), GFP_KERNEL);
+ if (!e)
+ return -ENOMEM;
+
+ e->inum = inum;
+ e->i_size = i_size;
+ e->d_size = d_size;
+ e->exists = exists;
+
+ rb_link_node(&e->rb, parent, p);
+ rb_insert_color(&e->rb, &c->size_tree);
+
+ return 0;
+}
+
+/**
+ * find_ino - find an entry on the size tree.
+ * @c: UBIFS file-system description object
+ * @inum: inode number
+ */
+static struct size_entry *find_ino(struct ubifs_info *c, ino_t inum)
+{
+ struct rb_node *p = c->size_tree.rb_node;
+ struct size_entry *e;
+
+ while (p) {
+ e = rb_entry(p, struct size_entry, rb);
+ if (inum < e->inum)
+ p = p->rb_left;
+ else if (inum > e->inum)
+ p = p->rb_right;
+ else
+ return e;
+ }
+ return NULL;
+}
+
+/**
+ * remove_ino - remove an entry from the size tree.
+ * @c: UBIFS file-system description object
+ * @inum: inode number
+ */
+static void remove_ino(struct ubifs_info *c, ino_t inum)
+{
+ struct size_entry *e = find_ino(c, inum);
+
+ if (!e)
+ return;
+ rb_erase(&e->rb, &c->size_tree);
+ kfree(e);
+}
+
+/**
+ * ubifs_recover_size_accum - accumulate inode sizes for recovery.
+ * @c: UBIFS file-system description object
+ * @key: node key
+ * @deletion: node is for a deletion
+ * @new_size: inode size
+ *
+ * This function has two purposes:
+ * 1) to ensure there are no data nodes that fall outside the inode size
+ * 2) to ensure there are no data nodes for inodes that do not exist
+ * To accomplish those purposes, a rb-tree is constructed containing an entry
+ * for each inode number in the journal that has not been deleted, and recording
+ * the size from the inode node, the maximum size of any data node (also altered
+ * by truncations) and a flag indicating a inode number for which no inode node
+ * was present in the journal.
+ *
+ * Note that there is still the possibility that there are data nodes that have
+ * been committed that are beyond the inode size, however the only way to find
+ * them would be to scan the entire index. Alternatively, some provision could
+ * be made to record the size of inodes at the start of commit, which would seem
+ * very cumbersome for a scenario that is quite unlikely and the only negative
+ * consequence of which is wasted space.
+ *
+ * This functions returns %0 on success and a negative error code on failure.
+ */
+int ubifs_recover_size_accum(struct ubifs_info *c, union ubifs_key *key,
+ int deletion, loff_t new_size)
+{
+ ino_t inum = key_inum(c, key);
+ struct size_entry *e;
+ int err;
+
+ switch (key_type(c, key)) {
+ case UBIFS_INO_KEY:
+ if (deletion)
+ remove_ino(c, inum);
+ else {
+ e = find_ino(c, inum);
+ if (e) {
+ e->i_size = new_size;
+ e->exists = 1;
+ } else {
+ err = add_ino(c, inum, new_size, 0, 1);
+ if (err)
+ return err;
+ }
+ }
+ break;
+ case UBIFS_DATA_KEY:
+ e = find_ino(c, inum);
+ if (e) {
+ if (new_size > e->d_size)
+ e->d_size = new_size;
+ } else {
+ err = add_ino(c, inum, 0, new_size, 0);
+ if (err)
+ return err;
+ }
+ break;
+ case UBIFS_TRUN_KEY:
+ e = find_ino(c, inum);
+ if (e)
+ e->d_size = new_size;
+ break;
+ }
+ return 0;
+}
+
+/**
+ * ubifs_recover_size - recover inode size.
+ * @c: UBIFS file-system description object
+ *
+ * This function attempts to fix inode size discrepancies identified by the
+ * 'ubifs_recover_size_accum()' function.
+ *
+ * This functions returns %0 on success and a negative error code on failure.
+ */
+int ubifs_recover_size(struct ubifs_info *c)
+{
+ struct rb_node *this = rb_first(&c->size_tree);
+
+ while (this) {
+ struct size_entry *e;
+ int err;
+
+ e = rb_entry(this, struct size_entry, rb);
+ if (!e->exists) {
+ union ubifs_key key;
+
+ ino_key_init(c, &key, e->inum);
+ err = ubifs_tnc_lookup(c, &key, c->sbuf);
+ if (err && err != -ENOENT)
+ return err;
+ if (err == -ENOENT) {
+ /* Remove data nodes that have no inode */
+ dbg_rcvry("removing ino %lu",
+ (unsigned long)e->inum);
+ err = ubifs_tnc_remove_ino(c, e->inum);
+ if (err)
+ return err;
+ } else {
+ struct ubifs_ino_node *ino = c->sbuf;
+
+ e->exists = 1;
+ e->i_size = le64_to_cpu(ino->size);
+ }
+ }
+ if (e->exists && e->i_size < e->d_size) {
+ if (!e->inode && (c->vfs_sb->s_flags & MS_RDONLY)) {
+ /* Fix the inode size and pin it in memory */
+ struct inode *inode;
+
+ inode = ubifs_iget(c->vfs_sb, e->inum);
+ if (IS_ERR(inode))
+ return PTR_ERR(inode);
+ if (inode->i_size < e->d_size) {
+ dbg_rcvry("ino %lu size %lld -> %lld",
+ (unsigned long)e->inum,
+ e->d_size, inode->i_size);
+ inode->i_size = e->d_size;
+ ubifs_inode(inode)->ui_size = e->d_size;
+ e->inode = inode;
+ this = rb_next(this);
+ continue;
+ }
+ iput(inode);
+ }
+ }
+ this = rb_next(this);
+ rb_erase(&e->rb, &c->size_tree);
+ kfree(e);
+ }
+ return 0;
+}
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
new file mode 100644
index 0000000..da33a14
--- /dev/null
+++ b/fs/ubifs/replay.c
@@ -0,0 +1,1070 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file contains journal replay code. It runs when the file-system is being
+ * mounted and requires no locking.
+ *
+ * The larger is the journal, the longer it takes to scan it, so the longer it
+ * takes to mount UBIFS. This is why the journal has limited size which may be
+ * changed depending on the system requirements. But a larger journal gives
+ * faster I/O speed because it writes the index less frequently. So this is a
+ * trade-off. Also, the journal is indexed by the in-memory index (TNC), so the
+ * larger is the journal, the more memory its index may consume.
+ */
+
+#include "ubifs.h"
+
+/*
+ * Replay flags.
+ *
+ * REPLAY_DELETION: node was deleted
+ * REPLAY_REF: node is a reference node
+ */
+enum {
+ REPLAY_DELETION = 1,
+ REPLAY_REF = 2,
+};
+
+/**
+ * struct replay_entry - replay tree entry.
+ * @lnum: logical eraseblock number of the node
+ * @offs: node offset
+ * @len: node length
+ * @sqnum: node sequence number
+ * @flags: replay flags
+ * @rb: links the replay tree
+ * @key: node key
+ * @nm: directory entry name
+ * @old_size: truncation old size
+ * @new_size: truncation new size
+ * @free: amount of free space in a bud
+ * @dirty: amount of dirty space in a bud from padding and deletion nodes
+ *
+ * UBIFS journal replay must compare node sequence numbers, which means it must
+ * build a tree of node information to insert into the TNC.
+ */
+struct replay_entry {
+ int lnum;
+ int offs;
+ int len;
+ unsigned long long sqnum;
+ int flags;
+ struct rb_node rb;
+ union ubifs_key key;
+ union {
+ struct qstr nm;
+ struct {
+ loff_t old_size;
+ loff_t new_size;
+ };
+ struct {
+ int free;
+ int dirty;
+ };
+ };
+};
+
+/**
+ * struct bud_entry - entry in the list of buds to replay.
+ * @list: next bud in the list
+ * @bud: bud description object
+ * @free: free bytes in the bud
+ * @sqnum: reference node sequence number
+ */
+struct bud_entry {
+ struct list_head list;
+ struct ubifs_bud *bud;
+ int free;
+ unsigned long long sqnum;
+};
+
+/**
+ * set_bud_lprops - set free and dirty space used by a bud.
+ * @c: UBIFS file-system description object
+ * @r: replay entry of bud
+ */
+static int set_bud_lprops(struct ubifs_info *c, struct replay_entry *r)
+{
+ const struct ubifs_lprops *lp;
+ int err = 0, dirty;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, r->lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ dirty = lp->dirty;
+ if (r->offs == 0 && (lp->free != c->leb_size || lp->dirty != 0)) {
+ /*
+ * The LEB was added to the journal with a starting offset of
+ * zero which means the LEB must have been empty. The LEB
+ * property values should be lp->free == c->leb_size and
+ * lp->dirty == 0, but that is not the case. The reason is that
+ * the LEB was garbage collected. The garbage collector resets
+ * the free and dirty space without recording it anywhere except
+ * lprops, so if there is not a commit then lprops does not have
+ * that information next time the file system is mounted.
+ *
+ * We do not need to adjust free space because the scan has told
+ * us the exact value which is recorded in the replay entry as
+ * r->free.
+ *
+ * However we do need to subtract from the dirty space the
+ * amount of space that the garbage collector reclaimed, which
+ * is the whole LEB minus the amount of space that was free.
+ */
+ dbg_mnt("bud LEB %d was GC'd (%d free, %d dirty)", r->lnum,
+ lp->free, lp->dirty);
+ dbg_gc("bud LEB %d was GC'd (%d free, %d dirty)", r->lnum,
+ lp->free, lp->dirty);
+ dirty -= c->leb_size - lp->free;
+ /*
+ * If the replay order was perfect the dirty space would now be
+ * zero. The order is not perfect because the the journal heads
+ * race with each other. This is not a problem but is does mean
+ * that the dirty space may temporarily exceed c->leb_size
+ * during the replay.
+ */
+ if (dirty != 0)
+ dbg_msg("LEB %d lp: %d free %d dirty "
+ "replay: %d free %d dirty", r->lnum, lp->free,
+ lp->dirty, r->free, r->dirty);
+ }
+ lp = ubifs_change_lp(c, lp, r->free, dirty + r->dirty,
+ lp->flags | LPROPS_TAKEN, 0);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * trun_remove_range - apply a replay entry for a truncation to the TNC.
+ * @c: UBIFS file-system description object
+ * @r: replay entry of truncation
+ */
+static int trun_remove_range(struct ubifs_info *c, struct replay_entry *r)
+{
+ unsigned min_blk, max_blk;
+ union ubifs_key min_key, max_key;
+ ino_t ino;
+
+ min_blk = r->new_size / UBIFS_BLOCK_SIZE;
+ if (r->new_size & (UBIFS_BLOCK_SIZE - 1))
+ min_blk += 1;
+
+ max_blk = r->old_size / UBIFS_BLOCK_SIZE;
+ if ((r->old_size & (UBIFS_BLOCK_SIZE - 1)) == 0)
+ max_blk -= 1;
+
+ ino = key_inum(c, &r->key);
+
+ data_key_init(c, &min_key, ino, min_blk);
+ data_key_init(c, &max_key, ino, max_blk);
+
+ return ubifs_tnc_remove_range(c, &min_key, &max_key);
+}
+
+/**
+ * apply_replay_entry - apply a replay entry to the TNC.
+ * @c: UBIFS file-system description object
+ * @r: replay entry to apply
+ *
+ * Apply a replay entry to the TNC.
+ */
+static int apply_replay_entry(struct ubifs_info *c, struct replay_entry *r)
+{
+ int err, deletion = ((r->flags & REPLAY_DELETION) != 0);
+
+ dbg_mnt("LEB %d:%d len %d flgs %d sqnum %llu %s", r->lnum,
+ r->offs, r->len, r->flags, r->sqnum, DBGKEY(&r->key));
+
+ /* Set c->replay_sqnum to help deal with dangling branches. */
+ c->replay_sqnum = r->sqnum;
+
+ if (r->flags & REPLAY_REF)
+ err = set_bud_lprops(c, r);
+ else if (is_hash_key(c, &r->key)) {
+ if (deletion)
+ err = ubifs_tnc_remove_nm(c, &r->key, &r->nm);
+ else
+ err = ubifs_tnc_add_nm(c, &r->key, r->lnum, r->offs,
+ r->len, &r->nm);
+ } else {
+ if (deletion)
+ switch (key_type(c, &r->key)) {
+ case UBIFS_INO_KEY:
+ {
+ ino_t inum = key_inum(c, &r->key);
+
+ err = ubifs_tnc_remove_ino(c, inum);
+ break;
+ }
+ case UBIFS_TRUN_KEY:
+ err = trun_remove_range(c, r);
+ break;
+ default:
+ err = ubifs_tnc_remove(c, &r->key);
+ break;
+ }
+ else
+ err = ubifs_tnc_add(c, &r->key, r->lnum, r->offs,
+ r->len);
+ if (err)
+ return err;
+
+ if (c->need_recovery)
+ err = ubifs_recover_size_accum(c, &r->key, deletion,
+ r->new_size);
+ }
+
+ return err;
+}
+
+/**
+ * destroy_replay_tree - destroy the replay.
+ * @c: UBIFS file-system description object
+ *
+ * Destroy the replay tree.
+ */
+static void destroy_replay_tree(struct ubifs_info *c)
+{
+ struct rb_node *this = c->replay_tree.rb_node;
+ struct replay_entry *r;
+
+ while (this) {
+ if (this->rb_left) {
+ this = this->rb_left;
+ continue;
+ } else if (this->rb_right) {
+ this = this->rb_right;
+ continue;
+ }
+ r = rb_entry(this, struct replay_entry, rb);
+ this = rb_parent(this);
+ if (this) {
+ if (this->rb_left == &r->rb)
+ this->rb_left = NULL;
+ else
+ this->rb_right = NULL;
+ }
+ if (is_hash_key(c, &r->key))
+ kfree((void *)r->nm.name);
+ kfree(r);
+ }
+ c->replay_tree = RB_ROOT;
+}
+
+/**
+ * apply_replay_tree - apply the replay tree to the TNC.
+ * @c: UBIFS file-system description object
+ *
+ * Apply the replay tree.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int apply_replay_tree(struct ubifs_info *c)
+{
+ struct rb_node *this = rb_first(&c->replay_tree);
+
+ while (this) {
+ struct replay_entry *r;
+ int err;
+
+ cond_resched();
+
+ r = rb_entry(this, struct replay_entry, rb);
+ err = apply_replay_entry(c, r);
+ if (err)
+ return err;
+ this = rb_next(this);
+ }
+ return 0;
+}
+
+/**
+ * insert_node - insert a node to the replay tree.
+ * @c: UBIFS file-system description object
+ * @lnum: node logical eraseblock number
+ * @offs: node offset
+ * @len: node length
+ * @key: node key
+ * @sqnum: sequence number
+ * @deletion: non-zero if this is a deletion
+ * @used: number of bytes in use in a LEB
+ * @old_size: truncation old size
+ * @new_size: truncation new size
+ *
+ * This function inserts a scanned non-direntry node to the replay tree. The
+ * replay tree is an RB-tree containing @struct replay_entry elements which are
+ * indexed by the sequence number. The replay tree is applied at the very end
+ * of the replay process. Since the tree is sorted in sequence number order,
+ * the older modifications are applied first. This function returns zero in
+ * case of success and a negative error code in case of failure.
+ */
+static int insert_node(struct ubifs_info *c, int lnum, int offs, int len,
+ union ubifs_key *key, unsigned long long sqnum,
+ int deletion, int *used, loff_t old_size,
+ loff_t new_size)
+{
+ struct rb_node **p = &c->replay_tree.rb_node, *parent = NULL;
+ struct replay_entry *r;
+
+ if (key_inum(c, key) >= c->highest_inum)
+ c->highest_inum = key_inum(c, key);
+
+ dbg_mnt("add LEB %d:%d, key %s", lnum, offs, DBGKEY(key));
+ while (*p) {
+ parent = *p;
+ r = rb_entry(parent, struct replay_entry, rb);
+ if (sqnum < r->sqnum) {
+ p = &(*p)->rb_left;
+ continue;
+ } else if (sqnum > r->sqnum) {
+ p = &(*p)->rb_right;
+ continue;
+ }
+ ubifs_err("duplicate sqnum in replay");
+ return -EINVAL;
+ }
+
+ r = kzalloc(sizeof(struct replay_entry), GFP_KERNEL);
+ if (!r)
+ return -ENOMEM;
+
+ if (!deletion)
+ *used += ALIGN(len, 8);
+ r->lnum = lnum;
+ r->offs = offs;
+ r->len = len;
+ r->sqnum = sqnum;
+ r->flags = (deletion ? REPLAY_DELETION : 0);
+ r->old_size = old_size;
+ r->new_size = new_size;
+ key_copy(c, key, &r->key);
+
+ rb_link_node(&r->rb, parent, p);
+ rb_insert_color(&r->rb, &c->replay_tree);
+ return 0;
+}
+
+/**
+ * insert_dent - insert a directory entry node into the replay tree.
+ * @c: UBIFS file-system description object
+ * @lnum: node logical eraseblock number
+ * @offs: node offset
+ * @len: node length
+ * @key: node key
+ * @name: directory entry name
+ * @nlen: directory entry name length
+ * @sqnum: sequence number
+ * @deletion: non-zero if this is a deletion
+ * @used: number of bytes in use in a LEB
+ *
+ * This function inserts a scanned directory entry node to the replay tree.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ *
+ * This function is also used for extended attribute entries because they are
+ * implemented as directory entry nodes.
+ */
+static int insert_dent(struct ubifs_info *c, int lnum, int offs, int len,
+ union ubifs_key *key, const char *name, int nlen,
+ unsigned long long sqnum, int deletion, int *used)
+{
+ struct rb_node **p = &c->replay_tree.rb_node, *parent = NULL;
+ struct replay_entry *r;
+ char *nbuf;
+
+ if (key_inum(c, key) >= c->highest_inum)
+ c->highest_inum = key_inum(c, key);
+
+ dbg_mnt("add LEB %d:%d, key %s", lnum, offs, DBGKEY(key));
+ while (*p) {
+ parent = *p;
+ r = rb_entry(parent, struct replay_entry, rb);
+ if (sqnum < r->sqnum) {
+ p = &(*p)->rb_left;
+ continue;
+ }
+ if (sqnum > r->sqnum) {
+ p = &(*p)->rb_right;
+ continue;
+ }
+ ubifs_err("duplicate sqnum in replay");
+ return -EINVAL;
+ }
+
+ r = kzalloc(sizeof(struct replay_entry), GFP_KERNEL);
+ if (!r)
+ return -ENOMEM;
+ nbuf = kmalloc(nlen + 1, GFP_KERNEL);
+ if (!nbuf) {
+ kfree(r);
+ return -ENOMEM;
+ }
+
+ if (!deletion)
+ *used += ALIGN(len, 8);
+ r->lnum = lnum;
+ r->offs = offs;
+ r->len = len;
+ r->sqnum = sqnum;
+ r->nm.len = nlen;
+ memcpy(nbuf, name, nlen);
+ nbuf[nlen] = '\0';
+ r->nm.name = nbuf;
+ r->flags = (deletion ? REPLAY_DELETION : 0);
+ key_copy(c, key, &r->key);
+
+ ubifs_assert(!*p);
+ rb_link_node(&r->rb, parent, p);
+ rb_insert_color(&r->rb, &c->replay_tree);
+ return 0;
+}
+
+/**
+ * ubifs_validate_entry - validate directory or extended attribute entry node.
+ * @c: UBIFS file-system description object
+ * @dent: the node to validate
+ *
+ * This function validates directory or extended attribute entry node @dent.
+ * Returns zero if the node is all right and a %-EINVAL if not.
+ */
+int ubifs_validate_entry(struct ubifs_info *c,
+ const struct ubifs_dent_node *dent)
+{
+ int key_type = key_type_flash(c, dent->key);
+ int nlen = le16_to_cpu(dent->nlen);
+
+ if (le32_to_cpu(dent->ch.len) != nlen + UBIFS_DENT_NODE_SZ + 1 ||
+ dent->type >= UBIFS_ITYPES_CNT ||
+ nlen > UBIFS_MAX_NLEN || dent->name[nlen] != 0 ||
+ strnlen((char *)dent->name, nlen) != nlen ||
+ le64_to_cpu(dent->inum) > MAX_INUM) {
+ ubifs_err("bad %s node", key_type == UBIFS_DENT_KEY ?
+ "directory entry" : "extended attribute entry");
+ return -EINVAL;
+ }
+
+ if (key_type != UBIFS_DENT_KEY && key_type != UBIFS_XENT_KEY) {
+ ubifs_err("bad key type %d", key_type);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/**
+ * replay_bud - replay a bud logical eraseblock.
+ * @c: UBIFS file-system description object
+ * @lnum: bud logical eraseblock number to replay
+ * @offs: bud start offset
+ * @jhead: journal head to which this bud belongs
+ * @free: amount of free space in the bud is returned here
+ * @dirty: amount of dirty space from padding and deletion nodes is returned
+ * here
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+static int replay_bud(struct ubifs_info *c, int lnum, int offs, int jhead,
+ int *free, int *dirty)
+{
+ int err = 0, used = 0;
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ struct ubifs_bud *bud;
+
+ dbg_mnt("replay bud LEB %d, head %d", lnum, jhead);
+ if (c->need_recovery)
+ sleb = ubifs_recover_leb(c, lnum, offs, c->sbuf, jhead != GCHD);
+ else
+ sleb = ubifs_scan(c, lnum, offs, c->sbuf);
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+
+ /*
+ * The bud does not have to start from offset zero - the beginning of
+ * the 'lnum' LEB may contain previously committed data. One of the
+ * things we have to do in replay is to correctly update lprops with
+ * newer information about this LEB.
+ *
+ * At this point lprops thinks that this LEB has 'c->leb_size - offs'
+ * bytes of free space because it only contain information about
+ * committed data.
+ *
+ * But we know that real amount of free space is 'c->leb_size -
+ * sleb->endpt', and the space in the 'lnum' LEB between 'offs' and
+ * 'sleb->endpt' is used by bud data. We have to correctly calculate
+ * how much of these data are dirty and update lprops with this
+ * information.
+ *
+ * The dirt in that LEB region is comprised of padding nodes, deletion
+ * nodes, truncation nodes and nodes which are obsoleted by subsequent
+ * nodes in this LEB. So instead of calculating clean space, we
+ * calculate used space ('used' variable).
+ */
+
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ int deletion = 0;
+
+ cond_resched();
+
+ if (snod->sqnum >= SQNUM_WATERMARK) {
+ ubifs_err("file system's life ended");
+ goto out_dump;
+ }
+
+ if (snod->sqnum > c->max_sqnum)
+ c->max_sqnum = snod->sqnum;
+
+ switch (snod->type) {
+ case UBIFS_INO_NODE:
+ {
+ struct ubifs_ino_node *ino = snod->node;
+ loff_t new_size = le64_to_cpu(ino->size);
+
+ if (le32_to_cpu(ino->nlink) == 0)
+ deletion = 1;
+ err = insert_node(c, lnum, snod->offs, snod->len,
+ &snod->key, snod->sqnum, deletion,
+ &used, 0, new_size);
+ break;
+ }
+ case UBIFS_DATA_NODE:
+ {
+ struct ubifs_data_node *dn = snod->node;
+ loff_t new_size = le32_to_cpu(dn->size) +
+ key_block(c, &snod->key) *
+ UBIFS_BLOCK_SIZE;
+
+ err = insert_node(c, lnum, snod->offs, snod->len,
+ &snod->key, snod->sqnum, deletion,
+ &used, 0, new_size);
+ break;
+ }
+ case UBIFS_DENT_NODE:
+ case UBIFS_XENT_NODE:
+ {
+ struct ubifs_dent_node *dent = snod->node;
+
+ err = ubifs_validate_entry(c, dent);
+ if (err)
+ goto out_dump;
+
+ err = insert_dent(c, lnum, snod->offs, snod->len,
+ &snod->key, (char *)dent->name,
+ le16_to_cpu(dent->nlen), snod->sqnum,
+ !le64_to_cpu(dent->inum), &used);
+ break;
+ }
+ case UBIFS_TRUN_NODE:
+ {
+ struct ubifs_trun_node *trun = snod->node;
+ loff_t old_size = le64_to_cpu(trun->old_size);
+ loff_t new_size = le64_to_cpu(trun->new_size);
+ union ubifs_key key;
+
+ /* Validate truncation node */
+ if (old_size < 0 || old_size > c->max_inode_sz ||
+ new_size < 0 || new_size > c->max_inode_sz ||
+ old_size <= new_size) {
+ ubifs_err("bad truncation node");
+ goto out_dump;
+ }
+
+ /*
+ * Create a fake truncation key just to use the same
+ * functions which expect nodes to have keys.
+ */
+ trun_key_init(c, &key, le32_to_cpu(trun->inum));
+ err = insert_node(c, lnum, snod->offs, snod->len,
+ &key, snod->sqnum, 1, &used,
+ old_size, new_size);
+ break;
+ }
+ default:
+ ubifs_err("unexpected node type %d in bud LEB %d:%d",
+ snod->type, lnum, snod->offs);
+ err = -EINVAL;
+ goto out_dump;
+ }
+ if (err)
+ goto out;
+ }
+
+ bud = ubifs_search_bud(c, lnum);
+ if (!bud)
+ BUG();
+
+ ubifs_assert(sleb->endpt - offs >= used);
+ ubifs_assert(sleb->endpt % c->min_io_size == 0);
+
+ *dirty = sleb->endpt - offs - used;
+ *free = c->leb_size - sleb->endpt;
+
+out:
+ ubifs_scan_destroy(sleb);
+ return err;
+
+out_dump:
+ ubifs_err("bad node is at LEB %d:%d", lnum, snod->offs);
+ dbg_dump_node(c, snod->node);
+ ubifs_scan_destroy(sleb);
+ return -EINVAL;
+}
+
+/**
+ * insert_ref_node - insert a reference node to the replay tree.
+ * @c: UBIFS file-system description object
+ * @lnum: node logical eraseblock number
+ * @offs: node offset
+ * @sqnum: sequence number
+ * @free: amount of free space in bud
+ * @dirty: amount of dirty space from padding and deletion nodes
+ *
+ * This function inserts a reference node to the replay tree and returns zero
+ * in case of success or a negative error code in case of failure.
+ */
+static int insert_ref_node(struct ubifs_info *c, int lnum, int offs,
+ unsigned long long sqnum, int free, int dirty)
+{
+ struct rb_node **p = &c->replay_tree.rb_node, *parent = NULL;
+ struct replay_entry *r;
+
+ dbg_mnt("add ref LEB %d:%d", lnum, offs);
+ while (*p) {
+ parent = *p;
+ r = rb_entry(parent, struct replay_entry, rb);
+ if (sqnum < r->sqnum) {
+ p = &(*p)->rb_left;
+ continue;
+ } else if (sqnum > r->sqnum) {
+ p = &(*p)->rb_right;
+ continue;
+ }
+ ubifs_err("duplicate sqnum in replay tree");
+ return -EINVAL;
+ }
+
+ r = kzalloc(sizeof(struct replay_entry), GFP_KERNEL);
+ if (!r)
+ return -ENOMEM;
+
+ r->lnum = lnum;
+ r->offs = offs;
+ r->sqnum = sqnum;
+ r->flags = REPLAY_REF;
+ r->free = free;
+ r->dirty = dirty;
+
+ rb_link_node(&r->rb, parent, p);
+ rb_insert_color(&r->rb, &c->replay_tree);
+ return 0;
+}
+
+/**
+ * replay_buds - replay all buds.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+static int replay_buds(struct ubifs_info *c)
+{
+ struct bud_entry *b;
+ int err, uninitialized_var(free), uninitialized_var(dirty);
+
+ list_for_each_entry(b, &c->replay_buds, list) {
+ err = replay_bud(c, b->bud->lnum, b->bud->start, b->bud->jhead,
+ &free, &dirty);
+ if (err)
+ return err;
+ err = insert_ref_node(c, b->bud->lnum, b->bud->start, b->sqnum,
+ free, dirty);
+ if (err)
+ return err;
+ }
+
+ return 0;
+}
+
+/**
+ * destroy_bud_list - destroy the list of buds to replay.
+ * @c: UBIFS file-system description object
+ */
+static void destroy_bud_list(struct ubifs_info *c)
+{
+ struct bud_entry *b;
+
+ while (!list_empty(&c->replay_buds)) {
+ b = list_entry(c->replay_buds.next, struct bud_entry, list);
+ list_del(&b->list);
+ kfree(b);
+ }
+}
+
+/**
+ * add_replay_bud - add a bud to the list of buds to replay.
+ * @c: UBIFS file-system description object
+ * @lnum: bud logical eraseblock number to replay
+ * @offs: bud start offset
+ * @jhead: journal head to which this bud belongs
+ * @sqnum: reference node sequence number
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+static int add_replay_bud(struct ubifs_info *c, int lnum, int offs, int jhead,
+ unsigned long long sqnum)
+{
+ struct ubifs_bud *bud;
+ struct bud_entry *b;
+
+ dbg_mnt("add replay bud LEB %d:%d, head %d", lnum, offs, jhead);
+
+ bud = kmalloc(sizeof(struct ubifs_bud), GFP_KERNEL);
+ if (!bud)
+ return -ENOMEM;
+
+ b = kmalloc(sizeof(struct bud_entry), GFP_KERNEL);
+ if (!b) {
+ kfree(bud);
+ return -ENOMEM;
+ }
+
+ bud->lnum = lnum;
+ bud->start = offs;
+ bud->jhead = jhead;
+ ubifs_add_bud(c, bud);
+
+ b->bud = bud;
+ b->sqnum = sqnum;
+ list_add_tail(&b->list, &c->replay_buds);
+
+ return 0;
+}
+
+/**
+ * validate_ref - validate a reference node.
+ * @c: UBIFS file-system description object
+ * @ref: the reference node to validate
+ * @ref_lnum: LEB number of the reference node
+ * @ref_offs: reference node offset
+ *
+ * This function returns %1 if a bud reference already exists for the LEB. %0 is
+ * returned if the reference node is new, otherwise %-EINVAL is returned if
+ * validation failed.
+ */
+static int validate_ref(struct ubifs_info *c, const struct ubifs_ref_node *ref)
+{
+ struct ubifs_bud *bud;
+ int lnum = le32_to_cpu(ref->lnum);
+ unsigned int offs = le32_to_cpu(ref->offs);
+ unsigned int jhead = le32_to_cpu(ref->jhead);
+
+ /*
+ * ref->offs may point to the end of LEB when the journal head points
+ * to the end of LEB and we write reference node for it during commit.
+ * So this is why we require 'offs > c->leb_size'.
+ */
+ if (jhead >= c->jhead_cnt || lnum >= c->leb_cnt ||
+ lnum < c->main_first || offs > c->leb_size ||
+ offs & (c->min_io_size - 1))
+ return -EINVAL;
+
+ /* Make sure we have not already looked at this bud */
+ bud = ubifs_search_bud(c, lnum);
+ if (bud) {
+ if (bud->jhead == jhead && bud->start <= offs)
+ return 1;
+ ubifs_err("bud at LEB %d:%d was already referred", lnum, offs);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/**
+ * replay_log_leb - replay a log logical eraseblock.
+ * @c: UBIFS file-system description object
+ * @lnum: log logical eraseblock to replay
+ * @offs: offset to start replaying from
+ * @sbuf: scan buffer
+ *
+ * This function replays a log LEB and returns zero in case of success, %1 if
+ * this is the last LEB in the log, and a negative error code in case of
+ * failure.
+ */
+static int replay_log_leb(struct ubifs_info *c, int lnum, int offs, void *sbuf)
+{
+ int err;
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ const struct ubifs_cs_node *node;
+
+ dbg_mnt("replay log LEB %d:%d", lnum, offs);
+ sleb = ubifs_scan(c, lnum, offs, sbuf);
+ if (IS_ERR(sleb)) {
+ if (c->need_recovery)
+ sleb = ubifs_recover_log_leb(c, lnum, offs, sbuf);
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+ }
+
+ if (sleb->nodes_cnt == 0) {
+ err = 1;
+ goto out;
+ }
+
+ node = sleb->buf;
+
+ snod = list_entry(sleb->nodes.next, struct ubifs_scan_node, list);
+ if (c->cs_sqnum == 0) {
+ /*
+ * This is the first log LEB we are looking at, make sure that
+ * the first node is a commit start node. Also record its
+ * sequence number so that UBIFS can determine where the log
+ * ends, because all nodes which were have higher sequence
+ * numbers.
+ */
+ if (snod->type != UBIFS_CS_NODE) {
+ dbg_err("first log node at LEB %d:%d is not CS node",
+ lnum, offs);
+ goto out_dump;
+ }
+ if (le64_to_cpu(node->cmt_no) != c->cmt_no) {
+ dbg_err("first CS node at LEB %d:%d has wrong "
+ "commit number %llu expected %llu",
+ lnum, offs,
+ (unsigned long long)le64_to_cpu(node->cmt_no),
+ c->cmt_no);
+ goto out_dump;
+ }
+
+ c->cs_sqnum = le64_to_cpu(node->ch.sqnum);
+ dbg_mnt("commit start sqnum %llu", c->cs_sqnum);
+ }
+
+ if (snod->sqnum < c->cs_sqnum) {
+ /*
+ * This means that we reached end of log and now
+ * look to the older log data, which was already
+ * committed but the eraseblock was not erased (UBIFS
+ * only un-maps it). So this basically means we have to
+ * exit with "end of log" code.
+ */
+ err = 1;
+ goto out;
+ }
+
+ /* Make sure the first node sits at offset zero of the LEB */
+ if (snod->offs != 0) {
+ dbg_err("first node is not at zero offset");
+ goto out_dump;
+ }
+
+ list_for_each_entry(snod, &sleb->nodes, list) {
+
+ cond_resched();
+
+ if (snod->sqnum >= SQNUM_WATERMARK) {
+ ubifs_err("file system's life ended");
+ goto out_dump;
+ }
+
+ if (snod->sqnum < c->cs_sqnum) {
+ dbg_err("bad sqnum %llu, commit sqnum %llu",
+ snod->sqnum, c->cs_sqnum);
+ goto out_dump;
+ }
+
+ if (snod->sqnum > c->max_sqnum)
+ c->max_sqnum = snod->sqnum;
+
+ switch (snod->type) {
+ case UBIFS_REF_NODE: {
+ const struct ubifs_ref_node *ref = snod->node;
+
+ err = validate_ref(c, ref);
+ if (err == 1)
+ break; /* Already have this bud */
+ if (err)
+ goto out_dump;
+
+ err = add_replay_bud(c, le32_to_cpu(ref->lnum),
+ le32_to_cpu(ref->offs),
+ le32_to_cpu(ref->jhead),
+ snod->sqnum);
+ if (err)
+ goto out;
+
+ break;
+ }
+ case UBIFS_CS_NODE:
+ /* Make sure it sits at the beginning of LEB */
+ if (snod->offs != 0) {
+ ubifs_err("unexpected node in log");
+ goto out_dump;
+ }
+ break;
+ default:
+ ubifs_err("unexpected node in log");
+ goto out_dump;
+ }
+ }
+
+ if (sleb->endpt || c->lhead_offs >= c->leb_size) {
+ c->lhead_lnum = lnum;
+ c->lhead_offs = sleb->endpt;
+ }
+
+ err = !sleb->endpt;
+out:
+ ubifs_scan_destroy(sleb);
+ return err;
+
+out_dump:
+ ubifs_err("log error detected while replying the log at LEB %d:%d",
+ lnum, offs + snod->offs);
+ dbg_dump_node(c, snod->node);
+ ubifs_scan_destroy(sleb);
+ return -EINVAL;
+}
+
+/**
+ * take_ihead - update the status of the index head in lprops to 'taken'.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns the amount of free space in the index head LEB or a
+ * negative error code.
+ */
+static int take_ihead(struct ubifs_info *c)
+{
+ const struct ubifs_lprops *lp;
+ int err, free;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, c->ihead_lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ free = lp->free;
+
+ lp = ubifs_change_lp(c, lp, LPROPS_NC, LPROPS_NC,
+ lp->flags | LPROPS_TAKEN, 0);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ err = free;
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_replay_journal - replay journal.
+ * @c: UBIFS file-system description object
+ *
+ * This function scans the journal, replays and cleans it up. It makes sure all
+ * memory data structures related to uncommitted journal are built (dirty TNC
+ * tree, tree of buds, modified lprops, etc).
+ */
+int ubifs_replay_journal(struct ubifs_info *c)
+{
+ int err, i, lnum, offs, _free;
+ void *sbuf = NULL;
+
+ BUILD_BUG_ON(UBIFS_TRUN_KEY > 5);
+
+ /* Update the status of the index head in lprops to 'taken' */
+ _free = take_ihead(c);
+ if (_free < 0)
+ return _free; /* Error code */
+
+ if (c->ihead_offs != c->leb_size - _free) {
+ ubifs_err("bad index head LEB %d:%d", c->ihead_lnum,
+ c->ihead_offs);
+ return -EINVAL;
+ }
+
+ sbuf = vmalloc(c->leb_size);
+ if (!sbuf)
+ return -ENOMEM;
+
+ dbg_mnt("start replaying the journal");
+
+ c->replaying = 1;
+
+ lnum = c->ltail_lnum = c->lhead_lnum;
+ offs = c->lhead_offs;
+
+ for (i = 0; i < c->log_lebs; i++, lnum++) {
+ if (lnum >= UBIFS_LOG_LNUM + c->log_lebs) {
+ /*
+ * The log is logically circular, we reached the last
+ * LEB, switch to the first one.
+ */
+ lnum = UBIFS_LOG_LNUM;
+ offs = 0;
+ }
+ err = replay_log_leb(c, lnum, offs, sbuf);
+ if (err == 1)
+ /* We hit the end of the log */
+ break;
+ if (err)
+ goto out;
+ offs = 0;
+ }
+
+ err = replay_buds(c);
+ if (err)
+ goto out;
+
+ err = apply_replay_tree(c);
+ if (err)
+ goto out;
+
+ ubifs_assert(c->bud_bytes <= c->max_bud_bytes || c->need_recovery);
+ dbg_mnt("finished, log head LEB %d:%d, max_sqnum %llu, "
+ "highest_inum %lu", c->lhead_lnum, c->lhead_offs, c->max_sqnum,
+ (unsigned long)c->highest_inum);
+out:
+ destroy_replay_tree(c);
+ destroy_bud_list(c);
+ vfree(sbuf);
+ c->replaying = 0;
+ return err;
+}
diff --git a/fs/ubifs/sb.c b/fs/ubifs/sb.c
new file mode 100644
index 0000000..9708fda
--- /dev/null
+++ b/fs/ubifs/sb.c
@@ -0,0 +1,324 @@
+/*
+ * 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 superblock. The superblock is stored at the first
+ * LEB of the volume and is never changed by UBIFS. Only user-space tools may
+ * change it. The superblock node mostly contains geometry information.
+ */
+
+#include "ubifs.h"
+
+/*
+ * Default journal size in logical eraseblocks as a percent of total
+ * flash size.
+ */
+#define DEFAULT_JNL_PERCENT 5
+
+/* Default maximum journal size in bytes */
+#define DEFAULT_MAX_JNL (32*1024*1024)
+
+/* Default indexing tree fanout */
+#define DEFAULT_FANOUT 8
+
+/* Default number of data journal heads */
+#define DEFAULT_JHEADS_CNT 1
+
+/* Default positions of different LEBs in the main area */
+#define DEFAULT_IDX_LEB 0
+#define DEFAULT_DATA_LEB 1
+#define DEFAULT_GC_LEB 2
+
+/* Default number of LEB numbers in LPT's save table */
+#define DEFAULT_LSAVE_CNT 256
+
+/* Default reserved pool size as a percent of maximum free space */
+#define DEFAULT_RP_PERCENT 5
+
+/* The default maximum size of reserved pool in bytes */
+#define DEFAULT_MAX_RP_SIZE (5*1024*1024)
+
+/* Default time granularity in nanoseconds */
+#define DEFAULT_TIME_GRAN 1000000000
+
+/**
+ * validate_sb - validate superblock node.
+ * @c: UBIFS file-system description object
+ * @sup: superblock node
+ *
+ * This function validates superblock node @sup. Since most of data was read
+ * from the superblock and stored in @c, the function validates fields in @c
+ * instead. Returns zero in case of success and %-EINVAL in case of validation
+ * failure.
+ */
+static int validate_sb(struct ubifs_info *c, struct ubifs_sb_node *sup)
+{
+ long long max_bytes;
+ int err = 1, min_leb_cnt;
+
+ if (!c->key_hash) {
+ err = 2;
+ goto failed;
+ }
+
+ if (sup->key_fmt != UBIFS_SIMPLE_KEY_FMT) {
+ err = 3;
+ goto failed;
+ }
+
+ if (le32_to_cpu(sup->min_io_size) != c->min_io_size) {
+ ubifs_err("min. I/O unit mismatch: %d in superblock, %d real",
+ le32_to_cpu(sup->min_io_size), c->min_io_size);
+ goto failed;
+ }
+
+ if (le32_to_cpu(sup->leb_size) != c->leb_size) {
+ ubifs_err("LEB size mismatch: %d in superblock, %d real",
+ le32_to_cpu(sup->leb_size), c->leb_size);
+ goto failed;
+ }
+
+ if (c->log_lebs < UBIFS_MIN_LOG_LEBS ||
+ c->lpt_lebs < UBIFS_MIN_LPT_LEBS ||
+ c->orph_lebs < UBIFS_MIN_ORPH_LEBS ||
+ c->main_lebs < UBIFS_MIN_MAIN_LEBS) {
+ err = 4;
+ goto failed;
+ }
+
+ /*
+ * Calculate minimum allowed amount of main area LEBs. This is very
+ * similar to %UBIFS_MIN_LEB_CNT, but we take into account real what we
+ * have just read from the superblock.
+ */
+ min_leb_cnt = UBIFS_SB_LEBS + UBIFS_MST_LEBS + c->log_lebs;
+ min_leb_cnt += c->lpt_lebs + c->orph_lebs + c->jhead_cnt + 6;
+
+ if (c->leb_cnt < min_leb_cnt || c->leb_cnt > c->vi.size) {
+ ubifs_err("bad LEB count: %d in superblock, %d on UBI volume, "
+ "%d minimum required", c->leb_cnt, c->vi.size,
+ min_leb_cnt);
+ goto failed;
+ }
+
+ if (c->max_leb_cnt < c->leb_cnt) {
+ ubifs_err("max. LEB count %d less than LEB count %d",
+ c->max_leb_cnt, c->leb_cnt);
+ goto failed;
+ }
+
+ if (c->main_lebs < UBIFS_MIN_MAIN_LEBS) {
+ err = 7;
+ goto failed;
+ }
+
+ if (c->max_bud_bytes < (long long)c->leb_size * UBIFS_MIN_BUD_LEBS ||
+ c->max_bud_bytes > (long long)c->leb_size * c->main_lebs) {
+ err = 8;
+ goto failed;
+ }
+
+ if (c->jhead_cnt < NONDATA_JHEADS_CNT + 1 ||
+ c->jhead_cnt > NONDATA_JHEADS_CNT + UBIFS_MAX_JHEADS) {
+ err = 9;
+ goto failed;
+ }
+
+ if (c->fanout < UBIFS_MIN_FANOUT ||
+ ubifs_idx_node_sz(c, c->fanout) > c->leb_size) {
+ err = 10;
+ goto failed;
+ }
+
+ if (c->lsave_cnt < 0 || (c->lsave_cnt > DEFAULT_LSAVE_CNT &&
+ c->lsave_cnt > c->max_leb_cnt - UBIFS_SB_LEBS - UBIFS_MST_LEBS -
+ c->log_lebs - c->lpt_lebs - c->orph_lebs)) {
+ err = 11;
+ goto failed;
+ }
+
+ if (UBIFS_SB_LEBS + UBIFS_MST_LEBS + c->log_lebs + c->lpt_lebs +
+ c->orph_lebs + c->main_lebs != c->leb_cnt) {
+ err = 12;
+ goto failed;
+ }
+
+ if (c->default_compr < 0 || c->default_compr >= UBIFS_COMPR_TYPES_CNT) {
+ err = 13;
+ goto failed;
+ }
+
+ max_bytes = c->main_lebs * (long long)c->leb_size;
+ if (c->rp_size < 0 || max_bytes < c->rp_size) {
+ err = 14;
+ goto failed;
+ }
+
+ if (le32_to_cpu(sup->time_gran) > 1000000000 ||
+ le32_to_cpu(sup->time_gran) < 1) {
+ err = 15;
+ goto failed;
+ }
+
+ return 0;
+
+failed:
+ ubifs_err("bad superblock, error %d", err);
+ dbg_dump_node(c, sup);
+ return -EINVAL;
+}
+
+/**
+ * ubifs_read_sb_node - read superblock node.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns a pointer to the superblock node or a negative error
+ * code.
+ */
+struct ubifs_sb_node *ubifs_read_sb_node(struct ubifs_info *c)
+{
+ struct ubifs_sb_node *sup;
+ int err;
+
+ sup = kmalloc(ALIGN(UBIFS_SB_NODE_SZ, c->min_io_size), GFP_NOFS);
+ if (!sup)
+ return ERR_PTR(-ENOMEM);
+
+ err = ubifs_read_node(c, sup, UBIFS_SB_NODE, UBIFS_SB_NODE_SZ,
+ UBIFS_SB_LNUM, 0);
+ if (err) {
+ kfree(sup);
+ return ERR_PTR(err);
+ }
+
+ return sup;
+}
+
+/**
+ * ubifs_read_superblock - read superblock.
+ * @c: UBIFS file-system description object
+ *
+ * This function finds, reads and checks the superblock. If an empty UBI volume
+ * is being mounted, this function creates default superblock. Returns zero in
+ * case of success, and a negative error code in case of failure.
+ */
+int ubifs_read_superblock(struct ubifs_info *c)
+{
+ int err, sup_flags;
+ struct ubifs_sb_node *sup;
+
+ if (c->empty) {
+ printf("No UBIFS filesystem found!\n");
+ return -1;
+ }
+
+ sup = ubifs_read_sb_node(c);
+ if (IS_ERR(sup))
+ return PTR_ERR(sup);
+
+ /*
+ * The software supports all previous versions but not future versions,
+ * due to the unavailability of time-travelling equipment.
+ */
+ c->fmt_version = le32_to_cpu(sup->fmt_version);
+ if (c->fmt_version > UBIFS_FORMAT_VERSION) {
+ ubifs_err("on-flash format version is %d, but software only "
+ "supports up to version %d", c->fmt_version,
+ UBIFS_FORMAT_VERSION);
+ err = -EINVAL;
+ goto out;
+ }
+
+ if (c->fmt_version < 3) {
+ ubifs_err("on-flash format version %d is not supported",
+ c->fmt_version);
+ err = -EINVAL;
+ goto out;
+ }
+
+ switch (sup->key_hash) {
+ case UBIFS_KEY_HASH_R5:
+ c->key_hash = key_r5_hash;
+ c->key_hash_type = UBIFS_KEY_HASH_R5;
+ break;
+
+ case UBIFS_KEY_HASH_TEST:
+ c->key_hash = key_test_hash;
+ c->key_hash_type = UBIFS_KEY_HASH_TEST;
+ break;
+ };
+
+ c->key_fmt = sup->key_fmt;
+
+ switch (c->key_fmt) {
+ case UBIFS_SIMPLE_KEY_FMT:
+ c->key_len = UBIFS_SK_LEN;
+ break;
+ default:
+ ubifs_err("unsupported key format");
+ err = -EINVAL;
+ goto out;
+ }
+
+ c->leb_cnt = le32_to_cpu(sup->leb_cnt);
+ c->max_leb_cnt = le32_to_cpu(sup->max_leb_cnt);
+ c->max_bud_bytes = le64_to_cpu(sup->max_bud_bytes);
+ c->log_lebs = le32_to_cpu(sup->log_lebs);
+ c->lpt_lebs = le32_to_cpu(sup->lpt_lebs);
+ c->orph_lebs = le32_to_cpu(sup->orph_lebs);
+ c->jhead_cnt = le32_to_cpu(sup->jhead_cnt) + NONDATA_JHEADS_CNT;
+ c->fanout = le32_to_cpu(sup->fanout);
+ c->lsave_cnt = le32_to_cpu(sup->lsave_cnt);
+ c->default_compr = le16_to_cpu(sup->default_compr);
+ c->rp_size = le64_to_cpu(sup->rp_size);
+ c->rp_uid = le32_to_cpu(sup->rp_uid);
+ c->rp_gid = le32_to_cpu(sup->rp_gid);
+ sup_flags = le32_to_cpu(sup->flags);
+
+ c->vfs_sb->s_time_gran = le32_to_cpu(sup->time_gran);
+ memcpy(&c->uuid, &sup->uuid, 16);
+ c->big_lpt = !!(sup_flags & UBIFS_FLG_BIGLPT);
+
+ /* Automatically increase file system size to the maximum size */
+ c->old_leb_cnt = c->leb_cnt;
+ if (c->leb_cnt < c->vi.size && c->leb_cnt < c->max_leb_cnt) {
+ c->leb_cnt = min_t(int, c->max_leb_cnt, c->vi.size);
+ dbg_mnt("Auto resizing (ro) from %d LEBs to %d LEBs",
+ c->old_leb_cnt, c->leb_cnt);
+ }
+
+ c->log_bytes = (long long)c->log_lebs * c->leb_size;
+ c->log_last = UBIFS_LOG_LNUM + c->log_lebs - 1;
+ c->lpt_first = UBIFS_LOG_LNUM + c->log_lebs;
+ c->lpt_last = c->lpt_first + c->lpt_lebs - 1;
+ c->orph_first = c->lpt_last + 1;
+ c->orph_last = c->orph_first + c->orph_lebs - 1;
+ c->main_lebs = c->leb_cnt - UBIFS_SB_LEBS - UBIFS_MST_LEBS;
+ c->main_lebs -= c->log_lebs + c->lpt_lebs + c->orph_lebs;
+ c->main_first = c->leb_cnt - c->main_lebs;
+ c->report_rp_size = ubifs_reported_space(c, c->rp_size);
+
+ err = validate_sb(c, sup);
+out:
+ kfree(sup);
+ return err;
+}
diff --git a/fs/ubifs/scan.c b/fs/ubifs/scan.c
new file mode 100644
index 0000000..0ed8247
--- /dev/null
+++ b/fs/ubifs/scan.c
@@ -0,0 +1,362 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements the scan which is a general-purpose function for
+ * determining what nodes are in an eraseblock. The scan is used to replay the
+ * journal, to do garbage collection. for the TNC in-the-gaps method, and by
+ * debugging functions.
+ */
+
+#include "ubifs.h"
+
+/**
+ * scan_padding_bytes - scan for padding bytes.
+ * @buf: buffer to scan
+ * @len: length of buffer
+ *
+ * This function returns the number of padding bytes on success and
+ * %SCANNED_GARBAGE on failure.
+ */
+static int scan_padding_bytes(void *buf, int len)
+{
+ int pad_len = 0, max_pad_len = min_t(int, UBIFS_PAD_NODE_SZ, len);
+ uint8_t *p = buf;
+
+ dbg_scan("not a node");
+
+ while (pad_len < max_pad_len && *p++ == UBIFS_PADDING_BYTE)
+ pad_len += 1;
+
+ if (!pad_len || (pad_len & 7))
+ return SCANNED_GARBAGE;
+
+ dbg_scan("%d padding bytes", pad_len);
+
+ return pad_len;
+}
+
+/**
+ * ubifs_scan_a_node - scan for a node or padding.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to scan
+ * @len: length of buffer
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ * @quiet: print no messages
+ *
+ * This function returns a scanning code to indicate what was scanned.
+ */
+int ubifs_scan_a_node(const struct ubifs_info *c, void *buf, int len, int lnum,
+ int offs, int quiet)
+{
+ struct ubifs_ch *ch = buf;
+ uint32_t magic;
+
+ magic = le32_to_cpu(ch->magic);
+
+ if (magic == 0xFFFFFFFF) {
+ dbg_scan("hit empty space");
+ return SCANNED_EMPTY_SPACE;
+ }
+
+ if (magic != UBIFS_NODE_MAGIC)
+ return scan_padding_bytes(buf, len);
+
+ if (len < UBIFS_CH_SZ)
+ return SCANNED_GARBAGE;
+
+ dbg_scan("scanning %s", dbg_ntype(ch->node_type));
+
+ if (ubifs_check_node(c, buf, lnum, offs, quiet, 1))
+ return SCANNED_A_CORRUPT_NODE;
+
+ if (ch->node_type == UBIFS_PAD_NODE) {
+ struct ubifs_pad_node *pad = buf;
+ int pad_len = le32_to_cpu(pad->pad_len);
+ int node_len = le32_to_cpu(ch->len);
+
+ /* Validate the padding node */
+ if (pad_len < 0 ||
+ offs + node_len + pad_len > c->leb_size) {
+ if (!quiet) {
+ ubifs_err("bad pad node at LEB %d:%d",
+ lnum, offs);
+ dbg_dump_node(c, pad);
+ }
+ return SCANNED_A_BAD_PAD_NODE;
+ }
+
+ /* Make the node pads to 8-byte boundary */
+ if ((node_len + pad_len) & 7) {
+ if (!quiet) {
+ dbg_err("bad padding length %d - %d",
+ offs, offs + node_len + pad_len);
+ }
+ return SCANNED_A_BAD_PAD_NODE;
+ }
+
+ dbg_scan("%d bytes padded, offset now %d",
+ pad_len, ALIGN(offs + node_len + pad_len, 8));
+
+ return node_len + pad_len;
+ }
+
+ return SCANNED_A_NODE;
+}
+
+/**
+ * ubifs_start_scan - create LEB scanning information at start of scan.
+ * @c: UBIFS file-system description object
+ * @lnum: logical eraseblock number
+ * @offs: offset to start at (usually zero)
+ * @sbuf: scan buffer (must be c->leb_size)
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+struct ubifs_scan_leb *ubifs_start_scan(const struct ubifs_info *c, int lnum,
+ int offs, void *sbuf)
+{
+ struct ubifs_scan_leb *sleb;
+ int err;
+
+ dbg_scan("scan LEB %d:%d", lnum, offs);
+
+ sleb = kzalloc(sizeof(struct ubifs_scan_leb), GFP_NOFS);
+ if (!sleb)
+ return ERR_PTR(-ENOMEM);
+
+ sleb->lnum = lnum;
+ INIT_LIST_HEAD(&sleb->nodes);
+ sleb->buf = sbuf;
+
+ err = ubi_read(c->ubi, lnum, sbuf + offs, offs, c->leb_size - offs);
+ if (err && err != -EBADMSG) {
+ ubifs_err("cannot read %d bytes from LEB %d:%d,"
+ " error %d", c->leb_size - offs, lnum, offs, err);
+ kfree(sleb);
+ return ERR_PTR(err);
+ }
+
+ if (err == -EBADMSG)
+ sleb->ecc = 1;
+
+ return sleb;
+}
+
+/**
+ * ubifs_end_scan - update LEB scanning information at end of scan.
+ * @c: UBIFS file-system description object
+ * @sleb: scanning information
+ * @lnum: logical eraseblock number
+ * @offs: offset to start at (usually zero)
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+void ubifs_end_scan(const struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ int lnum, int offs)
+{
+ lnum = lnum;
+ dbg_scan("stop scanning LEB %d at offset %d", lnum, offs);
+ ubifs_assert(offs % c->min_io_size == 0);
+
+ sleb->endpt = ALIGN(offs, c->min_io_size);
+}
+
+/**
+ * ubifs_add_snod - add a scanned node to LEB scanning information.
+ * @c: UBIFS file-system description object
+ * @sleb: scanning information
+ * @buf: buffer containing node
+ * @offs: offset of node on flash
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_add_snod(const struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ void *buf, int offs)
+{
+ struct ubifs_ch *ch = buf;
+ struct ubifs_ino_node *ino = buf;
+ struct ubifs_scan_node *snod;
+
+ snod = kzalloc(sizeof(struct ubifs_scan_node), GFP_NOFS);
+ if (!snod)
+ return -ENOMEM;
+
+ snod->sqnum = le64_to_cpu(ch->sqnum);
+ snod->type = ch->node_type;
+ snod->offs = offs;
+ snod->len = le32_to_cpu(ch->len);
+ snod->node = buf;
+
+ switch (ch->node_type) {
+ case UBIFS_INO_NODE:
+ case UBIFS_DENT_NODE:
+ case UBIFS_XENT_NODE:
+ case UBIFS_DATA_NODE:
+ case UBIFS_TRUN_NODE:
+ /*
+ * The key is in the same place in all keyed
+ * nodes.
+ */
+ key_read(c, &ino->key, &snod->key);
+ break;
+ }
+ list_add_tail(&snod->list, &sleb->nodes);
+ sleb->nodes_cnt += 1;
+ return 0;
+}
+
+/**
+ * ubifs_scanned_corruption - print information after UBIFS scanned corruption.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of corruption
+ * @offs: offset of corruption
+ * @buf: buffer containing corruption
+ */
+void ubifs_scanned_corruption(const struct ubifs_info *c, int lnum, int offs,
+ void *buf)
+{
+ int len;
+
+ ubifs_err("corrupted data at LEB %d:%d", lnum, offs);
+ if (dbg_failure_mode)
+ return;
+ len = c->leb_size - offs;
+ if (len > 4096)
+ len = 4096;
+ dbg_err("first %d bytes from LEB %d:%d", len, lnum, offs);
+ print_hex_dump(KERN_DEBUG, "", DUMP_PREFIX_OFFSET, 32, 4, buf, len, 1);
+}
+
+/**
+ * ubifs_scan - scan a logical eraseblock.
+ * @c: UBIFS file-system description object
+ * @lnum: logical eraseblock number
+ * @offs: offset to start at (usually zero)
+ * @sbuf: scan buffer (must be c->leb_size)
+ *
+ * This function scans LEB number @lnum and returns complete information about
+ * its contents. Returns an error code in case of failure.
+ */
+struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
+ int offs, void *sbuf)
+{
+ void *buf = sbuf + offs;
+ int err, len = c->leb_size - offs;
+ struct ubifs_scan_leb *sleb;
+
+ sleb = ubifs_start_scan(c, lnum, offs, sbuf);
+ if (IS_ERR(sleb))
+ return sleb;
+
+ while (len >= 8) {
+ struct ubifs_ch *ch = buf;
+ int node_len, ret;
+
+ dbg_scan("look at LEB %d:%d (%d bytes left)",
+ lnum, offs, len);
+
+ cond_resched();
+
+ ret = ubifs_scan_a_node(c, buf, len, lnum, offs, 0);
+
+ if (ret > 0) {
+ /* Padding bytes or a valid padding node */
+ offs += ret;
+ buf += ret;
+ len -= ret;
+ continue;
+ }
+
+ if (ret == SCANNED_EMPTY_SPACE)
+ /* Empty space is checked later */
+ break;
+
+ switch (ret) {
+ case SCANNED_GARBAGE:
+ dbg_err("garbage");
+ goto corrupted;
+ case SCANNED_A_NODE:
+ break;
+ case SCANNED_A_CORRUPT_NODE:
+ case SCANNED_A_BAD_PAD_NODE:
+ dbg_err("bad node");
+ goto corrupted;
+ default:
+ dbg_err("unknown");
+ goto corrupted;
+ }
+
+ err = ubifs_add_snod(c, sleb, buf, offs);
+ if (err)
+ goto error;
+
+ node_len = ALIGN(le32_to_cpu(ch->len), 8);
+ offs += node_len;
+ buf += node_len;
+ len -= node_len;
+ }
+
+ if (offs % c->min_io_size)
+ goto corrupted;
+
+ ubifs_end_scan(c, sleb, lnum, offs);
+
+ for (; len > 4; offs += 4, buf = buf + 4, len -= 4)
+ if (*(uint32_t *)buf != 0xffffffff)
+ break;
+ for (; len; offs++, buf++, len--)
+ if (*(uint8_t *)buf != 0xff) {
+ ubifs_err("corrupt empty space at LEB %d:%d",
+ lnum, offs);
+ goto corrupted;
+ }
+
+ return sleb;
+
+corrupted:
+ ubifs_scanned_corruption(c, lnum, offs, buf);
+ err = -EUCLEAN;
+error:
+ ubifs_err("LEB %d scanning failed", lnum);
+ ubifs_scan_destroy(sleb);
+ return ERR_PTR(err);
+}
+
+/**
+ * ubifs_scan_destroy - destroy LEB scanning information.
+ * @sleb: scanning information to free
+ */
+void ubifs_scan_destroy(struct ubifs_scan_leb *sleb)
+{
+ struct ubifs_scan_node *node;
+ struct list_head *head;
+
+ head = &sleb->nodes;
+ while (!list_empty(head)) {
+ node = list_entry(head->next, struct ubifs_scan_node, list);
+ list_del(&node->list);
+ kfree(node);
+ }
+ kfree(sleb);
+}
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
new file mode 100644
index 0000000..95f2a41
--- /dev/null
+++ b/fs/ubifs/super.c
@@ -0,0 +1,1189 @@
+/*
+ * 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 and VFS superblock operations. Some
+ * initialization stuff which is rather large and complex is placed at
+ * corresponding subsystems, but most of it is here.
+ */
+
+#include "ubifs.h"
+#include <linux/math64.h>
+
+#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 int sb_set(struct super_block *sb, 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));
+ if (inode) {
+ inode->i_ino = ino;
+ inode->i_sb = sb;
+ list_add(&inode->i_sb_list, &sb->s_inodes);
+ inode->i_state = I_LOCK | I_NEW;
+ }
+
+ return inode;
+}
+
+int ubifs_iput(struct inode *inode)
+{
+ list_del_init(&inode->i_sb_list);
+
+ free(inode);
+ return 0;
+}
+
+/*
+ * Lock (save) inode in inode array for readback after recovery
+ */
+void iput(struct inode *inode)
+{
+ int i;
+ struct inode *ino;
+
+ /*
+ * Search end of list
+ */
+ for (i = 0; i < INODE_LOCKED_MAX; i++) {
+ if (inodes_locked_down[i] == NULL)
+ break;
+ }
+
+ if (i >= INODE_LOCKED_MAX) {
+ ubifs_err("Error, can't lock (save) more inodes while recovery!!!");
+ return;
+ }
+
+ /*
+ * Allocate and use new inode
+ */
+ ino = (struct inode *)malloc(sizeof(struct ubifs_inode));
+ memcpy(ino, inode, sizeof(struct ubifs_inode));
+
+ /*
+ * Finally save inode in array
+ */
+ inodes_locked_down[i] = ino;
+}
+
+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;
+ int i;
+
+ dbg_gen("inode %lu", inum);
+
+ /*
+ * U-Boot special handling of locked down inodes via recovery
+ * e.g. ubifs_recover_size()
+ */
+ for (i = 0; i < INODE_LOCKED_MAX; i++) {
+ /*
+ * Exit on last entry (NULL), inode not found in list
+ */
+ if (inodes_locked_down[i] == NULL)
+ break;
+
+ if (inodes_locked_down[i]->i_ino == inum) {
+ /*
+ * We found the locked down inode in our array,
+ * so just return this pointer instead of creating
+ * a new one.
+ */
+ return inodes_locked_down[i];
+ }
+ }
+
+ 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->synced_i_size = ui->ui_size = inode->i_size;
+
+ err = validate_inode(c, inode);
+ if (err)
+ goto out_invalid;
+
+ if ((inode->i_mode & S_IFMT) == S_IFLNK) {
+ 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';
+ }
+
+ kfree(ino);
+ inode->i_state &= ~(I_LOCK | I_NEW);
+ 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);
+ 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)
+{
+ 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. See gc.c for comments
+ * about these values.
+ */
+ c->dead_wm = ALIGN(MIN_WRITE_SZ, c->min_io_size);
+ c->dark_wm = ALIGN(UBIFS_MAX_NODE_SZ, c->min_io_size);
+
+ /*
+ * Calculate how many bytes would be wasted at the end of LEB if it was
+ * fully filled with data nodes of maximum size. This is used in
+ * calculations when reporting free space.
+ */
+ c->leb_overhead = c->leb_size % UBIFS_MAX_DATA_NODE_SZ;
+
+ return 0;
+}
+
+/*
+ * init_constants_sb - 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_sb(struct ubifs_info *c)
+{
+ int tmp, err;
+ long long tmp64;
+
+ c->main_bytes = (long long)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 + c->leb_size - 1;
+ c->max_bud_cnt = div_u64(tmp64, c->leb_size);
+ 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 = UBIFS_MAX_DATA_NODE_SZ * UBIFS_BLOCKS_PER_PAGE;
+ 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 = (long long)(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;
+
+ return 0;
+}
+
+/*
+ * init_constants_master - initialize UBIFS constants.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which initializes various UBIFS constants after
+ * the master node has been read. It also checks various UBIFS parameters and
+ * makes sure they are all right.
+ */
+static void init_constants_master(struct ubifs_info *c)
+{
+ long long tmp64;
+
+ 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, the LEB which is reserved for
+ * deletions, minimum LEBs for the index, and assume only one journal
+ * head is available.
+ */
+ tmp64 = c->main_lebs - 1 - 1 - MIN_INDEX_LEBS - c->jhead_cnt + 1;
+ tmp64 *= (long long)c->leb_size - c->leb_overhead;
+ tmp64 = ubifs_reported_space(c, tmp64);
+ c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT;
+}
+
+/**
+ * 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;
+}
+
+/**
+ * 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);
+ long long x;
+ size_t sz;
+
+ 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 && (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");
+ err = -EROFS;
+ goto out_free;
+ }
+
+ if (c->ro_media && !mounted_read_only) {
+ ubifs_err("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
+ * height amount of integers. We assume the height if the TNC tree will
+ * never exceed 64.
+ */
+ err = -ENOMEM;
+ c->bottom_up_buf = kmalloc(BOTTOM_UP_HEIGHT * sizeof(int), GFP_KERNEL);
+ if (!c->bottom_up_buf)
+ goto out_free;
+
+ c->sbuf = vmalloc(c->leb_size);
+ 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;
+
+ err = ubifs_read_superblock(c);
+ if (err)
+ goto out_free;
+
+ /*
+ * 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_compr_name(c->default_compr));
+ goto out_free;
+ }
+
+ dbg_failure_mode_registration(c);
+
+ err = init_constants_sb(c);
+ if (err)
+ goto out_free;
+
+ 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) {
+ err = -ENOMEM;
+ goto out_free;
+ }
+
+ sprintf(c->bgt_name, BGT_NAME_PATTERN, c->vi.ubi_num, c->vi.vol_id);
+
+ err = ubifs_read_master(c);
+ if (err)
+ goto out_master;
+
+ init_constants_master(c);
+
+ if ((c->mst_node->flags & cpu_to_le32(UBIFS_MST_DIRTY)) != 0) {
+ ubifs_msg("recovery needed");
+ c->need_recovery = 1;
+ }
+
+ err = ubifs_lpt_init(c, 1, !mounted_read_only);
+ if (err)
+ goto out_lpt;
+
+ err = dbg_check_idx_size(c, c->old_idx_sz);
+ if (err)
+ goto out_lpt;
+
+ err = ubifs_replay_journal(c);
+ if (err)
+ goto out_journal;
+
+ err = ubifs_mount_orphans(c, c->need_recovery, mounted_read_only);
+ if (err)
+ goto out_orphans;
+
+ if (c->need_recovery) {
+ err = ubifs_recover_size(c);
+ if (err)
+ goto out_orphans;
+ }
+
+ 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");
+ }
+ }
+
+ err = dbg_check_filesystem(c);
+ if (err)
+ goto out_infos;
+
+ c->always_chk_crc = 0;
+
+ 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: %d (latest is %d)",
+ c->fmt_version, UBIFS_FORMAT_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);
+
+ 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:
+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 - 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)
+{
+ dbg_gen("un-mounting UBI device %d, volume %d", c->vi.ubi_num,
+ c->vi.vol_id);
+
+ spin_lock(&ubifs_infos_lock);
+ list_del(&c->infos_list);
+ spin_unlock(&ubifs_infos_lock);
+
+ if (c->bgt)
+ kthread_stop(c->bgt);
+
+ free_orphans(c);
+ ubifs_lpt_free(c, 0);
+
+ 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);
+
+ /* Finally free U-Boot's global copy of superblock */
+ free(ubifs_sb->s_fs_info);
+ free(ubifs_sb);
+}
+
+/**
+ * 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;
+ * 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.
+ */
+static struct ubi_volume_desc *open_ubi(const char *name, int mode)
+{
+ int dev, vol;
+ char *endptr;
+
+ if (name[0] != 'u' || name[1] != 'b' || name[2] != 'i')
+ return ERR_PTR(-EINVAL);
+
+ /* ubi:NAME method */
+ if ((name[3] == ':' || name[3] == '!') && name[4] != '\0')
+ return ubi_open_volume_nm(0, name + 4, mode);
+
+ if (!isdigit(name[3]))
+ return ERR_PTR(-EINVAL);
+
+ dev = simple_strtoul(name + 3, &endptr, 0);
+
+ /* ubiY method */
+ if (*endptr == '\0')
+ return ubi_open_volume(0, dev, mode);
+
+ /* ubiX_Y method */
+ if (*endptr == '_' && isdigit(endptr[1])) {
+ vol = simple_strtoul(endptr + 1, &endptr, 0);
+ if (*endptr != '\0')
+ return ERR_PTR(-EINVAL);
+ return ubi_open_volume(dev, vol, mode);
+ }
+
+ /* ubiX:NAME method */
+ if ((*endptr == ':' || *endptr == '!') && endptr[1] != '\0')
+ return ubi_open_volume_nm(dev, ++endptr, mode);
+
+ return ERR_PTR(-EINVAL);
+}
+
+static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
+{
+ 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;
+
+ 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);
+
+ /* Re-open the UBI device in read-write mode */
+ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READONLY);
+ 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 = key_max_inode_size(c);
+ if (c->max_inode_sz > MAX_LFS_FILESIZE)
+ sb->s_maxbytes = c->max_inode_sz = MAX_LFS_FILESIZE;
+
+ mutex_lock(&c->umount_mutex);
+ err = mount_ubifs(c);
+ if (err) {
+ ubifs_assert(err < 0);
+ goto out_unlock;
+ }
+
+ /* 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 = NULL;
+
+ mutex_unlock(&c->umount_mutex);
+ return 0;
+
+out_umount:
+ ubifs_umount(c);
+out_unlock:
+ mutex_unlock(&c->umount_mutex);
+ ubi_close_volume(c->ubi);
+out_free:
+ kfree(c);
+ return err;
+}
+
+static int sb_test(struct super_block *sb, void *data)
+{
+ dev_t *dev = data;
+
+ return sb->s_dev == *dev;
+}
+
+static int ubifs_get_sb(struct file_system_type *fs_type, int flags,
+ const char *name, void *data, struct vfsmount *mnt)
+{
+ struct ubi_volume_desc *ubi;
+ struct ubi_volume_info vi;
+ struct super_block *sb;
+ int err;
+
+ dbg_gen("name %s, flags %#x", name, flags);
+
+ /*
+ * 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.
+ */
+ 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);
+ }
+ ubi_get_volume_info(ubi, &vi);
+
+ dbg_gen("opened ubi%d_%d", vi.ubi_num, vi.vol_id);
+
+ sb = sget(fs_type, &sb_test, &sb_set, &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");
+ if ((flags ^ sb->s_flags) & MS_RDONLY) {
+ 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;
+ /* We do not support atime */
+ sb->s_flags |= MS_ACTIVE | MS_NOATIME;
+ }
+
+ /* 'fill_super()' opens ubi again so we must close it here */
+ ubi_close_volume(ubi);
+
+ ubifs_sb = sb;
+ return 0;
+
+out_deact:
+ up_write(&sb->s_umount);
+out_close:
+ ubi_close_volume(ubi);
+ return err;
+}
+
+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);
+
+ /* 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);
+
+ /*
+ * We use 2 bit wide bit-fields to store compression type, which should
+ * be amended if more compressors are added. The bit-fields are:
+ * @compr_type in 'struct ubifs_inode', @default_compr in
+ * 'struct ubifs_info' and @compr_type in 'struct ubifs_mount_opts'.
+ */
+ BUILD_BUG_ON(UBIFS_COMPR_TYPES_CNT > 4);
+
+ /*
+ * We require that PAGE_CACHE_SIZE is greater-than-or-equal-to
+ * 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);
+ return -EINVAL;
+ }
+
+ err = -ENOMEM;
+
+ err = ubifs_compressors_init();
+ if (err)
+ goto out_shrinker;
+
+ return 0;
+
+out_shrinker:
+ return err;
+}
+
+/*
+ * ubifsmount...
+ */
+
+static struct file_system_type ubifs_fs_type = {
+ .name = "ubifs",
+ .owner = THIS_MODULE,
+ .get_sb = ubifs_get_sb,
+};
+
+int ubifs_mount(char *vol_name)
+{
+ int flags;
+ char name[80] = "ubi:";
+ void *data;
+ struct vfsmount *mnt;
+ int ret;
+ struct ubifs_info *c;
+
+ /*
+ * First unmount if allready mounted
+ */
+ if (ubifs_sb)
+ ubifs_umount(ubifs_sb->s_fs_info);
+
+ INIT_LIST_HEAD(&ubifs_infos);
+
+ /*
+ * 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);
+ return -1;
+ }
+
+ c = ubifs_sb->s_fs_info;
+ ubi_close_volume(c->ubi);
+
+ return 0;
+}
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
new file mode 100644
index 0000000..ccda938
--- /dev/null
+++ b/fs/ubifs/tnc.c
@@ -0,0 +1,2767 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements TNC (Tree Node Cache) which caches indexing nodes of
+ * the UBIFS B-tree.
+ *
+ * At the moment the locking rules of the TNC tree are quite simple and
+ * straightforward. We just have a mutex and lock it when we traverse the
+ * tree. If a znode is not in memory, we read it from flash while still having
+ * the mutex locked.
+ */
+
+#include "ubifs.h"
+
+/*
+ * Returned codes of 'matches_name()' and 'fallible_matches_name()' functions.
+ * @NAME_LESS: name corresponding to the first argument is less than second
+ * @NAME_MATCHES: names match
+ * @NAME_GREATER: name corresponding to the second argument is greater than
+ * first
+ * @NOT_ON_MEDIA: node referred by zbranch does not exist on the media
+ *
+ * These constants were introduce to improve readability.
+ */
+enum {
+ NAME_LESS = 0,
+ NAME_MATCHES = 1,
+ NAME_GREATER = 2,
+ NOT_ON_MEDIA = 3,
+};
+
+/**
+ * insert_old_idx - record an index node obsoleted since the last commit start.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of obsoleted index node
+ * @offs: offset of obsoleted index node
+ *
+ * Returns %0 on success, and a negative error code on failure.
+ *
+ * For recovery, there must always be a complete intact version of the index on
+ * flash at all times. That is called the "old index". It is the index as at the
+ * time of the last successful commit. Many of the index nodes in the old index
+ * may be dirty, but they must not be erased until the next successful commit
+ * (at which point that index becomes the old index).
+ *
+ * That means that the garbage collection and the in-the-gaps method of
+ * committing must be able to determine if an index node is in the old index.
+ * Most of the old index nodes can be found by looking up the TNC using the
+ * 'lookup_znode()' function. However, some of the old index nodes may have
+ * been deleted from the current index or may have been changed so much that
+ * they cannot be easily found. In those cases, an entry is added to an RB-tree.
+ * That is what this function does. The RB-tree is ordered by LEB number and
+ * offset because they uniquely identify the old index node.
+ */
+static int insert_old_idx(struct ubifs_info *c, int lnum, int offs)
+{
+ struct ubifs_old_idx *old_idx, *o;
+ struct rb_node **p, *parent = NULL;
+
+ old_idx = kmalloc(sizeof(struct ubifs_old_idx), GFP_NOFS);
+ if (unlikely(!old_idx))
+ return -ENOMEM;
+ old_idx->lnum = lnum;
+ old_idx->offs = offs;
+
+ p = &c->old_idx.rb_node;
+ while (*p) {
+ parent = *p;
+ o = rb_entry(parent, struct ubifs_old_idx, rb);
+ if (lnum < o->lnum)
+ p = &(*p)->rb_left;
+ else if (lnum > o->lnum)
+ p = &(*p)->rb_right;
+ else if (offs < o->offs)
+ p = &(*p)->rb_left;
+ else if (offs > o->offs)
+ p = &(*p)->rb_right;
+ else {
+ ubifs_err("old idx added twice!");
+ kfree(old_idx);
+ return 0;
+ }
+ }
+ rb_link_node(&old_idx->rb, parent, p);
+ rb_insert_color(&old_idx->rb, &c->old_idx);
+ return 0;
+}
+
+/**
+ * insert_old_idx_znode - record a znode obsoleted since last commit start.
+ * @c: UBIFS file-system description object
+ * @znode: znode of obsoleted index node
+ *
+ * Returns %0 on success, and a negative error code on failure.
+ */
+int insert_old_idx_znode(struct ubifs_info *c, struct ubifs_znode *znode)
+{
+ if (znode->parent) {
+ struct ubifs_zbranch *zbr;
+
+ zbr = &znode->parent->zbranch[znode->iip];
+ if (zbr->len)
+ return insert_old_idx(c, zbr->lnum, zbr->offs);
+ } else
+ if (c->zroot.len)
+ return insert_old_idx(c, c->zroot.lnum,
+ c->zroot.offs);
+ return 0;
+}
+
+/**
+ * ins_clr_old_idx_znode - record a znode obsoleted since last commit start.
+ * @c: UBIFS file-system description object
+ * @znode: znode of obsoleted index node
+ *
+ * Returns %0 on success, and a negative error code on failure.
+ */
+static int ins_clr_old_idx_znode(struct ubifs_info *c,
+ struct ubifs_znode *znode)
+{
+ int err;
+
+ if (znode->parent) {
+ struct ubifs_zbranch *zbr;
+
+ zbr = &znode->parent->zbranch[znode->iip];
+ if (zbr->len) {
+ err = insert_old_idx(c, zbr->lnum, zbr->offs);
+ if (err)
+ return err;
+ zbr->lnum = 0;
+ zbr->offs = 0;
+ zbr->len = 0;
+ }
+ } else
+ if (c->zroot.len) {
+ err = insert_old_idx(c, c->zroot.lnum, c->zroot.offs);
+ if (err)
+ return err;
+ c->zroot.lnum = 0;
+ c->zroot.offs = 0;
+ c->zroot.len = 0;
+ }
+ return 0;
+}
+
+/**
+ * destroy_old_idx - destroy the old_idx RB-tree.
+ * @c: UBIFS file-system description object
+ *
+ * During start commit, the old_idx RB-tree is used to avoid overwriting index
+ * nodes that were in the index last commit but have since been deleted. This
+ * is necessary for recovery i.e. the old index must be kept intact until the
+ * new index is successfully written. The old-idx RB-tree is used for the
+ * in-the-gaps method of writing index nodes and is destroyed every commit.
+ */
+void destroy_old_idx(struct ubifs_info *c)
+{
+ struct rb_node *this = c->old_idx.rb_node;
+ struct ubifs_old_idx *old_idx;
+
+ while (this) {
+ if (this->rb_left) {
+ this = this->rb_left;
+ continue;
+ } else if (this->rb_right) {
+ this = this->rb_right;
+ continue;
+ }
+ old_idx = rb_entry(this, struct ubifs_old_idx, rb);
+ this = rb_parent(this);
+ if (this) {
+ if (this->rb_left == &old_idx->rb)
+ this->rb_left = NULL;
+ else
+ this->rb_right = NULL;
+ }
+ kfree(old_idx);
+ }
+ c->old_idx = RB_ROOT;
+}
+
+/**
+ * copy_znode - copy a dirty znode.
+ * @c: UBIFS file-system description object
+ * @znode: znode to copy
+ *
+ * A dirty znode being committed may not be changed, so it is copied.
+ */
+static struct ubifs_znode *copy_znode(struct ubifs_info *c,
+ struct ubifs_znode *znode)
+{
+ struct ubifs_znode *zn;
+
+ zn = kmalloc(c->max_znode_sz, GFP_NOFS);
+ if (unlikely(!zn))
+ return ERR_PTR(-ENOMEM);
+
+ memcpy(zn, znode, c->max_znode_sz);
+ zn->cnext = NULL;
+ __set_bit(DIRTY_ZNODE, &zn->flags);
+ __clear_bit(COW_ZNODE, &zn->flags);
+
+ ubifs_assert(!test_bit(OBSOLETE_ZNODE, &znode->flags));
+ __set_bit(OBSOLETE_ZNODE, &znode->flags);
+
+ if (znode->level != 0) {
+ int i;
+ const int n = zn->child_cnt;
+
+ /* The children now have new parent */
+ for (i = 0; i < n; i++) {
+ struct ubifs_zbranch *zbr = &zn->zbranch[i];
+
+ if (zbr->znode)
+ zbr->znode->parent = zn;
+ }
+ }
+
+ atomic_long_inc(&c->dirty_zn_cnt);
+ return zn;
+}
+
+/**
+ * add_idx_dirt - add dirt due to a dirty znode.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of index node
+ * @dirt: size of index node
+ *
+ * This function updates lprops dirty space and the new size of the index.
+ */
+static int add_idx_dirt(struct ubifs_info *c, int lnum, int dirt)
+{
+ c->calc_idx_sz -= ALIGN(dirt, 8);
+ return ubifs_add_dirt(c, lnum, dirt);
+}
+
+/**
+ * dirty_cow_znode - ensure a znode is not being committed.
+ * @c: UBIFS file-system description object
+ * @zbr: branch of znode to check
+ *
+ * Returns dirtied znode on success or negative error code on failure.
+ */
+static struct ubifs_znode *dirty_cow_znode(struct ubifs_info *c,
+ struct ubifs_zbranch *zbr)
+{
+ struct ubifs_znode *znode = zbr->znode;
+ struct ubifs_znode *zn;
+ int err;
+
+ if (!test_bit(COW_ZNODE, &znode->flags)) {
+ /* znode is not being committed */
+ if (!test_and_set_bit(DIRTY_ZNODE, &znode->flags)) {
+ atomic_long_inc(&c->dirty_zn_cnt);
+ atomic_long_dec(&c->clean_zn_cnt);
+ atomic_long_dec(&ubifs_clean_zn_cnt);
+ err = add_idx_dirt(c, zbr->lnum, zbr->len);
+ if (unlikely(err))
+ return ERR_PTR(err);
+ }
+ return znode;
+ }
+
+ zn = copy_znode(c, znode);
+ if (IS_ERR(zn))
+ return zn;
+
+ if (zbr->len) {
+ err = insert_old_idx(c, zbr->lnum, zbr->offs);
+ if (unlikely(err))
+ return ERR_PTR(err);
+ err = add_idx_dirt(c, zbr->lnum, zbr->len);
+ } else
+ err = 0;
+
+ zbr->znode = zn;
+ zbr->lnum = 0;
+ zbr->offs = 0;
+ zbr->len = 0;
+
+ if (unlikely(err))
+ return ERR_PTR(err);
+ return zn;
+}
+
+/**
+ * lnc_add - add a leaf node to the leaf node cache.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch of leaf node
+ * @node: leaf node
+ *
+ * Leaf nodes are non-index nodes directory entry nodes or data nodes. The
+ * purpose of the leaf node cache is to save re-reading the same leaf node over
+ * and over again. Most things are cached by VFS, however the file system must
+ * cache directory entries for readdir and for resolving hash collisions. The
+ * present implementation of the leaf node cache is extremely simple, and
+ * allows for error returns that are not used but that may be needed if a more
+ * complex implementation is created.
+ *
+ * Note, this function does not add the @node object to LNC directly, but
+ * allocates a copy of the object and adds the copy to LNC. The reason for this
+ * is that @node has been allocated outside of the TNC subsystem and will be
+ * used with @c->tnc_mutex unlock upon return from the TNC subsystem. But LNC
+ * may be changed at any time, e.g. freed by the shrinker.
+ */
+static int lnc_add(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ const void *node)
+{
+ int err;
+ void *lnc_node;
+ const struct ubifs_dent_node *dent = node;
+
+ ubifs_assert(!zbr->leaf);
+ ubifs_assert(zbr->len != 0);
+ ubifs_assert(is_hash_key(c, &zbr->key));
+
+ err = ubifs_validate_entry(c, dent);
+ if (err) {
+ dbg_dump_stack();
+ dbg_dump_node(c, dent);
+ return err;
+ }
+
+ lnc_node = kmalloc(zbr->len, GFP_NOFS);
+ if (!lnc_node)
+ /* We don't have to have the cache, so no error */
+ return 0;
+
+ memcpy(lnc_node, node, zbr->len);
+ zbr->leaf = lnc_node;
+ return 0;
+}
+
+ /**
+ * lnc_add_directly - add a leaf node to the leaf-node-cache.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch of leaf node
+ * @node: leaf node
+ *
+ * This function is similar to 'lnc_add()', but it does not create a copy of
+ * @node but inserts @node to TNC directly.
+ */
+static int lnc_add_directly(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ void *node)
+{
+ int err;
+
+ ubifs_assert(!zbr->leaf);
+ ubifs_assert(zbr->len != 0);
+
+ err = ubifs_validate_entry(c, node);
+ if (err) {
+ dbg_dump_stack();
+ dbg_dump_node(c, node);
+ return err;
+ }
+
+ zbr->leaf = node;
+ return 0;
+}
+
+/**
+ * lnc_free - remove a leaf node from the leaf node cache.
+ * @zbr: zbranch of leaf node
+ * @node: leaf node
+ */
+static void lnc_free(struct ubifs_zbranch *zbr)
+{
+ if (!zbr->leaf)
+ return;
+ kfree(zbr->leaf);
+ zbr->leaf = NULL;
+}
+
+/**
+ * tnc_read_node_nm - read a "hashed" leaf node.
+ * @c: UBIFS file-system description object
+ * @zbr: key and position of the node
+ * @node: node is returned here
+ *
+ * This function reads a "hashed" node defined by @zbr from the leaf node cache
+ * (in it is there) or from the hash media, in which case the node is also
+ * added to LNC. Returns zero in case of success or a negative negative error
+ * code in case of failure.
+ */
+static int tnc_read_node_nm(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ void *node)
+{
+ int err;
+
+ ubifs_assert(is_hash_key(c, &zbr->key));
+
+ if (zbr->leaf) {
+ /* Read from the leaf node cache */
+ ubifs_assert(zbr->len != 0);
+ memcpy(node, zbr->leaf, zbr->len);
+ return 0;
+ }
+
+ err = ubifs_tnc_read_node(c, zbr, node);
+ if (err)
+ return err;
+
+ /* Add the node to the leaf node cache */
+ err = lnc_add(c, zbr, node);
+ return err;
+}
+
+/**
+ * try_read_node - read a node if it is a node.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to read to
+ * @type: node type
+ * @len: node length (not aligned)
+ * @lnum: LEB number of node to read
+ * @offs: offset of node to read
+ *
+ * This function tries to read a node of known type and length, checks it and
+ * stores it in @buf. This function returns %1 if a node is present and %0 if
+ * a node is not present. A negative error code is returned for I/O errors.
+ * This function performs that same function as ubifs_read_node except that
+ * it does not require that there is actually a node present and instead
+ * the return code indicates if a node was read.
+ *
+ * Note, this function does not check CRC of data nodes if @c->no_chk_data_crc
+ * is true (it is controlled by corresponding mount option). However, if
+ * @c->always_chk_crc is true, @c->no_chk_data_crc is ignored and CRC is always
+ * checked.
+ */
+static int try_read_node(const struct ubifs_info *c, void *buf, int type,
+ int len, int lnum, int offs)
+{
+ int err, node_len;
+ struct ubifs_ch *ch = buf;
+ uint32_t crc, node_crc;
+
+ dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len);
+
+ err = ubi_read(c->ubi, lnum, buf, offs, len);
+ if (err) {
+ ubifs_err("cannot read node type %d from LEB %d:%d, error %d",
+ type, lnum, offs, err);
+ return err;
+ }
+
+ if (le32_to_cpu(ch->magic) != UBIFS_NODE_MAGIC)
+ return 0;
+
+ if (ch->node_type != type)
+ return 0;
+
+ node_len = le32_to_cpu(ch->len);
+ if (node_len != len)
+ return 0;
+
+ if (type == UBIFS_DATA_NODE && !c->always_chk_crc && c->no_chk_data_crc)
+ return 1;
+
+ crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
+ node_crc = le32_to_cpu(ch->crc);
+ if (crc != node_crc)
+ return 0;
+
+ return 1;
+}
+
+/**
+ * fallible_read_node - try to read a leaf node.
+ * @c: UBIFS file-system description object
+ * @key: key of node to read
+ * @zbr: position of node
+ * @node: node returned
+ *
+ * This function tries to read a node and returns %1 if the node is read, %0
+ * if the node is not present, and a negative error code in the case of error.
+ */
+static int fallible_read_node(struct ubifs_info *c, const union ubifs_key *key,
+ struct ubifs_zbranch *zbr, void *node)
+{
+ int ret;
+
+ dbg_tnc("LEB %d:%d, key %s", zbr->lnum, zbr->offs, DBGKEY(key));
+
+ ret = try_read_node(c, node, key_type(c, key), zbr->len, zbr->lnum,
+ zbr->offs);
+ if (ret == 1) {
+ union ubifs_key node_key;
+ struct ubifs_dent_node *dent = node;
+
+ /* All nodes have key in the same place */
+ key_read(c, &dent->key, &node_key);
+ if (keys_cmp(c, key, &node_key) != 0)
+ ret = 0;
+ }
+ if (ret == 0 && c->replaying)
+ dbg_mnt("dangling branch LEB %d:%d len %d, key %s",
+ zbr->lnum, zbr->offs, zbr->len, DBGKEY(key));
+ return ret;
+}
+
+/**
+ * matches_name - determine if a direntry or xattr entry matches a given name.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch of dent
+ * @nm: name to match
+ *
+ * This function checks if xentry/direntry referred by zbranch @zbr matches name
+ * @nm. Returns %NAME_MATCHES if it does, %NAME_LESS if the name referred by
+ * @zbr is less than @nm, and %NAME_GREATER if it is greater than @nm. In case
+ * of failure, a negative error code is returned.
+ */
+static int matches_name(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ const struct qstr *nm)
+{
+ struct ubifs_dent_node *dent;
+ int nlen, err;
+
+ /* If possible, match against the dent in the leaf node cache */
+ if (!zbr->leaf) {
+ dent = kmalloc(zbr->len, GFP_NOFS);
+ if (!dent)
+ return -ENOMEM;
+
+ err = ubifs_tnc_read_node(c, zbr, dent);
+ if (err)
+ goto out_free;
+
+ /* Add the node to the leaf node cache */
+ err = lnc_add_directly(c, zbr, dent);
+ if (err)
+ goto out_free;
+ } else
+ dent = zbr->leaf;
+
+ nlen = le16_to_cpu(dent->nlen);
+ err = memcmp(dent->name, nm->name, min_t(int, nlen, nm->len));
+ if (err == 0) {
+ if (nlen == nm->len)
+ return NAME_MATCHES;
+ else if (nlen < nm->len)
+ return NAME_LESS;
+ else
+ return NAME_GREATER;
+ } else if (err < 0)
+ return NAME_LESS;
+ else
+ return NAME_GREATER;
+
+out_free:
+ kfree(dent);
+ return err;
+}
+
+/**
+ * get_znode - get a TNC znode that may not be loaded yet.
+ * @c: UBIFS file-system description object
+ * @znode: parent znode
+ * @n: znode branch slot number
+ *
+ * This function returns the znode or a negative error code.
+ */
+static struct ubifs_znode *get_znode(struct ubifs_info *c,
+ struct ubifs_znode *znode, int n)
+{
+ struct ubifs_zbranch *zbr;
+
+ zbr = &znode->zbranch[n];
+ if (zbr->znode)
+ znode = zbr->znode;
+ else
+ znode = ubifs_load_znode(c, zbr, znode, n);
+ return znode;
+}
+
+/**
+ * tnc_next - find next TNC entry.
+ * @c: UBIFS file-system description object
+ * @zn: znode is passed and returned here
+ * @n: znode branch slot number is passed and returned here
+ *
+ * This function returns %0 if the next TNC entry is found, %-ENOENT if there is
+ * no next entry, or a negative error code otherwise.
+ */
+static int tnc_next(struct ubifs_info *c, struct ubifs_znode **zn, int *n)
+{
+ struct ubifs_znode *znode = *zn;
+ int nn = *n;
+
+ nn += 1;
+ if (nn < znode->child_cnt) {
+ *n = nn;
+ return 0;
+ }
+ while (1) {
+ struct ubifs_znode *zp;
+
+ zp = znode->parent;
+ if (!zp)
+ return -ENOENT;
+ nn = znode->iip + 1;
+ znode = zp;
+ if (nn < znode->child_cnt) {
+ znode = get_znode(c, znode, nn);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ while (znode->level != 0) {
+ znode = get_znode(c, znode, 0);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+ nn = 0;
+ break;
+ }
+ }
+ *zn = znode;
+ *n = nn;
+ return 0;
+}
+
+/**
+ * tnc_prev - find previous TNC entry.
+ * @c: UBIFS file-system description object
+ * @zn: znode is returned here
+ * @n: znode branch slot number is passed and returned here
+ *
+ * This function returns %0 if the previous TNC entry is found, %-ENOENT if
+ * there is no next entry, or a negative error code otherwise.
+ */
+static int tnc_prev(struct ubifs_info *c, struct ubifs_znode **zn, int *n)
+{
+ struct ubifs_znode *znode = *zn;
+ int nn = *n;
+
+ if (nn > 0) {
+ *n = nn - 1;
+ return 0;
+ }
+ while (1) {
+ struct ubifs_znode *zp;
+
+ zp = znode->parent;
+ if (!zp)
+ return -ENOENT;
+ nn = znode->iip - 1;
+ znode = zp;
+ if (nn >= 0) {
+ znode = get_znode(c, znode, nn);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ while (znode->level != 0) {
+ nn = znode->child_cnt - 1;
+ znode = get_znode(c, znode, nn);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+ nn = znode->child_cnt - 1;
+ break;
+ }
+ }
+ *zn = znode;
+ *n = nn;
+ return 0;
+}
+
+/**
+ * resolve_collision - resolve a collision.
+ * @c: UBIFS file-system description object
+ * @key: key of a directory or extended attribute entry
+ * @zn: znode is returned here
+ * @n: zbranch number is passed and returned here
+ * @nm: name of the entry
+ *
+ * This function is called for "hashed" keys to make sure that the found key
+ * really corresponds to the looked up node (directory or extended attribute
+ * entry). It returns %1 and sets @zn and @n if the collision is resolved.
+ * %0 is returned if @nm is not found and @zn and @n are set to the previous
+ * entry, i.e. to the entry after which @nm could follow if it were in TNC.
+ * This means that @n may be set to %-1 if the leftmost key in @zn is the
+ * previous one. A negative error code is returned on failures.
+ */
+static int resolve_collision(struct ubifs_info *c, const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n,
+ const struct qstr *nm)
+{
+ int err;
+
+ err = matches_name(c, &(*zn)->zbranch[*n], nm);
+ if (unlikely(err < 0))
+ return err;
+ if (err == NAME_MATCHES)
+ return 1;
+
+ if (err == NAME_GREATER) {
+ /* Look left */
+ while (1) {
+ err = tnc_prev(c, zn, n);
+ if (err == -ENOENT) {
+ ubifs_assert(*n == 0);
+ *n = -1;
+ return 0;
+ }
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &(*zn)->zbranch[*n].key, key)) {
+ /*
+ * We have found the branch after which we would
+ * like to insert, but inserting in this znode
+ * may still be wrong. Consider the following 3
+ * znodes, in the case where we are resolving a
+ * collision with Key2.
+ *
+ * znode zp
+ * ----------------------
+ * level 1 | Key0 | Key1 |
+ * -----------------------
+ * | |
+ * znode za | | znode zb
+ * ------------ ------------
+ * level 0 | Key0 | | Key2 |
+ * ------------ ------------
+ *
+ * The lookup finds Key2 in znode zb. Lets say
+ * there is no match and the name is greater so
+ * we look left. When we find Key0, we end up
+ * here. If we return now, we will insert into
+ * znode za at slot n = 1. But that is invalid
+ * according to the parent's keys. Key2 must
+ * be inserted into znode zb.
+ *
+ * Note, this problem is not relevant for the
+ * case when we go right, because
+ * 'tnc_insert()' would correct the parent key.
+ */
+ if (*n == (*zn)->child_cnt - 1) {
+ err = tnc_next(c, zn, n);
+ if (err) {
+ /* Should be impossible */
+ ubifs_assert(0);
+ if (err == -ENOENT)
+ err = -EINVAL;
+ return err;
+ }
+ ubifs_assert(*n == 0);
+ *n = -1;
+ }
+ return 0;
+ }
+ err = matches_name(c, &(*zn)->zbranch[*n], nm);
+ if (err < 0)
+ return err;
+ if (err == NAME_LESS)
+ return 0;
+ if (err == NAME_MATCHES)
+ return 1;
+ ubifs_assert(err == NAME_GREATER);
+ }
+ } else {
+ int nn = *n;
+ struct ubifs_znode *znode = *zn;
+
+ /* Look right */
+ while (1) {
+ err = tnc_next(c, &znode, &nn);
+ if (err == -ENOENT)
+ return 0;
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &znode->zbranch[nn].key, key))
+ return 0;
+ err = matches_name(c, &znode->zbranch[nn], nm);
+ if (err < 0)
+ return err;
+ if (err == NAME_GREATER)
+ return 0;
+ *zn = znode;
+ *n = nn;
+ if (err == NAME_MATCHES)
+ return 1;
+ ubifs_assert(err == NAME_LESS);
+ }
+ }
+}
+
+/**
+ * fallible_matches_name - determine if a dent matches a given name.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch of dent
+ * @nm: name to match
+ *
+ * This is a "fallible" version of 'matches_name()' function which does not
+ * panic if the direntry/xentry referred by @zbr does not exist on the media.
+ *
+ * This function checks if xentry/direntry referred by zbranch @zbr matches name
+ * @nm. Returns %NAME_MATCHES it does, %NAME_LESS if the name referred by @zbr
+ * is less than @nm, %NAME_GREATER if it is greater than @nm, and @NOT_ON_MEDIA
+ * if xentry/direntry referred by @zbr does not exist on the media. A negative
+ * error code is returned in case of failure.
+ */
+static int fallible_matches_name(struct ubifs_info *c,
+ struct ubifs_zbranch *zbr,
+ const struct qstr *nm)
+{
+ struct ubifs_dent_node *dent;
+ int nlen, err;
+
+ /* If possible, match against the dent in the leaf node cache */
+ if (!zbr->leaf) {
+ dent = kmalloc(zbr->len, GFP_NOFS);
+ if (!dent)
+ return -ENOMEM;
+
+ err = fallible_read_node(c, &zbr->key, zbr, dent);
+ if (err < 0)
+ goto out_free;
+ if (err == 0) {
+ /* The node was not present */
+ err = NOT_ON_MEDIA;
+ goto out_free;
+ }
+ ubifs_assert(err == 1);
+
+ err = lnc_add_directly(c, zbr, dent);
+ if (err)
+ goto out_free;
+ } else
+ dent = zbr->leaf;
+
+ nlen = le16_to_cpu(dent->nlen);
+ err = memcmp(dent->name, nm->name, min_t(int, nlen, nm->len));
+ if (err == 0) {
+ if (nlen == nm->len)
+ return NAME_MATCHES;
+ else if (nlen < nm->len)
+ return NAME_LESS;
+ else
+ return NAME_GREATER;
+ } else if (err < 0)
+ return NAME_LESS;
+ else
+ return NAME_GREATER;
+
+out_free:
+ kfree(dent);
+ return err;
+}
+
+/**
+ * fallible_resolve_collision - resolve a collision even if nodes are missing.
+ * @c: UBIFS file-system description object
+ * @key: key
+ * @zn: znode is returned here
+ * @n: branch number is passed and returned here
+ * @nm: name of directory entry
+ * @adding: indicates caller is adding a key to the TNC
+ *
+ * This is a "fallible" version of the 'resolve_collision()' function which
+ * does not panic if one of the nodes referred to by TNC does not exist on the
+ * media. This may happen when replaying the journal if a deleted node was
+ * Garbage-collected and the commit was not done. A branch that refers to a node
+ * that is not present is called a dangling branch. The following are the return
+ * codes for this function:
+ * o if @nm was found, %1 is returned and @zn and @n are set to the found
+ * branch;
+ * o if we are @adding and @nm was not found, %0 is returned;
+ * o if we are not @adding and @nm was not found, but a dangling branch was
+ * found, then %1 is returned and @zn and @n are set to the dangling branch;
+ * o a negative error code is returned in case of failure.
+ */
+static int fallible_resolve_collision(struct ubifs_info *c,
+ const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n,
+ const struct qstr *nm, int adding)
+{
+ struct ubifs_znode *o_znode = NULL, *znode = *zn;
+ int uninitialized_var(o_n), err, cmp, unsure = 0, nn = *n;
+
+ cmp = fallible_matches_name(c, &znode->zbranch[nn], nm);
+ if (unlikely(cmp < 0))
+ return cmp;
+ if (cmp == NAME_MATCHES)
+ return 1;
+ if (cmp == NOT_ON_MEDIA) {
+ o_znode = znode;
+ o_n = nn;
+ /*
+ * We are unlucky and hit a dangling branch straight away.
+ * Now we do not really know where to go to find the needed
+ * branch - to the left or to the right. Well, let's try left.
+ */
+ unsure = 1;
+ } else if (!adding)
+ unsure = 1; /* Remove a dangling branch wherever it is */
+
+ if (cmp == NAME_GREATER || unsure) {
+ /* Look left */
+ while (1) {
+ err = tnc_prev(c, zn, n);
+ if (err == -ENOENT) {
+ ubifs_assert(*n == 0);
+ *n = -1;
+ break;
+ }
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &(*zn)->zbranch[*n].key, key)) {
+ /* See comments in 'resolve_collision()' */
+ if (*n == (*zn)->child_cnt - 1) {
+ err = tnc_next(c, zn, n);
+ if (err) {
+ /* Should be impossible */
+ ubifs_assert(0);
+ if (err == -ENOENT)
+ err = -EINVAL;
+ return err;
+ }
+ ubifs_assert(*n == 0);
+ *n = -1;
+ }
+ break;
+ }
+ err = fallible_matches_name(c, &(*zn)->zbranch[*n], nm);
+ if (err < 0)
+ return err;
+ if (err == NAME_MATCHES)
+ return 1;
+ if (err == NOT_ON_MEDIA) {
+ o_znode = *zn;
+ o_n = *n;
+ continue;
+ }
+ if (!adding)
+ continue;
+ if (err == NAME_LESS)
+ break;
+ else
+ unsure = 0;
+ }
+ }
+
+ if (cmp == NAME_LESS || unsure) {
+ /* Look right */
+ *zn = znode;
+ *n = nn;
+ while (1) {
+ err = tnc_next(c, &znode, &nn);
+ if (err == -ENOENT)
+ break;
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &znode->zbranch[nn].key, key))
+ break;
+ err = fallible_matches_name(c, &znode->zbranch[nn], nm);
+ if (err < 0)
+ return err;
+ if (err == NAME_GREATER)
+ break;
+ *zn = znode;
+ *n = nn;
+ if (err == NAME_MATCHES)
+ return 1;
+ if (err == NOT_ON_MEDIA) {
+ o_znode = znode;
+ o_n = nn;
+ }
+ }
+ }
+
+ /* Never match a dangling branch when adding */
+ if (adding || !o_znode)
+ return 0;
+
+ dbg_mnt("dangling match LEB %d:%d len %d %s",
+ o_znode->zbranch[o_n].lnum, o_znode->zbranch[o_n].offs,
+ o_znode->zbranch[o_n].len, DBGKEY(key));
+ *zn = o_znode;
+ *n = o_n;
+ return 1;
+}
+
+/**
+ * matches_position - determine if a zbranch matches a given position.
+ * @zbr: zbranch of dent
+ * @lnum: LEB number of dent to match
+ * @offs: offset of dent to match
+ *
+ * This function returns %1 if @lnum:@offs matches, and %0 otherwise.
+ */
+static int matches_position(struct ubifs_zbranch *zbr, int lnum, int offs)
+{
+ if (zbr->lnum == lnum && zbr->offs == offs)
+ return 1;
+ else
+ return 0;
+}
+
+/**
+ * resolve_collision_directly - resolve a collision directly.
+ * @c: UBIFS file-system description object
+ * @key: key of directory entry
+ * @zn: znode is passed and returned here
+ * @n: zbranch number is passed and returned here
+ * @lnum: LEB number of dent node to match
+ * @offs: offset of dent node to match
+ *
+ * This function is used for "hashed" keys to make sure the found directory or
+ * extended attribute entry node is what was looked for. It is used when the
+ * flash address of the right node is known (@lnum:@offs) which makes it much
+ * easier to resolve collisions (no need to read entries and match full
+ * names). This function returns %1 and sets @zn and @n if the collision is
+ * resolved, %0 if @lnum:@offs is not found and @zn and @n are set to the
+ * previous directory entry. Otherwise a negative error code is returned.
+ */
+static int resolve_collision_directly(struct ubifs_info *c,
+ const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n,
+ int lnum, int offs)
+{
+ struct ubifs_znode *znode;
+ int nn, err;
+
+ znode = *zn;
+ nn = *n;
+ if (matches_position(&znode->zbranch[nn], lnum, offs))
+ return 1;
+
+ /* Look left */
+ while (1) {
+ err = tnc_prev(c, &znode, &nn);
+ if (err == -ENOENT)
+ break;
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &znode->zbranch[nn].key, key))
+ break;
+ if (matches_position(&znode->zbranch[nn], lnum, offs)) {
+ *zn = znode;
+ *n = nn;
+ return 1;
+ }
+ }
+
+ /* Look right */
+ znode = *zn;
+ nn = *n;
+ while (1) {
+ err = tnc_next(c, &znode, &nn);
+ if (err == -ENOENT)
+ return 0;
+ if (err < 0)
+ return err;
+ if (keys_cmp(c, &znode->zbranch[nn].key, key))
+ return 0;
+ *zn = znode;
+ *n = nn;
+ if (matches_position(&znode->zbranch[nn], lnum, offs))
+ return 1;
+ }
+}
+
+/**
+ * dirty_cow_bottom_up - dirty a znode and its ancestors.
+ * @c: UBIFS file-system description object
+ * @znode: znode to dirty
+ *
+ * If we do not have a unique key that resides in a znode, then we cannot
+ * dirty that znode from the top down (i.e. by using lookup_level0_dirty)
+ * This function records the path back to the last dirty ancestor, and then
+ * dirties the znodes on that path.
+ */
+static struct ubifs_znode *dirty_cow_bottom_up(struct ubifs_info *c,
+ struct ubifs_znode *znode)
+{
+ struct ubifs_znode *zp;
+ int *path = c->bottom_up_buf, p = 0;
+
+ ubifs_assert(c->zroot.znode);
+ ubifs_assert(znode);
+ if (c->zroot.znode->level > BOTTOM_UP_HEIGHT) {
+ kfree(c->bottom_up_buf);
+ c->bottom_up_buf = kmalloc(c->zroot.znode->level * sizeof(int),
+ GFP_NOFS);
+ if (!c->bottom_up_buf)
+ return ERR_PTR(-ENOMEM);
+ path = c->bottom_up_buf;
+ }
+ if (c->zroot.znode->level) {
+ /* Go up until parent is dirty */
+ while (1) {
+ int n;
+
+ zp = znode->parent;
+ if (!zp)
+ break;
+ n = znode->iip;
+ ubifs_assert(p < c->zroot.znode->level);
+ path[p++] = n;
+ if (!zp->cnext && ubifs_zn_dirty(znode))
+ break;
+ znode = zp;
+ }
+ }
+
+ /* Come back down, dirtying as we go */
+ while (1) {
+ struct ubifs_zbranch *zbr;
+
+ zp = znode->parent;
+ if (zp) {
+ ubifs_assert(path[p - 1] >= 0);
+ ubifs_assert(path[p - 1] < zp->child_cnt);
+ zbr = &zp->zbranch[path[--p]];
+ znode = dirty_cow_znode(c, zbr);
+ } else {
+ ubifs_assert(znode == c->zroot.znode);
+ znode = dirty_cow_znode(c, &c->zroot);
+ }
+ if (IS_ERR(znode) || !p)
+ break;
+ ubifs_assert(path[p - 1] >= 0);
+ ubifs_assert(path[p - 1] < znode->child_cnt);
+ znode = znode->zbranch[path[p - 1]].znode;
+ }
+
+ return znode;
+}
+
+/**
+ * ubifs_lookup_level0 - search for zero-level znode.
+ * @c: UBIFS file-system description object
+ * @key: key to lookup
+ * @zn: znode is returned here
+ * @n: znode branch slot number is returned here
+ *
+ * This function looks up the TNC tree and search for zero-level znode which
+ * refers key @key. The found zero-level znode is returned in @zn. There are 3
+ * cases:
+ * o exact match, i.e. the found zero-level znode contains key @key, then %1
+ * is returned and slot number of the matched branch is stored in @n;
+ * o not exact match, which means that zero-level znode does not contain
+ * @key, then %0 is returned and slot number of the closed branch is stored
+ * in @n;
+ * o @key is so small that it is even less than the lowest key of the
+ * leftmost zero-level node, then %0 is returned and %0 is stored in @n.
+ *
+ * Note, when the TNC tree is traversed, some znodes may be absent, then this
+ * function reads corresponding indexing nodes and inserts them to TNC. In
+ * case of failure, a negative error code is returned.
+ */
+int ubifs_lookup_level0(struct ubifs_info *c, const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n)
+{
+ int err, exact;
+ struct ubifs_znode *znode;
+ unsigned long time = get_seconds();
+
+ dbg_tnc("search key %s", DBGKEY(key));
+
+ znode = c->zroot.znode;
+ if (unlikely(!znode)) {
+ znode = ubifs_load_znode(c, &c->zroot, NULL, 0);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+
+ znode->time = time;
+
+ while (1) {
+ struct ubifs_zbranch *zbr;
+
+ exact = ubifs_search_zbranch(c, znode, key, n);
+
+ if (znode->level == 0)
+ break;
+
+ if (*n < 0)
+ *n = 0;
+ zbr = &znode->zbranch[*n];
+
+ if (zbr->znode) {
+ znode->time = time;
+ znode = zbr->znode;
+ continue;
+ }
+
+ /* znode is not in TNC cache, load it from the media */
+ znode = ubifs_load_znode(c, zbr, znode, *n);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+
+ *zn = znode;
+ if (exact || !is_hash_key(c, key) || *n != -1) {
+ dbg_tnc("found %d, lvl %d, n %d", exact, znode->level, *n);
+ return exact;
+ }
+
+ /*
+ * Here is a tricky place. We have not found the key and this is a
+ * "hashed" key, which may collide. The rest of the code deals with
+ * situations like this:
+ *
+ * | 3 | 5 |
+ * / \
+ * | 3 | 5 | | 6 | 7 | (x)
+ *
+ * Or more a complex example:
+ *
+ * | 1 | 5 |
+ * / \
+ * | 1 | 3 | | 5 | 8 |
+ * \ /
+ * | 5 | 5 | | 6 | 7 | (x)
+ *
+ * In the examples, if we are looking for key "5", we may reach nodes
+ * marked with "(x)". In this case what we have do is to look at the
+ * left and see if there is "5" key there. If there is, we have to
+ * return it.
+ *
+ * Note, this whole situation is possible because we allow to have
+ * elements which are equivalent to the next key in the parent in the
+ * children of current znode. For example, this happens if we split a
+ * znode like this: | 3 | 5 | 5 | 6 | 7 |, which results in something
+ * like this:
+ * | 3 | 5 |
+ * / \
+ * | 3 | 5 | | 5 | 6 | 7 |
+ * ^
+ * And this becomes what is at the first "picture" after key "5" marked
+ * with "^" is removed. What could be done is we could prohibit
+ * splitting in the middle of the colliding sequence. Also, when
+ * removing the leftmost key, we would have to correct the key of the
+ * parent node, which would introduce additional complications. Namely,
+ * if we changed the the leftmost key of the parent znode, the garbage
+ * collector would be unable to find it (GC is doing this when GC'ing
+ * indexing LEBs). Although we already have an additional RB-tree where
+ * we save such changed znodes (see 'ins_clr_old_idx_znode()') until
+ * after the commit. But anyway, this does not look easy to implement
+ * so we did not try this.
+ */
+ err = tnc_prev(c, &znode, n);
+ if (err == -ENOENT) {
+ dbg_tnc("found 0, lvl %d, n -1", znode->level);
+ *n = -1;
+ return 0;
+ }
+ if (unlikely(err < 0))
+ return err;
+ if (keys_cmp(c, key, &znode->zbranch[*n].key)) {
+ dbg_tnc("found 0, lvl %d, n -1", znode->level);
+ *n = -1;
+ return 0;
+ }
+
+ dbg_tnc("found 1, lvl %d, n %d", znode->level, *n);
+ *zn = znode;
+ return 1;
+}
+
+/**
+ * lookup_level0_dirty - search for zero-level znode dirtying.
+ * @c: UBIFS file-system description object
+ * @key: key to lookup
+ * @zn: znode is returned here
+ * @n: znode branch slot number is returned here
+ *
+ * This function looks up the TNC tree and search for zero-level znode which
+ * refers key @key. The found zero-level znode is returned in @zn. There are 3
+ * cases:
+ * o exact match, i.e. the found zero-level znode contains key @key, then %1
+ * is returned and slot number of the matched branch is stored in @n;
+ * o not exact match, which means that zero-level znode does not contain @key
+ * then %0 is returned and slot number of the closed branch is stored in
+ * @n;
+ * o @key is so small that it is even less than the lowest key of the
+ * leftmost zero-level node, then %0 is returned and %-1 is stored in @n.
+ *
+ * Additionally all znodes in the path from the root to the located zero-level
+ * znode are marked as dirty.
+ *
+ * Note, when the TNC tree is traversed, some znodes may be absent, then this
+ * function reads corresponding indexing nodes and inserts them to TNC. In
+ * case of failure, a negative error code is returned.
+ */
+static int lookup_level0_dirty(struct ubifs_info *c, const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n)
+{
+ int err, exact;
+ struct ubifs_znode *znode;
+ unsigned long time = get_seconds();
+
+ dbg_tnc("search and dirty key %s", DBGKEY(key));
+
+ znode = c->zroot.znode;
+ if (unlikely(!znode)) {
+ znode = ubifs_load_znode(c, &c->zroot, NULL, 0);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+
+ znode = dirty_cow_znode(c, &c->zroot);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+
+ znode->time = time;
+
+ while (1) {
+ struct ubifs_zbranch *zbr;
+
+ exact = ubifs_search_zbranch(c, znode, key, n);
+
+ if (znode->level == 0)
+ break;
+
+ if (*n < 0)
+ *n = 0;
+ zbr = &znode->zbranch[*n];
+
+ if (zbr->znode) {
+ znode->time = time;
+ znode = dirty_cow_znode(c, zbr);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ continue;
+ }
+
+ /* znode is not in TNC cache, load it from the media */
+ znode = ubifs_load_znode(c, zbr, znode, *n);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ znode = dirty_cow_znode(c, zbr);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+
+ *zn = znode;
+ if (exact || !is_hash_key(c, key) || *n != -1) {
+ dbg_tnc("found %d, lvl %d, n %d", exact, znode->level, *n);
+ return exact;
+ }
+
+ /*
+ * See huge comment at 'lookup_level0_dirty()' what is the rest of the
+ * code.
+ */
+ err = tnc_prev(c, &znode, n);
+ if (err == -ENOENT) {
+ *n = -1;
+ dbg_tnc("found 0, lvl %d, n -1", znode->level);
+ return 0;
+ }
+ if (unlikely(err < 0))
+ return err;
+ if (keys_cmp(c, key, &znode->zbranch[*n].key)) {
+ *n = -1;
+ dbg_tnc("found 0, lvl %d, n -1", znode->level);
+ return 0;
+ }
+
+ if (znode->cnext || !ubifs_zn_dirty(znode)) {
+ znode = dirty_cow_bottom_up(c, znode);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ }
+
+ dbg_tnc("found 1, lvl %d, n %d", znode->level, *n);
+ *zn = znode;
+ return 1;
+}
+
+/**
+ * maybe_leb_gced - determine if a LEB may have been garbage collected.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number
+ * @gc_seq1: garbage collection sequence number
+ *
+ * This function determines if @lnum may have been garbage collected since
+ * sequence number @gc_seq1. If it may have been then %1 is returned, otherwise
+ * %0 is returned.
+ */
+static int maybe_leb_gced(struct ubifs_info *c, int lnum, int gc_seq1)
+{
+ /*
+ * No garbage collection in the read-only U-Boot implementation
+ */
+ return 0;
+}
+
+/**
+ * ubifs_tnc_locate - look up a file-system node and return it and its location.
+ * @c: UBIFS file-system description object
+ * @key: node key to lookup
+ * @node: the node is returned here
+ * @lnum: LEB number is returned here
+ * @offs: offset is returned here
+ *
+ * This function look up and reads node with key @key. The caller has to make
+ * sure the @node buffer is large enough to fit the node. Returns zero in case
+ * of success, %-ENOENT if the node was not found, and a negative error code in
+ * case of failure. The node location can be returned in @lnum and @offs.
+ */
+int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key,
+ void *node, int *lnum, int *offs)
+{
+ int found, n, err, safely = 0, gc_seq1;
+ struct ubifs_znode *znode;
+ struct ubifs_zbranch zbr, *zt;
+
+again:
+ mutex_lock(&c->tnc_mutex);
+ found = ubifs_lookup_level0(c, key, &znode, &n);
+ if (!found) {
+ err = -ENOENT;
+ goto out;
+ } else if (found < 0) {
+ err = found;
+ goto out;
+ }
+ zt = &znode->zbranch[n];
+ if (lnum) {
+ *lnum = zt->lnum;
+ *offs = zt->offs;
+ }
+ if (is_hash_key(c, key)) {
+ /*
+ * In this case the leaf node cache gets used, so we pass the
+ * address of the zbranch and keep the mutex locked
+ */
+ err = tnc_read_node_nm(c, zt, node);
+ goto out;
+ }
+ if (safely) {
+ err = ubifs_tnc_read_node(c, zt, node);
+ goto out;
+ }
+ /* Drop the TNC mutex prematurely and race with garbage collection */
+ zbr = znode->zbranch[n];
+ gc_seq1 = c->gc_seq;
+ mutex_unlock(&c->tnc_mutex);
+
+ err = fallible_read_node(c, key, &zbr, node);
+ if (err <= 0 || maybe_leb_gced(c, zbr.lnum, gc_seq1)) {
+ /*
+ * The node may have been GC'ed out from under us so try again
+ * while keeping the TNC mutex locked.
+ */
+ safely = 1;
+ goto again;
+ }
+ return 0;
+
+out:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * ubifs_tnc_get_bu_keys - lookup keys for bulk-read.
+ * @c: UBIFS file-system description object
+ * @bu: bulk-read parameters and results
+ *
+ * Lookup consecutive data node keys for the same inode that reside
+ * consecutively in the same LEB. This function returns zero in case of success
+ * and a negative error code in case of failure.
+ *
+ * Note, if the bulk-read buffer length (@bu->buf_len) is known, this function
+ * makes sure bulk-read nodes fit the buffer. Otherwise, this function prepares
+ * maximum possible amount of nodes for bulk-read.
+ */
+int ubifs_tnc_get_bu_keys(struct ubifs_info *c, struct bu_info *bu)
+{
+ int n, err = 0, lnum = -1, uninitialized_var(offs);
+ int uninitialized_var(len);
+ unsigned int block = key_block(c, &bu->key);
+ struct ubifs_znode *znode;
+
+ bu->cnt = 0;
+ bu->blk_cnt = 0;
+ bu->eof = 0;
+
+ mutex_lock(&c->tnc_mutex);
+ /* Find first key */
+ err = ubifs_lookup_level0(c, &bu->key, &znode, &n);
+ if (err < 0)
+ goto out;
+ if (err) {
+ /* Key found */
+ len = znode->zbranch[n].len;
+ /* The buffer must be big enough for at least 1 node */
+ if (len > bu->buf_len) {
+ err = -EINVAL;
+ goto out;
+ }
+ /* Add this key */
+ bu->zbranch[bu->cnt++] = znode->zbranch[n];
+ bu->blk_cnt += 1;
+ lnum = znode->zbranch[n].lnum;
+ offs = ALIGN(znode->zbranch[n].offs + len, 8);
+ }
+ while (1) {
+ struct ubifs_zbranch *zbr;
+ union ubifs_key *key;
+ unsigned int next_block;
+
+ /* Find next key */
+ err = tnc_next(c, &znode, &n);
+ if (err)
+ goto out;
+ zbr = &znode->zbranch[n];
+ key = &zbr->key;
+ /* See if there is another data key for this file */
+ if (key_inum(c, key) != key_inum(c, &bu->key) ||
+ key_type(c, key) != UBIFS_DATA_KEY) {
+ err = -ENOENT;
+ goto out;
+ }
+ if (lnum < 0) {
+ /* First key found */
+ lnum = zbr->lnum;
+ offs = ALIGN(zbr->offs + zbr->len, 8);
+ len = zbr->len;
+ if (len > bu->buf_len) {
+ err = -EINVAL;
+ goto out;
+ }
+ } else {
+ /*
+ * The data nodes must be in consecutive positions in
+ * the same LEB.
+ */
+ if (zbr->lnum != lnum || zbr->offs != offs)
+ goto out;
+ offs += ALIGN(zbr->len, 8);
+ len = ALIGN(len, 8) + zbr->len;
+ /* Must not exceed buffer length */
+ if (len > bu->buf_len)
+ goto out;
+ }
+ /* Allow for holes */
+ next_block = key_block(c, key);
+ bu->blk_cnt += (next_block - block - 1);
+ if (bu->blk_cnt >= UBIFS_MAX_BULK_READ)
+ goto out;
+ block = next_block;
+ /* Add this key */
+ bu->zbranch[bu->cnt++] = *zbr;
+ bu->blk_cnt += 1;
+ /* See if we have room for more */
+ if (bu->cnt >= UBIFS_MAX_BULK_READ)
+ goto out;
+ if (bu->blk_cnt >= UBIFS_MAX_BULK_READ)
+ goto out;
+ }
+out:
+ if (err == -ENOENT) {
+ bu->eof = 1;
+ err = 0;
+ }
+ bu->gc_seq = c->gc_seq;
+ mutex_unlock(&c->tnc_mutex);
+ if (err)
+ return err;
+ /*
+ * An enormous hole could cause bulk-read to encompass too many
+ * page cache pages, so limit the number here.
+ */
+ if (bu->blk_cnt > UBIFS_MAX_BULK_READ)
+ bu->blk_cnt = UBIFS_MAX_BULK_READ;
+ /*
+ * Ensure that bulk-read covers a whole number of page cache
+ * pages.
+ */
+ if (UBIFS_BLOCKS_PER_PAGE == 1 ||
+ !(bu->blk_cnt & (UBIFS_BLOCKS_PER_PAGE - 1)))
+ return 0;
+ if (bu->eof) {
+ /* At the end of file we can round up */
+ bu->blk_cnt += UBIFS_BLOCKS_PER_PAGE - 1;
+ return 0;
+ }
+ /* Exclude data nodes that do not make up a whole page cache page */
+ block = key_block(c, &bu->key) + bu->blk_cnt;
+ block &= ~(UBIFS_BLOCKS_PER_PAGE - 1);
+ while (bu->cnt) {
+ if (key_block(c, &bu->zbranch[bu->cnt - 1].key) < block)
+ break;
+ bu->cnt -= 1;
+ }
+ return 0;
+}
+
+/**
+ * validate_data_node - validate data nodes for bulk-read.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing data node to validate
+ * @zbr: zbranch of data node to validate
+ *
+ * This functions returns %0 on success or a negative error code on failure.
+ */
+static int validate_data_node(struct ubifs_info *c, void *buf,
+ struct ubifs_zbranch *zbr)
+{
+ union ubifs_key key1;
+ struct ubifs_ch *ch = buf;
+ int err, len;
+
+ if (ch->node_type != UBIFS_DATA_NODE) {
+ ubifs_err("bad node type (%d but expected %d)",
+ ch->node_type, UBIFS_DATA_NODE);
+ goto out_err;
+ }
+
+ err = ubifs_check_node(c, buf, zbr->lnum, zbr->offs, 0, 0);
+ if (err) {
+ ubifs_err("expected node type %d", UBIFS_DATA_NODE);
+ goto out;
+ }
+
+ len = le32_to_cpu(ch->len);
+ if (len != zbr->len) {
+ ubifs_err("bad node length %d, expected %d", len, zbr->len);
+ goto out_err;
+ }
+
+ /* Make sure the key of the read node is correct */
+ key_read(c, buf + UBIFS_KEY_OFFSET, &key1);
+ if (!keys_eq(c, &zbr->key, &key1)) {
+ ubifs_err("bad key in node at LEB %d:%d",
+ zbr->lnum, zbr->offs);
+ dbg_tnc("looked for key %s found node's key %s",
+ DBGKEY(&zbr->key), DBGKEY1(&key1));
+ goto out_err;
+ }
+
+ return 0;
+
+out_err:
+ err = -EINVAL;
+out:
+ ubifs_err("bad node at LEB %d:%d", zbr->lnum, zbr->offs);
+ dbg_dump_node(c, buf);
+ dbg_dump_stack();
+ return err;
+}
+
+/**
+ * ubifs_tnc_bulk_read - read a number of data nodes in one go.
+ * @c: UBIFS file-system description object
+ * @bu: bulk-read parameters and results
+ *
+ * This functions reads and validates the data nodes that were identified by the
+ * 'ubifs_tnc_get_bu_keys()' function. This functions returns %0 on success,
+ * -EAGAIN to indicate a race with GC, or another negative error code on
+ * failure.
+ */
+int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu)
+{
+ int lnum = bu->zbranch[0].lnum, offs = bu->zbranch[0].offs, len, err, i;
+ void *buf;
+
+ len = bu->zbranch[bu->cnt - 1].offs;
+ len += bu->zbranch[bu->cnt - 1].len - offs;
+ if (len > bu->buf_len) {
+ ubifs_err("buffer too small %d vs %d", bu->buf_len, len);
+ return -EINVAL;
+ }
+
+ /* Do the read */
+ err = ubi_read(c->ubi, lnum, bu->buf, offs, len);
+
+ /* Check for a race with GC */
+ if (maybe_leb_gced(c, lnum, bu->gc_seq))
+ return -EAGAIN;
+
+ if (err && err != -EBADMSG) {
+ ubifs_err("failed to read from LEB %d:%d, error %d",
+ lnum, offs, err);
+ dbg_dump_stack();
+ dbg_tnc("key %s", DBGKEY(&bu->key));
+ return err;
+ }
+
+ /* Validate the nodes read */
+ buf = bu->buf;
+ for (i = 0; i < bu->cnt; i++) {
+ err = validate_data_node(c, buf, &bu->zbranch[i]);
+ if (err)
+ return err;
+ buf = buf + ALIGN(bu->zbranch[i].len, 8);
+ }
+
+ return 0;
+}
+
+/**
+ * do_lookup_nm- look up a "hashed" node.
+ * @c: UBIFS file-system description object
+ * @key: node key to lookup
+ * @node: the node is returned here
+ * @nm: node name
+ *
+ * This function look up and reads a node which contains name hash in the key.
+ * Since the hash may have collisions, there may be many nodes with the same
+ * key, so we have to sequentially look to all of them until the needed one is
+ * found. This function returns zero in case of success, %-ENOENT if the node
+ * was not found, and a negative error code in case of failure.
+ */
+static int do_lookup_nm(struct ubifs_info *c, const union ubifs_key *key,
+ void *node, const struct qstr *nm)
+{
+ int found, n, err;
+ struct ubifs_znode *znode;
+
+ dbg_tnc("name '%.*s' key %s", nm->len, nm->name, DBGKEY(key));
+ mutex_lock(&c->tnc_mutex);
+ found = ubifs_lookup_level0(c, key, &znode, &n);
+ if (!found) {
+ err = -ENOENT;
+ goto out_unlock;
+ } else if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+
+ ubifs_assert(n >= 0);
+
+ err = resolve_collision(c, key, &znode, &n, nm);
+ dbg_tnc("rc returned %d, znode %p, n %d", err, znode, n);
+ if (unlikely(err < 0))
+ goto out_unlock;
+ if (err == 0) {
+ err = -ENOENT;
+ goto out_unlock;
+ }
+
+ err = tnc_read_node_nm(c, &znode->zbranch[n], node);
+
+out_unlock:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * ubifs_tnc_lookup_nm - look up a "hashed" node.
+ * @c: UBIFS file-system description object
+ * @key: node key to lookup
+ * @node: the node is returned here
+ * @nm: node name
+ *
+ * This function look up and reads a node which contains name hash in the key.
+ * Since the hash may have collisions, there may be many nodes with the same
+ * key, so we have to sequentially look to all of them until the needed one is
+ * found. This function returns zero in case of success, %-ENOENT if the node
+ * was not found, and a negative error code in case of failure.
+ */
+int ubifs_tnc_lookup_nm(struct ubifs_info *c, const union ubifs_key *key,
+ void *node, const struct qstr *nm)
+{
+ int err, len;
+ const struct ubifs_dent_node *dent = node;
+
+ /*
+ * We assume that in most of the cases there are no name collisions and
+ * 'ubifs_tnc_lookup()' returns us the right direntry.
+ */
+ err = ubifs_tnc_lookup(c, key, node);
+ if (err)
+ return err;
+
+ len = le16_to_cpu(dent->nlen);
+ if (nm->len == len && !memcmp(dent->name, nm->name, len))
+ return 0;
+
+ /*
+ * Unluckily, there are hash collisions and we have to iterate over
+ * them look at each direntry with colliding name hash sequentially.
+ */
+ return do_lookup_nm(c, key, node, nm);
+}
+
+/**
+ * correct_parent_keys - correct parent znodes' keys.
+ * @c: UBIFS file-system description object
+ * @znode: znode to correct parent znodes for
+ *
+ * This is a helper function for 'tnc_insert()'. When the key of the leftmost
+ * zbranch changes, keys of parent znodes have to be corrected. This helper
+ * function is called in such situations and corrects the keys if needed.
+ */
+static void correct_parent_keys(const struct ubifs_info *c,
+ struct ubifs_znode *znode)
+{
+ union ubifs_key *key, *key1;
+
+ ubifs_assert(znode->parent);
+ ubifs_assert(znode->iip == 0);
+
+ key = &znode->zbranch[0].key;
+ key1 = &znode->parent->zbranch[0].key;
+
+ while (keys_cmp(c, key, key1) < 0) {
+ key_copy(c, key, key1);
+ znode = znode->parent;
+ znode->alt = 1;
+ if (!znode->parent || znode->iip)
+ break;
+ key1 = &znode->parent->zbranch[0].key;
+ }
+}
+
+/**
+ * insert_zbranch - insert a zbranch into a znode.
+ * @znode: znode into which to insert
+ * @zbr: zbranch to insert
+ * @n: slot number to insert to
+ *
+ * This is a helper function for 'tnc_insert()'. UBIFS does not allow "gaps" in
+ * znode's array of zbranches and keeps zbranches consolidated, so when a new
+ * zbranch has to be inserted to the @znode->zbranches[]' array at the @n-th
+ * slot, zbranches starting from @n have to be moved right.
+ */
+static void insert_zbranch(struct ubifs_znode *znode,
+ const struct ubifs_zbranch *zbr, int n)
+{
+ int i;
+
+ ubifs_assert(ubifs_zn_dirty(znode));
+
+ if (znode->level) {
+ for (i = znode->child_cnt; i > n; i--) {
+ znode->zbranch[i] = znode->zbranch[i - 1];
+ if (znode->zbranch[i].znode)
+ znode->zbranch[i].znode->iip = i;
+ }
+ if (zbr->znode)
+ zbr->znode->iip = n;
+ } else
+ for (i = znode->child_cnt; i > n; i--)
+ znode->zbranch[i] = znode->zbranch[i - 1];
+
+ znode->zbranch[n] = *zbr;
+ znode->child_cnt += 1;
+
+ /*
+ * After inserting at slot zero, the lower bound of the key range of
+ * this znode may have changed. If this znode is subsequently split
+ * then the upper bound of the key range may change, and furthermore
+ * it could change to be lower than the original lower bound. If that
+ * happens, then it will no longer be possible to find this znode in the
+ * TNC using the key from the index node on flash. That is bad because
+ * if it is not found, we will assume it is obsolete and may overwrite
+ * it. Then if there is an unclean unmount, we will start using the
+ * old index which will be broken.
+ *
+ * So we first mark znodes that have insertions at slot zero, and then
+ * if they are split we add their lnum/offs to the old_idx tree.
+ */
+ if (n == 0)
+ znode->alt = 1;
+}
+
+/**
+ * tnc_insert - insert a node into TNC.
+ * @c: UBIFS file-system description object
+ * @znode: znode to insert into
+ * @zbr: branch to insert
+ * @n: slot number to insert new zbranch to
+ *
+ * This function inserts a new node described by @zbr into znode @znode. If
+ * znode does not have a free slot for new zbranch, it is split. Parent znodes
+ * are splat as well if needed. Returns zero in case of success or a negative
+ * error code in case of failure.
+ */
+static int tnc_insert(struct ubifs_info *c, struct ubifs_znode *znode,
+ struct ubifs_zbranch *zbr, int n)
+{
+ struct ubifs_znode *zn, *zi, *zp;
+ int i, keep, move, appending = 0;
+ union ubifs_key *key = &zbr->key, *key1;
+
+ ubifs_assert(n >= 0 && n <= c->fanout);
+
+ /* Implement naive insert for now */
+again:
+ zp = znode->parent;
+ if (znode->child_cnt < c->fanout) {
+ ubifs_assert(n != c->fanout);
+ dbg_tnc("inserted at %d level %d, key %s", n, znode->level,
+ DBGKEY(key));
+
+ insert_zbranch(znode, zbr, n);
+
+ /* Ensure parent's key is correct */
+ if (n == 0 && zp && znode->iip == 0)
+ correct_parent_keys(c, znode);
+
+ return 0;
+ }
+
+ /*
+ * Unfortunately, @znode does not have more empty slots and we have to
+ * split it.
+ */
+ dbg_tnc("splitting level %d, key %s", znode->level, DBGKEY(key));
+
+ if (znode->alt)
+ /*
+ * We can no longer be sure of finding this znode by key, so we
+ * record it in the old_idx tree.
+ */
+ ins_clr_old_idx_znode(c, znode);
+
+ zn = kzalloc(c->max_znode_sz, GFP_NOFS);
+ if (!zn)
+ return -ENOMEM;
+ zn->parent = zp;
+ zn->level = znode->level;
+
+ /* Decide where to split */
+ if (znode->level == 0 && key_type(c, key) == UBIFS_DATA_KEY) {
+ /* Try not to split consecutive data keys */
+ if (n == c->fanout) {
+ key1 = &znode->zbranch[n - 1].key;
+ if (key_inum(c, key1) == key_inum(c, key) &&
+ key_type(c, key1) == UBIFS_DATA_KEY)
+ appending = 1;
+ } else
+ goto check_split;
+ } else if (appending && n != c->fanout) {
+ /* Try not to split consecutive data keys */
+ appending = 0;
+check_split:
+ if (n >= (c->fanout + 1) / 2) {
+ key1 = &znode->zbranch[0].key;
+ if (key_inum(c, key1) == key_inum(c, key) &&
+ key_type(c, key1) == UBIFS_DATA_KEY) {
+ key1 = &znode->zbranch[n].key;
+ if (key_inum(c, key1) != key_inum(c, key) ||
+ key_type(c, key1) != UBIFS_DATA_KEY) {
+ keep = n;
+ move = c->fanout - keep;
+ zi = znode;
+ goto do_split;
+ }
+ }
+ }
+ }
+
+ if (appending) {
+ keep = c->fanout;
+ move = 0;
+ } else {
+ keep = (c->fanout + 1) / 2;
+ move = c->fanout - keep;
+ }
+
+ /*
+ * Although we don't at present, we could look at the neighbors and see
+ * if we can move some zbranches there.
+ */
+
+ if (n < keep) {
+ /* Insert into existing znode */
+ zi = znode;
+ move += 1;
+ keep -= 1;
+ } else {
+ /* Insert into new znode */
+ zi = zn;
+ n -= keep;
+ /* Re-parent */
+ if (zn->level != 0)
+ zbr->znode->parent = zn;
+ }
+
+do_split:
+
+ __set_bit(DIRTY_ZNODE, &zn->flags);
+ atomic_long_inc(&c->dirty_zn_cnt);
+
+ zn->child_cnt = move;
+ znode->child_cnt = keep;
+
+ dbg_tnc("moving %d, keeping %d", move, keep);
+
+ /* Move zbranch */
+ for (i = 0; i < move; i++) {
+ zn->zbranch[i] = znode->zbranch[keep + i];
+ /* Re-parent */
+ if (zn->level != 0)
+ if (zn->zbranch[i].znode) {
+ zn->zbranch[i].znode->parent = zn;
+ zn->zbranch[i].znode->iip = i;
+ }
+ }
+
+ /* Insert new key and branch */
+ dbg_tnc("inserting at %d level %d, key %s", n, zn->level, DBGKEY(key));
+
+ insert_zbranch(zi, zbr, n);
+
+ /* Insert new znode (produced by spitting) into the parent */
+ if (zp) {
+ if (n == 0 && zi == znode && znode->iip == 0)
+ correct_parent_keys(c, znode);
+
+ /* Locate insertion point */
+ n = znode->iip + 1;
+
+ /* Tail recursion */
+ zbr->key = zn->zbranch[0].key;
+ zbr->znode = zn;
+ zbr->lnum = 0;
+ zbr->offs = 0;
+ zbr->len = 0;
+ znode = zp;
+
+ goto again;
+ }
+
+ /* We have to split root znode */
+ dbg_tnc("creating new zroot at level %d", znode->level + 1);
+
+ zi = kzalloc(c->max_znode_sz, GFP_NOFS);
+ if (!zi)
+ return -ENOMEM;
+
+ zi->child_cnt = 2;
+ zi->level = znode->level + 1;
+
+ __set_bit(DIRTY_ZNODE, &zi->flags);
+ atomic_long_inc(&c->dirty_zn_cnt);
+
+ zi->zbranch[0].key = znode->zbranch[0].key;
+ zi->zbranch[0].znode = znode;
+ zi->zbranch[0].lnum = c->zroot.lnum;
+ zi->zbranch[0].offs = c->zroot.offs;
+ zi->zbranch[0].len = c->zroot.len;
+ zi->zbranch[1].key = zn->zbranch[0].key;
+ zi->zbranch[1].znode = zn;
+
+ c->zroot.lnum = 0;
+ c->zroot.offs = 0;
+ c->zroot.len = 0;
+ c->zroot.znode = zi;
+
+ zn->parent = zi;
+ zn->iip = 1;
+ znode->parent = zi;
+ znode->iip = 0;
+
+ return 0;
+}
+
+/**
+ * ubifs_tnc_add - add a node to TNC.
+ * @c: UBIFS file-system description object
+ * @key: key to add
+ * @lnum: LEB number of node
+ * @offs: node offset
+ * @len: node length
+ *
+ * This function adds a node with key @key to TNC. The node may be new or it may
+ * obsolete some existing one. Returns %0 on success or negative error code on
+ * failure.
+ */
+int ubifs_tnc_add(struct ubifs_info *c, const union ubifs_key *key, int lnum,
+ int offs, int len)
+{
+ int found, n, err = 0;
+ struct ubifs_znode *znode;
+
+ mutex_lock(&c->tnc_mutex);
+ dbg_tnc("%d:%d, len %d, key %s", lnum, offs, len, DBGKEY(key));
+ found = lookup_level0_dirty(c, key, &znode, &n);
+ if (!found) {
+ struct ubifs_zbranch zbr;
+
+ zbr.znode = NULL;
+ zbr.lnum = lnum;
+ zbr.offs = offs;
+ zbr.len = len;
+ key_copy(c, key, &zbr.key);
+ err = tnc_insert(c, znode, &zbr, n + 1);
+ } else if (found == 1) {
+ struct ubifs_zbranch *zbr = &znode->zbranch[n];
+
+ lnc_free(zbr);
+ err = ubifs_add_dirt(c, zbr->lnum, zbr->len);
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ } else
+ err = found;
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+ mutex_unlock(&c->tnc_mutex);
+
+ return err;
+}
+
+/**
+ * ubifs_tnc_replace - replace a node in the TNC only if the old node is found.
+ * @c: UBIFS file-system description object
+ * @key: key to add
+ * @old_lnum: LEB number of old node
+ * @old_offs: old node offset
+ * @lnum: LEB number of node
+ * @offs: node offset
+ * @len: node length
+ *
+ * This function replaces a node with key @key in the TNC only if the old node
+ * is found. This function is called by garbage collection when node are moved.
+ * Returns %0 on success or negative error code on failure.
+ */
+int ubifs_tnc_replace(struct ubifs_info *c, const union ubifs_key *key,
+ int old_lnum, int old_offs, int lnum, int offs, int len)
+{
+ int found, n, err = 0;
+ struct ubifs_znode *znode;
+
+ mutex_lock(&c->tnc_mutex);
+ dbg_tnc("old LEB %d:%d, new LEB %d:%d, len %d, key %s", old_lnum,
+ old_offs, lnum, offs, len, DBGKEY(key));
+ found = lookup_level0_dirty(c, key, &znode, &n);
+ if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+
+ if (found == 1) {
+ struct ubifs_zbranch *zbr = &znode->zbranch[n];
+
+ found = 0;
+ if (zbr->lnum == old_lnum && zbr->offs == old_offs) {
+ lnc_free(zbr);
+ err = ubifs_add_dirt(c, zbr->lnum, zbr->len);
+ if (err)
+ goto out_unlock;
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ found = 1;
+ } else if (is_hash_key(c, key)) {
+ found = resolve_collision_directly(c, key, &znode, &n,
+ old_lnum, old_offs);
+ dbg_tnc("rc returned %d, znode %p, n %d, LEB %d:%d",
+ found, znode, n, old_lnum, old_offs);
+ if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+
+ if (found) {
+ /* Ensure the znode is dirtied */
+ if (znode->cnext || !ubifs_zn_dirty(znode)) {
+ znode = dirty_cow_bottom_up(c, znode);
+ if (IS_ERR(znode)) {
+ err = PTR_ERR(znode);
+ goto out_unlock;
+ }
+ }
+ zbr = &znode->zbranch[n];
+ lnc_free(zbr);
+ err = ubifs_add_dirt(c, zbr->lnum,
+ zbr->len);
+ if (err)
+ goto out_unlock;
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ }
+ }
+ }
+
+ if (!found)
+ err = ubifs_add_dirt(c, lnum, len);
+
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+
+out_unlock:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * ubifs_tnc_add_nm - add a "hashed" node to TNC.
+ * @c: UBIFS file-system description object
+ * @key: key to add
+ * @lnum: LEB number of node
+ * @offs: node offset
+ * @len: node length
+ * @nm: node name
+ *
+ * This is the same as 'ubifs_tnc_add()' but it should be used with keys which
+ * may have collisions, like directory entry keys.
+ */
+int ubifs_tnc_add_nm(struct ubifs_info *c, const union ubifs_key *key,
+ int lnum, int offs, int len, const struct qstr *nm)
+{
+ int found, n, err = 0;
+ struct ubifs_znode *znode;
+
+ mutex_lock(&c->tnc_mutex);
+ dbg_tnc("LEB %d:%d, name '%.*s', key %s", lnum, offs, nm->len, nm->name,
+ DBGKEY(key));
+ found = lookup_level0_dirty(c, key, &znode, &n);
+ if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+
+ if (found == 1) {
+ if (c->replaying)
+ found = fallible_resolve_collision(c, key, &znode, &n,
+ nm, 1);
+ else
+ found = resolve_collision(c, key, &znode, &n, nm);
+ dbg_tnc("rc returned %d, znode %p, n %d", found, znode, n);
+ if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+
+ /* Ensure the znode is dirtied */
+ if (znode->cnext || !ubifs_zn_dirty(znode)) {
+ znode = dirty_cow_bottom_up(c, znode);
+ if (IS_ERR(znode)) {
+ err = PTR_ERR(znode);
+ goto out_unlock;
+ }
+ }
+
+ if (found == 1) {
+ struct ubifs_zbranch *zbr = &znode->zbranch[n];
+
+ lnc_free(zbr);
+ err = ubifs_add_dirt(c, zbr->lnum, zbr->len);
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ goto out_unlock;
+ }
+ }
+
+ if (!found) {
+ struct ubifs_zbranch zbr;
+
+ zbr.znode = NULL;
+ zbr.lnum = lnum;
+ zbr.offs = offs;
+ zbr.len = len;
+ key_copy(c, key, &zbr.key);
+ err = tnc_insert(c, znode, &zbr, n + 1);
+ if (err)
+ goto out_unlock;
+ if (c->replaying) {
+ /*
+ * We did not find it in the index so there may be a
+ * dangling branch still in the index. So we remove it
+ * by passing 'ubifs_tnc_remove_nm()' the same key but
+ * an unmatchable name.
+ */
+ struct qstr noname = { .len = 0, .name = "" };
+
+ err = dbg_check_tnc(c, 0);
+ mutex_unlock(&c->tnc_mutex);
+ if (err)
+ return err;
+ return ubifs_tnc_remove_nm(c, key, &noname);
+ }
+ }
+
+out_unlock:
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * tnc_delete - delete a znode form TNC.
+ * @c: UBIFS file-system description object
+ * @znode: znode to delete from
+ * @n: zbranch slot number to delete
+ *
+ * This function deletes a leaf node from @n-th slot of @znode. Returns zero in
+ * case of success and a negative error code in case of failure.
+ */
+static int tnc_delete(struct ubifs_info *c, struct ubifs_znode *znode, int n)
+{
+ struct ubifs_zbranch *zbr;
+ struct ubifs_znode *zp;
+ int i, err;
+
+ /* Delete without merge for now */
+ ubifs_assert(znode->level == 0);
+ ubifs_assert(n >= 0 && n < c->fanout);
+ dbg_tnc("deleting %s", DBGKEY(&znode->zbranch[n].key));
+
+ zbr = &znode->zbranch[n];
+ lnc_free(zbr);
+
+ err = ubifs_add_dirt(c, zbr->lnum, zbr->len);
+ if (err) {
+ dbg_dump_znode(c, znode);
+ return err;
+ }
+
+ /* We do not "gap" zbranch slots */
+ for (i = n; i < znode->child_cnt - 1; i++)
+ znode->zbranch[i] = znode->zbranch[i + 1];
+ znode->child_cnt -= 1;
+
+ if (znode->child_cnt > 0)
+ return 0;
+
+ /*
+ * This was the last zbranch, we have to delete this znode from the
+ * parent.
+ */
+
+ do {
+ ubifs_assert(!test_bit(OBSOLETE_ZNODE, &znode->flags));
+ ubifs_assert(ubifs_zn_dirty(znode));
+
+ zp = znode->parent;
+ n = znode->iip;
+
+ atomic_long_dec(&c->dirty_zn_cnt);
+
+ err = insert_old_idx_znode(c, znode);
+ if (err)
+ return err;
+
+ if (znode->cnext) {
+ __set_bit(OBSOLETE_ZNODE, &znode->flags);
+ atomic_long_inc(&c->clean_zn_cnt);
+ atomic_long_inc(&ubifs_clean_zn_cnt);
+ } else
+ kfree(znode);
+ znode = zp;
+ } while (znode->child_cnt == 1); /* while removing last child */
+
+ /* Remove from znode, entry n - 1 */
+ znode->child_cnt -= 1;
+ ubifs_assert(znode->level != 0);
+ for (i = n; i < znode->child_cnt; i++) {
+ znode->zbranch[i] = znode->zbranch[i + 1];
+ if (znode->zbranch[i].znode)
+ znode->zbranch[i].znode->iip = i;
+ }
+
+ /*
+ * If this is the root and it has only 1 child then
+ * collapse the tree.
+ */
+ if (!znode->parent) {
+ while (znode->child_cnt == 1 && znode->level != 0) {
+ zp = znode;
+ zbr = &znode->zbranch[0];
+ znode = get_znode(c, znode, 0);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ znode = dirty_cow_znode(c, zbr);
+ if (IS_ERR(znode))
+ return PTR_ERR(znode);
+ znode->parent = NULL;
+ znode->iip = 0;
+ if (c->zroot.len) {
+ err = insert_old_idx(c, c->zroot.lnum,
+ c->zroot.offs);
+ if (err)
+ return err;
+ }
+ c->zroot.lnum = zbr->lnum;
+ c->zroot.offs = zbr->offs;
+ c->zroot.len = zbr->len;
+ c->zroot.znode = znode;
+ ubifs_assert(!test_bit(OBSOLETE_ZNODE,
+ &zp->flags));
+ ubifs_assert(test_bit(DIRTY_ZNODE, &zp->flags));
+ atomic_long_dec(&c->dirty_zn_cnt);
+
+ if (zp->cnext) {
+ __set_bit(OBSOLETE_ZNODE, &zp->flags);
+ atomic_long_inc(&c->clean_zn_cnt);
+ atomic_long_inc(&ubifs_clean_zn_cnt);
+ } else
+ kfree(zp);
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_tnc_remove - remove an index entry of a node.
+ * @c: UBIFS file-system description object
+ * @key: key of node
+ *
+ * Returns %0 on success or negative error code on failure.
+ */
+int ubifs_tnc_remove(struct ubifs_info *c, const union ubifs_key *key)
+{
+ int found, n, err = 0;
+ struct ubifs_znode *znode;
+
+ mutex_lock(&c->tnc_mutex);
+ dbg_tnc("key %s", DBGKEY(key));
+ found = lookup_level0_dirty(c, key, &znode, &n);
+ if (found < 0) {
+ err = found;
+ goto out_unlock;
+ }
+ if (found == 1)
+ err = tnc_delete(c, znode, n);
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+
+out_unlock:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * ubifs_tnc_remove_nm - remove an index entry for a "hashed" node.
+ * @c: UBIFS file-system description object
+ * @key: key of node
+ * @nm: directory entry name
+ *
+ * Returns %0 on success or negative error code on failure.
+ */
+int ubifs_tnc_remove_nm(struct ubifs_info *c, const union ubifs_key *key,
+ const struct qstr *nm)
+{
+ int n, err;
+ struct ubifs_znode *znode;
+
+ mutex_lock(&c->tnc_mutex);
+ dbg_tnc("%.*s, key %s", nm->len, nm->name, DBGKEY(key));
+ err = lookup_level0_dirty(c, key, &znode, &n);
+ if (err < 0)
+ goto out_unlock;
+
+ if (err) {
+ if (c->replaying)
+ err = fallible_resolve_collision(c, key, &znode, &n,
+ nm, 0);
+ else
+ err = resolve_collision(c, key, &znode, &n, nm);
+ dbg_tnc("rc returned %d, znode %p, n %d", err, znode, n);
+ if (err < 0)
+ goto out_unlock;
+ if (err) {
+ /* Ensure the znode is dirtied */
+ if (znode->cnext || !ubifs_zn_dirty(znode)) {
+ znode = dirty_cow_bottom_up(c, znode);
+ if (IS_ERR(znode)) {
+ err = PTR_ERR(znode);
+ goto out_unlock;
+ }
+ }
+ err = tnc_delete(c, znode, n);
+ }
+ }
+
+out_unlock:
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * key_in_range - determine if a key falls within a range of keys.
+ * @c: UBIFS file-system description object
+ * @key: key to check
+ * @from_key: lowest key in range
+ * @to_key: highest key in range
+ *
+ * This function returns %1 if the key is in range and %0 otherwise.
+ */
+static int key_in_range(struct ubifs_info *c, union ubifs_key *key,
+ union ubifs_key *from_key, union ubifs_key *to_key)
+{
+ if (keys_cmp(c, key, from_key) < 0)
+ return 0;
+ if (keys_cmp(c, key, to_key) > 0)
+ return 0;
+ return 1;
+}
+
+/**
+ * ubifs_tnc_remove_range - remove index entries in range.
+ * @c: UBIFS file-system description object
+ * @from_key: lowest key to remove
+ * @to_key: highest key to remove
+ *
+ * This function removes index entries starting at @from_key and ending at
+ * @to_key. This function returns zero in case of success and a negative error
+ * code in case of failure.
+ */
+int ubifs_tnc_remove_range(struct ubifs_info *c, union ubifs_key *from_key,
+ union ubifs_key *to_key)
+{
+ int i, n, k, err = 0;
+ struct ubifs_znode *znode;
+ union ubifs_key *key;
+
+ mutex_lock(&c->tnc_mutex);
+ while (1) {
+ /* Find first level 0 znode that contains keys to remove */
+ err = ubifs_lookup_level0(c, from_key, &znode, &n);
+ if (err < 0)
+ goto out_unlock;
+
+ if (err)
+ key = from_key;
+ else {
+ err = tnc_next(c, &znode, &n);
+ if (err == -ENOENT) {
+ err = 0;
+ goto out_unlock;
+ }
+ if (err < 0)
+ goto out_unlock;
+ key = &znode->zbranch[n].key;
+ if (!key_in_range(c, key, from_key, to_key)) {
+ err = 0;
+ goto out_unlock;
+ }
+ }
+
+ /* Ensure the znode is dirtied */
+ if (znode->cnext || !ubifs_zn_dirty(znode)) {
+ znode = dirty_cow_bottom_up(c, znode);
+ if (IS_ERR(znode)) {
+ err = PTR_ERR(znode);
+ goto out_unlock;
+ }
+ }
+
+ /* Remove all keys in range except the first */
+ for (i = n + 1, k = 0; i < znode->child_cnt; i++, k++) {
+ key = &znode->zbranch[i].key;
+ if (!key_in_range(c, key, from_key, to_key))
+ break;
+ lnc_free(&znode->zbranch[i]);
+ err = ubifs_add_dirt(c, znode->zbranch[i].lnum,
+ znode->zbranch[i].len);
+ if (err) {
+ dbg_dump_znode(c, znode);
+ goto out_unlock;
+ }
+ dbg_tnc("removing %s", DBGKEY(key));
+ }
+ if (k) {
+ for (i = n + 1 + k; i < znode->child_cnt; i++)
+ znode->zbranch[i - k] = znode->zbranch[i];
+ znode->child_cnt -= k;
+ }
+
+ /* Now delete the first */
+ err = tnc_delete(c, znode, n);
+ if (err)
+ goto out_unlock;
+ }
+
+out_unlock:
+ if (!err)
+ err = dbg_check_tnc(c, 0);
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * ubifs_tnc_remove_ino - remove an inode from TNC.
+ * @c: UBIFS file-system description object
+ * @inum: inode number to remove
+ *
+ * This function remove inode @inum and all the extended attributes associated
+ * with the anode from TNC and returns zero in case of success or a negative
+ * error code in case of failure.
+ */
+int ubifs_tnc_remove_ino(struct ubifs_info *c, ino_t inum)
+{
+ union ubifs_key key1, key2;
+ struct ubifs_dent_node *xent, *pxent = NULL;
+ struct qstr nm = { .name = NULL };
+
+ dbg_tnc("ino %lu", (unsigned long)inum);
+
+ /*
+ * Walk all extended attribute entries and remove them together with
+ * corresponding extended attribute inodes.
+ */
+ lowest_xent_key(c, &key1, inum);
+ while (1) {
+ ino_t xattr_inum;
+ int err;
+
+ xent = ubifs_tnc_next_ent(c, &key1, &nm);
+ if (IS_ERR(xent)) {
+ err = PTR_ERR(xent);
+ if (err == -ENOENT)
+ break;
+ return err;
+ }
+
+ xattr_inum = le64_to_cpu(xent->inum);
+ dbg_tnc("xent '%s', ino %lu", xent->name,
+ (unsigned long)xattr_inum);
+
+ nm.name = (char *)xent->name;
+ nm.len = le16_to_cpu(xent->nlen);
+ err = ubifs_tnc_remove_nm(c, &key1, &nm);
+ if (err) {
+ kfree(xent);
+ return err;
+ }
+
+ lowest_ino_key(c, &key1, xattr_inum);
+ highest_ino_key(c, &key2, xattr_inum);
+ err = ubifs_tnc_remove_range(c, &key1, &key2);
+ if (err) {
+ kfree(xent);
+ return err;
+ }
+
+ kfree(pxent);
+ pxent = xent;
+ key_read(c, &xent->key, &key1);
+ }
+
+ kfree(pxent);
+ lowest_ino_key(c, &key1, inum);
+ highest_ino_key(c, &key2, inum);
+
+ return ubifs_tnc_remove_range(c, &key1, &key2);
+}
+
+/**
+ * ubifs_tnc_next_ent - walk directory or extended attribute entries.
+ * @c: UBIFS file-system description object
+ * @key: key of last entry
+ * @nm: name of last entry found or %NULL
+ *
+ * This function finds and reads the next directory or extended attribute entry
+ * after the given key (@key) if there is one. @nm is used to resolve
+ * collisions.
+ *
+ * If the name of the current entry is not known and only the key is known,
+ * @nm->name has to be %NULL. In this case the semantics of this function is a
+ * little bit different and it returns the entry corresponding to this key, not
+ * the next one. If the key was not found, the closest "right" entry is
+ * returned.
+ *
+ * If the fist entry has to be found, @key has to contain the lowest possible
+ * key value for this inode and @name has to be %NULL.
+ *
+ * This function returns the found directory or extended attribute entry node
+ * in case of success, %-ENOENT is returned if no entry was found, and a
+ * negative error code is returned in case of failure.
+ */
+struct ubifs_dent_node *ubifs_tnc_next_ent(struct ubifs_info *c,
+ union ubifs_key *key,
+ const struct qstr *nm)
+{
+ int n, err, type = key_type(c, key);
+ struct ubifs_znode *znode;
+ struct ubifs_dent_node *dent;
+ struct ubifs_zbranch *zbr;
+ union ubifs_key *dkey;
+
+ dbg_tnc("%s %s", nm->name ? (char *)nm->name : "(lowest)", DBGKEY(key));
+ ubifs_assert(is_hash_key(c, key));
+
+ mutex_lock(&c->tnc_mutex);
+ err = ubifs_lookup_level0(c, key, &znode, &n);
+ if (unlikely(err < 0))
+ goto out_unlock;
+
+ if (nm->name) {
+ if (err) {
+ /* Handle collisions */
+ err = resolve_collision(c, key, &znode, &n, nm);
+ dbg_tnc("rc returned %d, znode %p, n %d",
+ err, znode, n);
+ if (unlikely(err < 0))
+ goto out_unlock;
+ }
+
+ /* Now find next entry */
+ err = tnc_next(c, &znode, &n);
+ if (unlikely(err))
+ goto out_unlock;
+ } else {
+ /*
+ * The full name of the entry was not given, in which case the
+ * behavior of this function is a little different and it
+ * returns current entry, not the next one.
+ */
+ if (!err) {
+ /*
+ * However, the given key does not exist in the TNC
+ * tree and @znode/@n variables contain the closest
+ * "preceding" element. Switch to the next one.
+ */
+ err = tnc_next(c, &znode, &n);
+ if (err)
+ goto out_unlock;
+ }
+ }
+
+ zbr = &znode->zbranch[n];
+ dent = kmalloc(zbr->len, GFP_NOFS);
+ if (unlikely(!dent)) {
+ err = -ENOMEM;
+ goto out_unlock;
+ }
+
+ /*
+ * The above 'tnc_next()' call could lead us to the next inode, check
+ * this.
+ */
+ dkey = &zbr->key;
+ if (key_inum(c, dkey) != key_inum(c, key) ||
+ key_type(c, dkey) != type) {
+ err = -ENOENT;
+ goto out_free;
+ }
+
+ err = tnc_read_node_nm(c, zbr, dent);
+ if (unlikely(err))
+ goto out_free;
+
+ mutex_unlock(&c->tnc_mutex);
+ return dent;
+
+out_free:
+ kfree(dent);
+out_unlock:
+ mutex_unlock(&c->tnc_mutex);
+ return ERR_PTR(err);
+}
diff --git a/fs/ubifs/tnc_commit.c b/fs/ubifs/tnc_commit.c
new file mode 100644
index 0000000..8ac76b1
--- /dev/null
+++ b/fs/ubifs/tnc_commit.c
@@ -0,0 +1,1102 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/* This file implements TNC functions for committing */
+
+#include "ubifs.h"
+
+/**
+ * make_idx_node - make an index node for fill-the-gaps method of TNC commit.
+ * @c: UBIFS file-system description object
+ * @idx: buffer in which to place new index node
+ * @znode: znode from which to make new index node
+ * @lnum: LEB number where new index node will be written
+ * @offs: offset where new index node will be written
+ * @len: length of new index node
+ */
+static int make_idx_node(struct ubifs_info *c, struct ubifs_idx_node *idx,
+ struct ubifs_znode *znode, int lnum, int offs, int len)
+{
+ struct ubifs_znode *zp;
+ int i, err;
+
+ /* Make index node */
+ idx->ch.node_type = UBIFS_IDX_NODE;
+ idx->child_cnt = cpu_to_le16(znode->child_cnt);
+ idx->level = cpu_to_le16(znode->level);
+ for (i = 0; i < znode->child_cnt; i++) {
+ struct ubifs_branch *br = ubifs_idx_branch(c, idx, i);
+ struct ubifs_zbranch *zbr = &znode->zbranch[i];
+
+ key_write_idx(c, &zbr->key, &br->key);
+ br->lnum = cpu_to_le32(zbr->lnum);
+ br->offs = cpu_to_le32(zbr->offs);
+ br->len = cpu_to_le32(zbr->len);
+ if (!zbr->lnum || !zbr->len) {
+ ubifs_err("bad ref in znode");
+ dbg_dump_znode(c, znode);
+ if (zbr->znode)
+ dbg_dump_znode(c, zbr->znode);
+ }
+ }
+ ubifs_prepare_node(c, idx, len, 0);
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ znode->lnum = lnum;
+ znode->offs = offs;
+ znode->len = len;
+#endif
+
+ err = insert_old_idx_znode(c, znode);
+
+ /* Update the parent */
+ zp = znode->parent;
+ if (zp) {
+ struct ubifs_zbranch *zbr;
+
+ zbr = &zp->zbranch[znode->iip];
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ } else {
+ c->zroot.lnum = lnum;
+ c->zroot.offs = offs;
+ c->zroot.len = len;
+ }
+ c->calc_idx_sz += ALIGN(len, 8);
+
+ atomic_long_dec(&c->dirty_zn_cnt);
+
+ ubifs_assert(ubifs_zn_dirty(znode));
+ ubifs_assert(test_bit(COW_ZNODE, &znode->flags));
+
+ __clear_bit(DIRTY_ZNODE, &znode->flags);
+ __clear_bit(COW_ZNODE, &znode->flags);
+
+ return err;
+}
+
+/**
+ * fill_gap - make index nodes in gaps in dirty index LEBs.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number that gap appears in
+ * @gap_start: offset of start of gap
+ * @gap_end: offset of end of gap
+ * @dirt: adds dirty space to this
+ *
+ * This function returns the number of index nodes written into the gap.
+ */
+static int fill_gap(struct ubifs_info *c, int lnum, int gap_start, int gap_end,
+ int *dirt)
+{
+ int len, gap_remains, gap_pos, written, pad_len;
+
+ ubifs_assert((gap_start & 7) == 0);
+ ubifs_assert((gap_end & 7) == 0);
+ ubifs_assert(gap_end >= gap_start);
+
+ gap_remains = gap_end - gap_start;
+ if (!gap_remains)
+ return 0;
+ gap_pos = gap_start;
+ written = 0;
+ while (c->enext) {
+ len = ubifs_idx_node_sz(c, c->enext->child_cnt);
+ if (len < gap_remains) {
+ struct ubifs_znode *znode = c->enext;
+ const int alen = ALIGN(len, 8);
+ int err;
+
+ ubifs_assert(alen <= gap_remains);
+ err = make_idx_node(c, c->ileb_buf + gap_pos, znode,
+ lnum, gap_pos, len);
+ if (err)
+ return err;
+ gap_remains -= alen;
+ gap_pos += alen;
+ c->enext = znode->cnext;
+ if (c->enext == c->cnext)
+ c->enext = NULL;
+ written += 1;
+ } else
+ break;
+ }
+ if (gap_end == c->leb_size) {
+ c->ileb_len = ALIGN(gap_pos, c->min_io_size);
+ /* Pad to end of min_io_size */
+ pad_len = c->ileb_len - gap_pos;
+ } else
+ /* Pad to end of gap */
+ pad_len = gap_remains;
+ dbg_gc("LEB %d:%d to %d len %d nodes written %d wasted bytes %d",
+ lnum, gap_start, gap_end, gap_end - gap_start, written, pad_len);
+ ubifs_pad(c, c->ileb_buf + gap_pos, pad_len);
+ *dirt += pad_len;
+ return written;
+}
+
+/**
+ * find_old_idx - find an index node obsoleted since the last commit start.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number of obsoleted index node
+ * @offs: offset of obsoleted index node
+ *
+ * Returns %1 if found and %0 otherwise.
+ */
+static int find_old_idx(struct ubifs_info *c, int lnum, int offs)
+{
+ struct ubifs_old_idx *o;
+ struct rb_node *p;
+
+ p = c->old_idx.rb_node;
+ while (p) {
+ o = rb_entry(p, struct ubifs_old_idx, rb);
+ if (lnum < o->lnum)
+ p = p->rb_left;
+ else if (lnum > o->lnum)
+ p = p->rb_right;
+ else if (offs < o->offs)
+ p = p->rb_left;
+ else if (offs > o->offs)
+ p = p->rb_right;
+ else
+ return 1;
+ }
+ return 0;
+}
+
+/**
+ * is_idx_node_in_use - determine if an index node can be overwritten.
+ * @c: UBIFS file-system description object
+ * @key: key of index node
+ * @level: index node level
+ * @lnum: LEB number of index node
+ * @offs: offset of index node
+ *
+ * If @key / @lnum / @offs identify an index node that was not part of the old
+ * index, then this function returns %0 (obsolete). Else if the index node was
+ * part of the old index but is now dirty %1 is returned, else if it is clean %2
+ * is returned. A negative error code is returned on failure.
+ */
+static int is_idx_node_in_use(struct ubifs_info *c, union ubifs_key *key,
+ int level, int lnum, int offs)
+{
+ int ret;
+
+ ret = is_idx_node_in_tnc(c, key, level, lnum, offs);
+ if (ret < 0)
+ return ret; /* Error code */
+ if (ret == 0)
+ if (find_old_idx(c, lnum, offs))
+ return 1;
+ return ret;
+}
+
+/**
+ * layout_leb_in_gaps - layout index nodes using in-the-gaps method.
+ * @c: UBIFS file-system description object
+ * @p: return LEB number here
+ *
+ * This function lays out new index nodes for dirty znodes using in-the-gaps
+ * method of TNC commit.
+ * This function merely puts the next znode into the next gap, making no attempt
+ * to try to maximise the number of znodes that fit.
+ * This function returns the number of index nodes written into the gaps, or a
+ * negative error code on failure.
+ */
+static int layout_leb_in_gaps(struct ubifs_info *c, int *p)
+{
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ int lnum, dirt = 0, gap_start, gap_end, err, written, tot_written;
+
+ tot_written = 0;
+ /* Get an index LEB with lots of obsolete index nodes */
+ lnum = ubifs_find_dirty_idx_leb(c);
+ if (lnum < 0)
+ /*
+ * There also may be dirt in the index head that could be
+ * filled, however we do not check there at present.
+ */
+ return lnum; /* Error code */
+ *p = lnum;
+ dbg_gc("LEB %d", lnum);
+ /*
+ * Scan the index LEB. We use the generic scan for this even though
+ * it is more comprehensive and less efficient than is needed for this
+ * purpose.
+ */
+ sleb = ubifs_scan(c, lnum, 0, c->ileb_buf);
+ c->ileb_len = 0;
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+ gap_start = 0;
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ struct ubifs_idx_node *idx;
+ int in_use, level;
+
+ ubifs_assert(snod->type == UBIFS_IDX_NODE);
+ idx = snod->node;
+ key_read(c, ubifs_idx_key(c, idx), &snod->key);
+ level = le16_to_cpu(idx->level);
+ /* Determine if the index node is in use (not obsolete) */
+ in_use = is_idx_node_in_use(c, &snod->key, level, lnum,
+ snod->offs);
+ if (in_use < 0) {
+ ubifs_scan_destroy(sleb);
+ return in_use; /* Error code */
+ }
+ if (in_use) {
+ if (in_use == 1)
+ dirt += ALIGN(snod->len, 8);
+ /*
+ * The obsolete index nodes form gaps that can be
+ * overwritten. This gap has ended because we have
+ * found an index node that is still in use
+ * i.e. not obsolete
+ */
+ gap_end = snod->offs;
+ /* Try to fill gap */
+ written = fill_gap(c, lnum, gap_start, gap_end, &dirt);
+ if (written < 0) {
+ ubifs_scan_destroy(sleb);
+ return written; /* Error code */
+ }
+ tot_written += written;
+ gap_start = ALIGN(snod->offs + snod->len, 8);
+ }
+ }
+ ubifs_scan_destroy(sleb);
+ c->ileb_len = c->leb_size;
+ gap_end = c->leb_size;
+ /* Try to fill gap */
+ written = fill_gap(c, lnum, gap_start, gap_end, &dirt);
+ if (written < 0)
+ return written; /* Error code */
+ tot_written += written;
+ if (tot_written == 0) {
+ struct ubifs_lprops lp;
+
+ dbg_gc("LEB %d wrote %d index nodes", lnum, tot_written);
+ err = ubifs_read_one_lp(c, lnum, &lp);
+ if (err)
+ return err;
+ if (lp.free == c->leb_size) {
+ /*
+ * We must have snatched this LEB from the idx_gc list
+ * so we need to correct the free and dirty space.
+ */
+ err = ubifs_change_one_lp(c, lnum,
+ c->leb_size - c->ileb_len,
+ dirt, 0, 0, 0);
+ if (err)
+ return err;
+ }
+ return 0;
+ }
+ err = ubifs_change_one_lp(c, lnum, c->leb_size - c->ileb_len, dirt,
+ 0, 0, 0);
+ if (err)
+ return err;
+ err = ubifs_leb_change(c, lnum, c->ileb_buf, c->ileb_len,
+ UBI_SHORTTERM);
+ if (err)
+ return err;
+ dbg_gc("LEB %d wrote %d index nodes", lnum, tot_written);
+ return tot_written;
+}
+
+/**
+ * get_leb_cnt - calculate the number of empty LEBs needed to commit.
+ * @c: UBIFS file-system description object
+ * @cnt: number of znodes to commit
+ *
+ * This function returns the number of empty LEBs needed to commit @cnt znodes
+ * to the current index head. The number is not exact and may be more than
+ * needed.
+ */
+static int get_leb_cnt(struct ubifs_info *c, int cnt)
+{
+ int d;
+
+ /* Assume maximum index node size (i.e. overestimate space needed) */
+ cnt -= (c->leb_size - c->ihead_offs) / c->max_idx_node_sz;
+ if (cnt < 0)
+ cnt = 0;
+ d = c->leb_size / c->max_idx_node_sz;
+ return DIV_ROUND_UP(cnt, d);
+}
+
+/**
+ * layout_in_gaps - in-the-gaps method of committing TNC.
+ * @c: UBIFS file-system description object
+ * @cnt: number of dirty znodes to commit.
+ *
+ * This function lays out new index nodes for dirty znodes using in-the-gaps
+ * method of TNC commit.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int layout_in_gaps(struct ubifs_info *c, int cnt)
+{
+ int err, leb_needed_cnt, written, *p;
+
+ dbg_gc("%d znodes to write", cnt);
+
+ c->gap_lebs = kmalloc(sizeof(int) * (c->lst.idx_lebs + 1), GFP_NOFS);
+ if (!c->gap_lebs)
+ return -ENOMEM;
+
+ p = c->gap_lebs;
+ do {
+ ubifs_assert(p < c->gap_lebs + sizeof(int) * c->lst.idx_lebs);
+ written = layout_leb_in_gaps(c, p);
+ if (written < 0) {
+ err = written;
+ if (err != -ENOSPC) {
+ kfree(c->gap_lebs);
+ c->gap_lebs = NULL;
+ return err;
+ }
+ if (!dbg_force_in_the_gaps_enabled) {
+ /*
+ * Do not print scary warnings if the debugging
+ * option which forces in-the-gaps is enabled.
+ */
+ ubifs_err("out of space");
+ spin_lock(&c->space_lock);
+ dbg_dump_budg(c);
+ spin_unlock(&c->space_lock);
+ dbg_dump_lprops(c);
+ }
+ /* Try to commit anyway */
+ err = 0;
+ break;
+ }
+ p++;
+ cnt -= written;
+ leb_needed_cnt = get_leb_cnt(c, cnt);
+ dbg_gc("%d znodes remaining, need %d LEBs, have %d", cnt,
+ leb_needed_cnt, c->ileb_cnt);
+ } while (leb_needed_cnt > c->ileb_cnt);
+
+ *p = -1;
+ return 0;
+}
+
+/**
+ * layout_in_empty_space - layout index nodes in empty space.
+ * @c: UBIFS file-system description object
+ *
+ * This function lays out new index nodes for dirty znodes using empty LEBs.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int layout_in_empty_space(struct ubifs_info *c)
+{
+ struct ubifs_znode *znode, *cnext, *zp;
+ int lnum, offs, len, next_len, buf_len, buf_offs, used, avail;
+ int wlen, blen, err;
+
+ cnext = c->enext;
+ if (!cnext)
+ return 0;
+
+ lnum = c->ihead_lnum;
+ buf_offs = c->ihead_offs;
+
+ buf_len = ubifs_idx_node_sz(c, c->fanout);
+ buf_len = ALIGN(buf_len, c->min_io_size);
+ used = 0;
+ avail = buf_len;
+
+ /* Ensure there is enough room for first write */
+ next_len = ubifs_idx_node_sz(c, cnext->child_cnt);
+ if (buf_offs + next_len > c->leb_size)
+ lnum = -1;
+
+ while (1) {
+ znode = cnext;
+
+ len = ubifs_idx_node_sz(c, znode->child_cnt);
+
+ /* Determine the index node position */
+ if (lnum == -1) {
+ if (c->ileb_nxt >= c->ileb_cnt) {
+ ubifs_err("out of space");
+ return -ENOSPC;
+ }
+ lnum = c->ilebs[c->ileb_nxt++];
+ buf_offs = 0;
+ used = 0;
+ avail = buf_len;
+ }
+
+ offs = buf_offs + used;
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ znode->lnum = lnum;
+ znode->offs = offs;
+ znode->len = len;
+#endif
+
+ /* Update the parent */
+ zp = znode->parent;
+ if (zp) {
+ struct ubifs_zbranch *zbr;
+ int i;
+
+ i = znode->iip;
+ zbr = &zp->zbranch[i];
+ zbr->lnum = lnum;
+ zbr->offs = offs;
+ zbr->len = len;
+ } else {
+ c->zroot.lnum = lnum;
+ c->zroot.offs = offs;
+ c->zroot.len = len;
+ }
+ c->calc_idx_sz += ALIGN(len, 8);
+
+ /*
+ * Once lprops is updated, we can decrease the dirty znode count
+ * but it is easier to just do it here.
+ */
+ atomic_long_dec(&c->dirty_zn_cnt);
+
+ /*
+ * Calculate the next index node length to see if there is
+ * enough room for it
+ */
+ cnext = znode->cnext;
+ if (cnext == c->cnext)
+ next_len = 0;
+ else
+ next_len = ubifs_idx_node_sz(c, cnext->child_cnt);
+
+ if (c->min_io_size == 1) {
+ buf_offs += ALIGN(len, 8);
+ if (next_len) {
+ if (buf_offs + next_len <= c->leb_size)
+ continue;
+ err = ubifs_update_one_lp(c, lnum, 0,
+ c->leb_size - buf_offs, 0, 0);
+ if (err)
+ return err;
+ lnum = -1;
+ continue;
+ }
+ err = ubifs_update_one_lp(c, lnum,
+ c->leb_size - buf_offs, 0, 0, 0);
+ if (err)
+ return err;
+ break;
+ }
+
+ /* Update buffer positions */
+ wlen = used + len;
+ used += ALIGN(len, 8);
+ avail -= ALIGN(len, 8);
+
+ if (next_len != 0 &&
+ buf_offs + used + next_len <= c->leb_size &&
+ avail > 0)
+ continue;
+
+ if (avail <= 0 && next_len &&
+ buf_offs + used + next_len <= c->leb_size)
+ blen = buf_len;
+ else
+ blen = ALIGN(wlen, c->min_io_size);
+
+ /* The buffer is full or there are no more znodes to do */
+ buf_offs += blen;
+ if (next_len) {
+ if (buf_offs + next_len > c->leb_size) {
+ err = ubifs_update_one_lp(c, lnum,
+ c->leb_size - buf_offs, blen - used,
+ 0, 0);
+ if (err)
+ return err;
+ lnum = -1;
+ }
+ used -= blen;
+ if (used < 0)
+ used = 0;
+ avail = buf_len - used;
+ continue;
+ }
+ err = ubifs_update_one_lp(c, lnum, c->leb_size - buf_offs,
+ blen - used, 0, 0);
+ if (err)
+ return err;
+ break;
+ }
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ c->new_ihead_lnum = lnum;
+ c->new_ihead_offs = buf_offs;
+#endif
+
+ return 0;
+}
+
+/**
+ * layout_commit - determine positions of index nodes to commit.
+ * @c: UBIFS file-system description object
+ * @no_space: indicates that insufficient empty LEBs were allocated
+ * @cnt: number of znodes to commit
+ *
+ * Calculate and update the positions of index nodes to commit. If there were
+ * an insufficient number of empty LEBs allocated, then index nodes are placed
+ * into the gaps created by obsolete index nodes in non-empty index LEBs. For
+ * this purpose, an obsolete index node is one that was not in the index as at
+ * the end of the last commit. To write "in-the-gaps" requires that those index
+ * LEBs are updated atomically in-place.
+ */
+static int layout_commit(struct ubifs_info *c, int no_space, int cnt)
+{
+ int err;
+
+ if (no_space) {
+ err = layout_in_gaps(c, cnt);
+ if (err)
+ return err;
+ }
+ err = layout_in_empty_space(c);
+ return err;
+}
+
+/**
+ * find_first_dirty - find first dirty znode.
+ * @znode: znode to begin searching from
+ */
+static struct ubifs_znode *find_first_dirty(struct ubifs_znode *znode)
+{
+ int i, cont;
+
+ if (!znode)
+ return NULL;
+
+ while (1) {
+ if (znode->level == 0) {
+ if (ubifs_zn_dirty(znode))
+ return znode;
+ return NULL;
+ }
+ cont = 0;
+ for (i = 0; i < znode->child_cnt; i++) {
+ struct ubifs_zbranch *zbr = &znode->zbranch[i];
+
+ if (zbr->znode && ubifs_zn_dirty(zbr->znode)) {
+ znode = zbr->znode;
+ cont = 1;
+ break;
+ }
+ }
+ if (!cont) {
+ if (ubifs_zn_dirty(znode))
+ return znode;
+ return NULL;
+ }
+ }
+}
+
+/**
+ * find_next_dirty - find next dirty znode.
+ * @znode: znode to begin searching from
+ */
+static struct ubifs_znode *find_next_dirty(struct ubifs_znode *znode)
+{
+ int n = znode->iip + 1;
+
+ znode = znode->parent;
+ if (!znode)
+ return NULL;
+ for (; n < znode->child_cnt; n++) {
+ struct ubifs_zbranch *zbr = &znode->zbranch[n];
+
+ if (zbr->znode && ubifs_zn_dirty(zbr->znode))
+ return find_first_dirty(zbr->znode);
+ }
+ return znode;
+}
+
+/**
+ * get_znodes_to_commit - create list of dirty znodes to commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns the number of znodes to commit.
+ */
+static int get_znodes_to_commit(struct ubifs_info *c)
+{
+ struct ubifs_znode *znode, *cnext;
+ int cnt = 0;
+
+ c->cnext = find_first_dirty(c->zroot.znode);
+ znode = c->enext = c->cnext;
+ if (!znode) {
+ dbg_cmt("no znodes to commit");
+ return 0;
+ }
+ cnt += 1;
+ while (1) {
+ ubifs_assert(!test_bit(COW_ZNODE, &znode->flags));
+ __set_bit(COW_ZNODE, &znode->flags);
+ znode->alt = 0;
+ cnext = find_next_dirty(znode);
+ if (!cnext) {
+ znode->cnext = c->cnext;
+ break;
+ }
+ znode->cnext = cnext;
+ znode = cnext;
+ cnt += 1;
+ }
+ dbg_cmt("committing %d znodes", cnt);
+ ubifs_assert(cnt == atomic_long_read(&c->dirty_zn_cnt));
+ return cnt;
+}
+
+/**
+ * alloc_idx_lebs - allocate empty LEBs to be used to commit.
+ * @c: UBIFS file-system description object
+ * @cnt: number of znodes to commit
+ *
+ * This function returns %-ENOSPC if it cannot allocate a sufficient number of
+ * empty LEBs. %0 is returned on success, otherwise a negative error code
+ * is returned.
+ */
+static int alloc_idx_lebs(struct ubifs_info *c, int cnt)
+{
+ int i, leb_cnt, lnum;
+
+ c->ileb_cnt = 0;
+ c->ileb_nxt = 0;
+ leb_cnt = get_leb_cnt(c, cnt);
+ dbg_cmt("need about %d empty LEBS for TNC commit", leb_cnt);
+ if (!leb_cnt)
+ return 0;
+ c->ilebs = kmalloc(leb_cnt * sizeof(int), GFP_NOFS);
+ if (!c->ilebs)
+ return -ENOMEM;
+ for (i = 0; i < leb_cnt; i++) {
+ lnum = ubifs_find_free_leb_for_idx(c);
+ if (lnum < 0)
+ return lnum;
+ c->ilebs[c->ileb_cnt++] = lnum;
+ dbg_cmt("LEB %d", lnum);
+ }
+ if (dbg_force_in_the_gaps())
+ return -ENOSPC;
+ return 0;
+}
+
+/**
+ * free_unused_idx_lebs - free unused LEBs that were allocated for the commit.
+ * @c: UBIFS file-system description object
+ *
+ * It is possible that we allocate more empty LEBs for the commit than we need.
+ * This functions frees the surplus.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int free_unused_idx_lebs(struct ubifs_info *c)
+{
+ int i, err = 0, lnum, er;
+
+ for (i = c->ileb_nxt; i < c->ileb_cnt; i++) {
+ lnum = c->ilebs[i];
+ dbg_cmt("LEB %d", lnum);
+ er = ubifs_change_one_lp(c, lnum, LPROPS_NC, LPROPS_NC, 0,
+ LPROPS_INDEX | LPROPS_TAKEN, 0);
+ if (!err)
+ err = er;
+ }
+ return err;
+}
+
+/**
+ * free_idx_lebs - free unused LEBs after commit end.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int free_idx_lebs(struct ubifs_info *c)
+{
+ int err;
+
+ err = free_unused_idx_lebs(c);
+ kfree(c->ilebs);
+ c->ilebs = NULL;
+ return err;
+}
+
+/**
+ * ubifs_tnc_start_commit - start TNC commit.
+ * @c: UBIFS file-system description object
+ * @zroot: new index root position is returned here
+ *
+ * This function prepares the list of indexing nodes to commit and lays out
+ * their positions on flash. If there is not enough free space it uses the
+ * in-gap commit method. Returns zero in case of success and a negative error
+ * code in case of failure.
+ */
+int ubifs_tnc_start_commit(struct ubifs_info *c, struct ubifs_zbranch *zroot)
+{
+ int err = 0, cnt;
+
+ mutex_lock(&c->tnc_mutex);
+ err = dbg_check_tnc(c, 1);
+ if (err)
+ goto out;
+ cnt = get_znodes_to_commit(c);
+ if (cnt != 0) {
+ int no_space = 0;
+
+ err = alloc_idx_lebs(c, cnt);
+ if (err == -ENOSPC)
+ no_space = 1;
+ else if (err)
+ goto out_free;
+ err = layout_commit(c, no_space, cnt);
+ if (err)
+ goto out_free;
+ ubifs_assert(atomic_long_read(&c->dirty_zn_cnt) == 0);
+ err = free_unused_idx_lebs(c);
+ if (err)
+ goto out;
+ }
+ destroy_old_idx(c);
+ memcpy(zroot, &c->zroot, sizeof(struct ubifs_zbranch));
+
+ err = ubifs_save_dirty_idx_lnums(c);
+ if (err)
+ goto out;
+
+ spin_lock(&c->space_lock);
+ /*
+ * Although we have not finished committing yet, update size of the
+ * committed index ('c->old_idx_sz') and zero out the index growth
+ * budget. It is OK to do this now, because we've reserved all the
+ * space which is needed to commit the index, and it is save for the
+ * budgeting subsystem to assume the index is already committed,
+ * even though it is not.
+ */
+ c->old_idx_sz = c->calc_idx_sz;
+ c->budg_uncommitted_idx = 0;
+ spin_unlock(&c->space_lock);
+ mutex_unlock(&c->tnc_mutex);
+
+ dbg_cmt("number of index LEBs %d", c->lst.idx_lebs);
+ dbg_cmt("size of index %llu", c->calc_idx_sz);
+ return err;
+
+out_free:
+ free_idx_lebs(c);
+out:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * write_index - write index nodes.
+ * @c: UBIFS file-system description object
+ *
+ * This function writes the index nodes whose positions were laid out in the
+ * layout_in_empty_space function.
+ */
+static int write_index(struct ubifs_info *c)
+{
+ struct ubifs_idx_node *idx;
+ struct ubifs_znode *znode, *cnext;
+ int i, lnum, offs, len, next_len, buf_len, buf_offs, used;
+ int avail, wlen, err, lnum_pos = 0;
+
+ cnext = c->enext;
+ if (!cnext)
+ return 0;
+
+ /*
+ * Always write index nodes to the index head so that index nodes and
+ * other types of nodes are never mixed in the same erase block.
+ */
+ lnum = c->ihead_lnum;
+ buf_offs = c->ihead_offs;
+
+ /* Allocate commit buffer */
+ buf_len = ALIGN(c->max_idx_node_sz, c->min_io_size);
+ used = 0;
+ avail = buf_len;
+
+ /* Ensure there is enough room for first write */
+ next_len = ubifs_idx_node_sz(c, cnext->child_cnt);
+ if (buf_offs + next_len > c->leb_size) {
+ err = ubifs_update_one_lp(c, lnum, LPROPS_NC, 0, 0,
+ LPROPS_TAKEN);
+ if (err)
+ return err;
+ lnum = -1;
+ }
+
+ while (1) {
+ cond_resched();
+
+ znode = cnext;
+ idx = c->cbuf + used;
+
+ /* Make index node */
+ idx->ch.node_type = UBIFS_IDX_NODE;
+ idx->child_cnt = cpu_to_le16(znode->child_cnt);
+ idx->level = cpu_to_le16(znode->level);
+ for (i = 0; i < znode->child_cnt; i++) {
+ struct ubifs_branch *br = ubifs_idx_branch(c, idx, i);
+ struct ubifs_zbranch *zbr = &znode->zbranch[i];
+
+ key_write_idx(c, &zbr->key, &br->key);
+ br->lnum = cpu_to_le32(zbr->lnum);
+ br->offs = cpu_to_le32(zbr->offs);
+ br->len = cpu_to_le32(zbr->len);
+ if (!zbr->lnum || !zbr->len) {
+ ubifs_err("bad ref in znode");
+ dbg_dump_znode(c, znode);
+ if (zbr->znode)
+ dbg_dump_znode(c, zbr->znode);
+ }
+ }
+ len = ubifs_idx_node_sz(c, znode->child_cnt);
+ ubifs_prepare_node(c, idx, len, 0);
+
+ /* Determine the index node position */
+ if (lnum == -1) {
+ lnum = c->ilebs[lnum_pos++];
+ buf_offs = 0;
+ used = 0;
+ avail = buf_len;
+ }
+ offs = buf_offs + used;
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ if (lnum != znode->lnum || offs != znode->offs ||
+ len != znode->len) {
+ ubifs_err("inconsistent znode posn");
+ return -EINVAL;
+ }
+#endif
+
+ /* Grab some stuff from znode while we still can */
+ cnext = znode->cnext;
+
+ ubifs_assert(ubifs_zn_dirty(znode));
+ ubifs_assert(test_bit(COW_ZNODE, &znode->flags));
+
+ /*
+ * It is important that other threads should see %DIRTY_ZNODE
+ * flag cleared before %COW_ZNODE. Specifically, it matters in
+ * the 'dirty_cow_znode()' function. This is the reason for the
+ * first barrier. Also, we want the bit changes to be seen to
+ * other threads ASAP, to avoid unnecesarry copying, which is
+ * the reason for the second barrier.
+ */
+ clear_bit(DIRTY_ZNODE, &znode->flags);
+ smp_mb__before_clear_bit();
+ clear_bit(COW_ZNODE, &znode->flags);
+ smp_mb__after_clear_bit();
+
+ /* Do not access znode from this point on */
+
+ /* Update buffer positions */
+ wlen = used + len;
+ used += ALIGN(len, 8);
+ avail -= ALIGN(len, 8);
+
+ /*
+ * Calculate the next index node length to see if there is
+ * enough room for it
+ */
+ if (cnext == c->cnext)
+ next_len = 0;
+ else
+ next_len = ubifs_idx_node_sz(c, cnext->child_cnt);
+
+ if (c->min_io_size == 1) {
+ /*
+ * Write the prepared index node immediately if there is
+ * no minimum IO size
+ */
+ err = ubifs_leb_write(c, lnum, c->cbuf, buf_offs,
+ wlen, UBI_SHORTTERM);
+ if (err)
+ return err;
+ buf_offs += ALIGN(wlen, 8);
+ if (next_len) {
+ used = 0;
+ avail = buf_len;
+ if (buf_offs + next_len > c->leb_size) {
+ err = ubifs_update_one_lp(c, lnum,
+ LPROPS_NC, 0, 0, LPROPS_TAKEN);
+ if (err)
+ return err;
+ lnum = -1;
+ }
+ continue;
+ }
+ } else {
+ int blen, nxt_offs = buf_offs + used + next_len;
+
+ if (next_len && nxt_offs <= c->leb_size) {
+ if (avail > 0)
+ continue;
+ else
+ blen = buf_len;
+ } else {
+ wlen = ALIGN(wlen, 8);
+ blen = ALIGN(wlen, c->min_io_size);
+ ubifs_pad(c, c->cbuf + wlen, blen - wlen);
+ }
+ /*
+ * The buffer is full or there are no more znodes
+ * to do
+ */
+ err = ubifs_leb_write(c, lnum, c->cbuf, buf_offs,
+ blen, UBI_SHORTTERM);
+ if (err)
+ return err;
+ buf_offs += blen;
+ if (next_len) {
+ if (nxt_offs > c->leb_size) {
+ err = ubifs_update_one_lp(c, lnum,
+ LPROPS_NC, 0, 0, LPROPS_TAKEN);
+ if (err)
+ return err;
+ lnum = -1;
+ }
+ used -= blen;
+ if (used < 0)
+ used = 0;
+ avail = buf_len - used;
+ memmove(c->cbuf, c->cbuf + blen, used);
+ continue;
+ }
+ }
+ break;
+ }
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ if (lnum != c->new_ihead_lnum || buf_offs != c->new_ihead_offs) {
+ ubifs_err("inconsistent ihead");
+ return -EINVAL;
+ }
+#endif
+
+ c->ihead_lnum = lnum;
+ c->ihead_offs = buf_offs;
+
+ return 0;
+}
+
+/**
+ * free_obsolete_znodes - free obsolete znodes.
+ * @c: UBIFS file-system description object
+ *
+ * At the end of commit end, obsolete znodes are freed.
+ */
+static void free_obsolete_znodes(struct ubifs_info *c)
+{
+ struct ubifs_znode *znode, *cnext;
+
+ cnext = c->cnext;
+ do {
+ znode = cnext;
+ cnext = znode->cnext;
+ if (test_bit(OBSOLETE_ZNODE, &znode->flags))
+ kfree(znode);
+ else {
+ znode->cnext = NULL;
+ atomic_long_inc(&c->clean_zn_cnt);
+ atomic_long_inc(&ubifs_clean_zn_cnt);
+ }
+ } while (cnext != c->cnext);
+}
+
+/**
+ * return_gap_lebs - return LEBs used by the in-gap commit method.
+ * @c: UBIFS file-system description object
+ *
+ * This function clears the "taken" flag for the LEBs which were used by the
+ * "commit in-the-gaps" method.
+ */
+static int return_gap_lebs(struct ubifs_info *c)
+{
+ int *p, err;
+
+ if (!c->gap_lebs)
+ return 0;
+
+ dbg_cmt("");
+ for (p = c->gap_lebs; *p != -1; p++) {
+ err = ubifs_change_one_lp(c, *p, LPROPS_NC, LPROPS_NC, 0,
+ LPROPS_TAKEN, 0);
+ if (err)
+ return err;
+ }
+
+ kfree(c->gap_lebs);
+ c->gap_lebs = NULL;
+ return 0;
+}
+
+/**
+ * ubifs_tnc_end_commit - update the TNC for commit end.
+ * @c: UBIFS file-system description object
+ *
+ * Write the dirty znodes.
+ */
+int ubifs_tnc_end_commit(struct ubifs_info *c)
+{
+ int err;
+
+ if (!c->cnext)
+ return 0;
+
+ err = return_gap_lebs(c);
+ if (err)
+ return err;
+
+ err = write_index(c);
+ if (err)
+ return err;
+
+ mutex_lock(&c->tnc_mutex);
+
+ dbg_cmt("TNC height is %d", c->zroot.znode->level + 1);
+
+ free_obsolete_znodes(c);
+
+ c->cnext = NULL;
+ kfree(c->ilebs);
+ c->ilebs = NULL;
+
+ mutex_unlock(&c->tnc_mutex);
+
+ return 0;
+}
diff --git a/fs/ubifs/tnc_misc.c b/fs/ubifs/tnc_misc.c
new file mode 100644
index 0000000..955219f
--- /dev/null
+++ b/fs/ubifs/tnc_misc.c
@@ -0,0 +1,435 @@
+/*
+ * 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: Adrian Hunter
+ * Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file contains miscelanious TNC-related functions shared betweend
+ * different files. This file does not form any logically separate TNC
+ * sub-system. The file was created because there is a lot of TNC code and
+ * putting it all in one file would make that file too big and unreadable.
+ */
+
+#include "ubifs.h"
+
+/**
+ * ubifs_tnc_levelorder_next - next TNC tree element in levelorder traversal.
+ * @zr: root of the subtree to traverse
+ * @znode: previous znode
+ *
+ * This function implements levelorder TNC traversal. The LNC is ignored.
+ * Returns the next element or %NULL if @znode is already the last one.
+ */
+struct ubifs_znode *ubifs_tnc_levelorder_next(struct ubifs_znode *zr,
+ struct ubifs_znode *znode)
+{
+ int level, iip, level_search = 0;
+ struct ubifs_znode *zn;
+
+ ubifs_assert(zr);
+
+ if (unlikely(!znode))
+ return zr;
+
+ if (unlikely(znode == zr)) {
+ if (znode->level == 0)
+ return NULL;
+ return ubifs_tnc_find_child(zr, 0);
+ }
+
+ level = znode->level;
+
+ iip = znode->iip;
+ while (1) {
+ ubifs_assert(znode->level <= zr->level);
+
+ /*
+ * First walk up until there is a znode with next branch to
+ * look at.
+ */
+ while (znode->parent != zr && iip >= znode->parent->child_cnt) {
+ znode = znode->parent;
+ iip = znode->iip;
+ }
+
+ if (unlikely(znode->parent == zr &&
+ iip >= znode->parent->child_cnt)) {
+ /* This level is done, switch to the lower one */
+ level -= 1;
+ if (level_search || level < 0)
+ /*
+ * We were already looking for znode at lower
+ * level ('level_search'). As we are here
+ * again, it just does not exist. Or all levels
+ * were finished ('level < 0').
+ */
+ return NULL;
+
+ level_search = 1;
+ iip = -1;
+ znode = ubifs_tnc_find_child(zr, 0);
+ ubifs_assert(znode);
+ }
+
+ /* Switch to the next index */
+ zn = ubifs_tnc_find_child(znode->parent, iip + 1);
+ if (!zn) {
+ /* No more children to look at, we have walk up */
+ iip = znode->parent->child_cnt;
+ continue;
+ }
+
+ /* Walk back down to the level we came from ('level') */
+ while (zn->level != level) {
+ znode = zn;
+ zn = ubifs_tnc_find_child(zn, 0);
+ if (!zn) {
+ /*
+ * This path is not too deep so it does not
+ * reach 'level'. Try next path.
+ */
+ iip = znode->iip;
+ break;
+ }
+ }
+
+ if (zn) {
+ ubifs_assert(zn->level >= 0);
+ return zn;
+ }
+ }
+}
+
+/**
+ * ubifs_search_zbranch - search znode branch.
+ * @c: UBIFS file-system description object
+ * @znode: znode to search in
+ * @key: key to search for
+ * @n: znode branch slot number is returned here
+ *
+ * This is a helper function which search branch with key @key in @znode using
+ * binary search. The result of the search may be:
+ * o exact match, then %1 is returned, and the slot number of the branch is
+ * stored in @n;
+ * o no exact match, then %0 is returned and the slot number of the left
+ * closest branch is returned in @n; the slot if all keys in this znode are
+ * greater than @key, then %-1 is returned in @n.
+ */
+int ubifs_search_zbranch(const struct ubifs_info *c,
+ const struct ubifs_znode *znode,
+ const union ubifs_key *key, int *n)
+{
+ int beg = 0, end = znode->child_cnt, uninitialized_var(mid);
+ int uninitialized_var(cmp);
+ const struct ubifs_zbranch *zbr = &znode->zbranch[0];
+
+ ubifs_assert(end > beg);
+
+ while (end > beg) {
+ mid = (beg + end) >> 1;
+ cmp = keys_cmp(c, key, &zbr[mid].key);
+ if (cmp > 0)
+ beg = mid + 1;
+ else if (cmp < 0)
+ end = mid;
+ else {
+ *n = mid;
+ return 1;
+ }
+ }
+
+ *n = end - 1;
+
+ /* The insert point is after *n */
+ ubifs_assert(*n >= -1 && *n < znode->child_cnt);
+ if (*n == -1)
+ ubifs_assert(keys_cmp(c, key, &zbr[0].key) < 0);
+ else
+ ubifs_assert(keys_cmp(c, key, &zbr[*n].key) > 0);
+ if (*n + 1 < znode->child_cnt)
+ ubifs_assert(keys_cmp(c, key, &zbr[*n + 1].key) < 0);
+
+ return 0;
+}
+
+/**
+ * ubifs_tnc_postorder_first - find first znode to do postorder tree traversal.
+ * @znode: znode to start at (root of the sub-tree to traverse)
+ *
+ * Find the lowest leftmost znode in a subtree of the TNC tree. The LNC is
+ * ignored.
+ */
+struct ubifs_znode *ubifs_tnc_postorder_first(struct ubifs_znode *znode)
+{
+ if (unlikely(!znode))
+ return NULL;
+
+ while (znode->level > 0) {
+ struct ubifs_znode *child;
+
+ child = ubifs_tnc_find_child(znode, 0);
+ if (!child)
+ return znode;
+ znode = child;
+ }
+
+ return znode;
+}
+
+/**
+ * ubifs_tnc_postorder_next - next TNC tree element in postorder traversal.
+ * @znode: previous znode
+ *
+ * This function implements postorder TNC traversal. The LNC is ignored.
+ * Returns the next element or %NULL if @znode is already the last one.
+ */
+struct ubifs_znode *ubifs_tnc_postorder_next(struct ubifs_znode *znode)
+{
+ struct ubifs_znode *zn;
+
+ ubifs_assert(znode);
+ if (unlikely(!znode->parent))
+ return NULL;
+
+ /* Switch to the next index in the parent */
+ zn = ubifs_tnc_find_child(znode->parent, znode->iip + 1);
+ if (!zn)
+ /* This is in fact the last child, return parent */
+ return znode->parent;
+
+ /* Go to the first znode in this new subtree */
+ return ubifs_tnc_postorder_first(zn);
+}
+
+/**
+ * read_znode - read an indexing node from flash and fill znode.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB of the indexing node to read
+ * @offs: node offset
+ * @len: node length
+ * @znode: znode to read to
+ *
+ * This function reads an indexing node from the flash media and fills znode
+ * with the read data. Returns zero in case of success and a negative error
+ * code in case of failure. The read indexing node is validated and if anything
+ * is wrong with it, this function prints complaint messages and returns
+ * %-EINVAL.
+ */
+static int read_znode(struct ubifs_info *c, int lnum, int offs, int len,
+ struct ubifs_znode *znode)
+{
+ int i, err, type, cmp;
+ struct ubifs_idx_node *idx;
+
+ idx = kmalloc(c->max_idx_node_sz, GFP_NOFS);
+ if (!idx)
+ return -ENOMEM;
+
+ err = ubifs_read_node(c, idx, UBIFS_IDX_NODE, len, lnum, offs);
+ if (err < 0) {
+ kfree(idx);
+ return err;
+ }
+
+ znode->child_cnt = le16_to_cpu(idx->child_cnt);
+ znode->level = le16_to_cpu(idx->level);
+
+ dbg_tnc("LEB %d:%d, level %d, %d branch",
+ lnum, offs, znode->level, znode->child_cnt);
+
+ if (znode->child_cnt > c->fanout || znode->level > UBIFS_MAX_LEVELS) {
+ dbg_err("current fanout %d, branch count %d",
+ c->fanout, znode->child_cnt);
+ dbg_err("max levels %d, znode level %d",
+ UBIFS_MAX_LEVELS, znode->level);
+ err = 1;
+ goto out_dump;
+ }
+
+ for (i = 0; i < znode->child_cnt; i++) {
+ const struct ubifs_branch *br = ubifs_idx_branch(c, idx, i);
+ struct ubifs_zbranch *zbr = &znode->zbranch[i];
+
+ key_read(c, &br->key, &zbr->key);
+ zbr->lnum = le32_to_cpu(br->lnum);
+ zbr->offs = le32_to_cpu(br->offs);
+ zbr->len = le32_to_cpu(br->len);
+ zbr->znode = NULL;
+
+ /* Validate branch */
+
+ if (zbr->lnum < c->main_first ||
+ zbr->lnum >= c->leb_cnt || zbr->offs < 0 ||
+ zbr->offs + zbr->len > c->leb_size || zbr->offs & 7) {
+ dbg_err("bad branch %d", i);
+ err = 2;
+ goto out_dump;
+ }
+
+ switch (key_type(c, &zbr->key)) {
+ case UBIFS_INO_KEY:
+ case UBIFS_DATA_KEY:
+ case UBIFS_DENT_KEY:
+ case UBIFS_XENT_KEY:
+ break;
+ default:
+ dbg_msg("bad key type at slot %d: %s", i,
+ DBGKEY(&zbr->key));
+ err = 3;
+ goto out_dump;
+ }
+
+ if (znode->level)
+ continue;
+
+ type = key_type(c, &zbr->key);
+ if (c->ranges[type].max_len == 0) {
+ if (zbr->len != c->ranges[type].len) {
+ dbg_err("bad target node (type %d) length (%d)",
+ type, zbr->len);
+ dbg_err("have to be %d", c->ranges[type].len);
+ err = 4;
+ goto out_dump;
+ }
+ } else if (zbr->len < c->ranges[type].min_len ||
+ zbr->len > c->ranges[type].max_len) {
+ dbg_err("bad target node (type %d) length (%d)",
+ type, zbr->len);
+ dbg_err("have to be in range of %d-%d",
+ c->ranges[type].min_len,
+ c->ranges[type].max_len);
+ err = 5;
+ goto out_dump;
+ }
+ }
+
+ /*
+ * Ensure that the next key is greater or equivalent to the
+ * previous one.
+ */
+ for (i = 0; i < znode->child_cnt - 1; i++) {
+ const union ubifs_key *key1, *key2;
+
+ key1 = &znode->zbranch[i].key;
+ key2 = &znode->zbranch[i + 1].key;
+
+ cmp = keys_cmp(c, key1, key2);
+ if (cmp > 0) {
+ dbg_err("bad key order (keys %d and %d)", i, i + 1);
+ err = 6;
+ goto out_dump;
+ } else if (cmp == 0 && !is_hash_key(c, key1)) {
+ /* These can only be keys with colliding hash */
+ dbg_err("keys %d and %d are not hashed but equivalent",
+ i, i + 1);
+ err = 7;
+ goto out_dump;
+ }
+ }
+
+ kfree(idx);
+ return 0;
+
+out_dump:
+ ubifs_err("bad indexing node at LEB %d:%d, error %d", lnum, offs, err);
+ dbg_dump_node(c, idx);
+ kfree(idx);
+ return -EINVAL;
+}
+
+/**
+ * ubifs_load_znode - load znode to TNC cache.
+ * @c: UBIFS file-system description object
+ * @zbr: znode branch
+ * @parent: znode's parent
+ * @iip: index in parent
+ *
+ * This function loads znode pointed to by @zbr into the TNC cache and
+ * returns pointer to it in case of success and a negative error code in case
+ * of failure.
+ */
+struct ubifs_znode *ubifs_load_znode(struct ubifs_info *c,
+ struct ubifs_zbranch *zbr,
+ struct ubifs_znode *parent, int iip)
+{
+ int err;
+ struct ubifs_znode *znode;
+
+ ubifs_assert(!zbr->znode);
+ /*
+ * A slab cache is not presently used for znodes because the znode size
+ * depends on the fanout which is stored in the superblock.
+ */
+ znode = kzalloc(c->max_znode_sz, GFP_NOFS);
+ if (!znode)
+ return ERR_PTR(-ENOMEM);
+
+ err = read_znode(c, zbr->lnum, zbr->offs, zbr->len, znode);
+ if (err)
+ goto out;
+
+ zbr->znode = znode;
+ znode->parent = parent;
+ znode->time = get_seconds();
+ znode->iip = iip;
+
+ return znode;
+
+out:
+ kfree(znode);
+ return ERR_PTR(err);
+}
+
+/**
+ * ubifs_tnc_read_node - read a leaf node from the flash media.
+ * @c: UBIFS file-system description object
+ * @zbr: key and position of the node
+ * @node: node is returned here
+ *
+ * This function reads a node defined by @zbr from the flash media. Returns
+ * zero in case of success or a negative negative error code in case of
+ * failure.
+ */
+int ubifs_tnc_read_node(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ void *node)
+{
+ union ubifs_key key1, *key = &zbr->key;
+ int err, type = key_type(c, key);
+
+ err = ubifs_read_node(c, node, type, zbr->len, zbr->lnum, zbr->offs);
+
+ if (err) {
+ dbg_tnc("key %s", DBGKEY(key));
+ return err;
+ }
+
+ /* Make sure the key of the read node is correct */
+ key_read(c, node + UBIFS_KEY_OFFSET, &key1);
+ if (!keys_eq(c, key, &key1)) {
+ ubifs_err("bad key in node at LEB %d:%d",
+ zbr->lnum, zbr->offs);
+ dbg_tnc("looked for key %s found node's key %s",
+ DBGKEY(key), DBGKEY1(&key1));
+ dbg_dump_node(c, node);
+ return -EINVAL;
+ }
+
+ return 0;
+}
diff --git a/fs/ubifs/ubifs-media.h b/fs/ubifs/ubifs-media.h
new file mode 100644
index 0000000..b25fc36
--- /dev/null
+++ b/fs/ubifs/ubifs-media.h
@@ -0,0 +1,751 @@
+/*
+ * 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 describes UBIFS on-flash format and contains definitions of all the
+ * relevant data structures and constants.
+ *
+ * All UBIFS on-flash objects are stored in the form of nodes. All nodes start
+ * with the UBIFS node magic number and have the same common header. Nodes
+ * always sit at 8-byte aligned positions on the media and node header sizes are
+ * also 8-byte aligned (except for the indexing node and the padding node).
+ */
+
+#ifndef __UBIFS_MEDIA_H__
+#define __UBIFS_MEDIA_H__
+
+/* UBIFS node magic number (must not have the padding byte first or last) */
+#define UBIFS_NODE_MAGIC 0x06101831
+
+/* UBIFS on-flash format version */
+#define UBIFS_FORMAT_VERSION 4
+
+/* Minimum logical eraseblock size in bytes */
+#define UBIFS_MIN_LEB_SZ (15*1024)
+
+/* Initial CRC32 value used when calculating CRC checksums */
+#define UBIFS_CRC32_INIT 0xFFFFFFFFU
+
+/*
+ * UBIFS does not try to compress data if its length is less than the below
+ * constant.
+ */
+#define UBIFS_MIN_COMPR_LEN 128
+
+/*
+ * If compressed data length is less than %UBIFS_MIN_COMPRESS_DIFF bytes
+ * shorter than uncompressed data length, UBIFS preferes to leave this data
+ * node uncompress, because it'll be read faster.
+ */
+#define UBIFS_MIN_COMPRESS_DIFF 64
+
+/* Root inode number */
+#define UBIFS_ROOT_INO 1
+
+/* Lowest inode number used for regular inodes (not UBIFS-only internal ones) */
+#define UBIFS_FIRST_INO 64
+
+/*
+ * Maximum file name and extended attribute length (must be a multiple of 8,
+ * minus 1).
+ */
+#define UBIFS_MAX_NLEN 255
+
+/* Maximum number of data journal heads */
+#define UBIFS_MAX_JHEADS 1
+
+/*
+ * Size of UBIFS data block. Note, UBIFS is not a block oriented file-system,
+ * which means that it does not treat the underlying media as consisting of
+ * blocks like in case of hard drives. Do not be confused. UBIFS block is just
+ * the maximum amount of data which one data node can have or which can be
+ * attached to an inode node.
+ */
+#define UBIFS_BLOCK_SIZE 4096
+#define UBIFS_BLOCK_SHIFT 12
+
+/* UBIFS padding byte pattern (must not be first or last byte of node magic) */
+#define UBIFS_PADDING_BYTE 0xCE
+
+/* Maximum possible key length */
+#define UBIFS_MAX_KEY_LEN 16
+
+/* Key length ("simple" format) */
+#define UBIFS_SK_LEN 8
+
+/* Minimum index tree fanout */
+#define UBIFS_MIN_FANOUT 3
+
+/* Maximum number of levels in UBIFS indexing B-tree */
+#define UBIFS_MAX_LEVELS 512
+
+/* Maximum amount of data attached to an inode in bytes */
+#define UBIFS_MAX_INO_DATA UBIFS_BLOCK_SIZE
+
+/* LEB Properties Tree fanout (must be power of 2) and fanout shift */
+#define UBIFS_LPT_FANOUT 4
+#define UBIFS_LPT_FANOUT_SHIFT 2
+
+/* LEB Properties Tree bit field sizes */
+#define UBIFS_LPT_CRC_BITS 16
+#define UBIFS_LPT_CRC_BYTES 2
+#define UBIFS_LPT_TYPE_BITS 4
+
+/* The key is always at the same position in all keyed nodes */
+#define UBIFS_KEY_OFFSET offsetof(struct ubifs_ino_node, key)
+
+/*
+ * LEB Properties Tree node types.
+ *
+ * UBIFS_LPT_PNODE: LPT leaf node (contains LEB properties)
+ * UBIFS_LPT_NNODE: LPT internal node
+ * UBIFS_LPT_LTAB: LPT's own lprops table
+ * UBIFS_LPT_LSAVE: LPT's save table (big model only)
+ * UBIFS_LPT_NODE_CNT: count of LPT node types
+ * UBIFS_LPT_NOT_A_NODE: all ones (15 for 4 bits) is never a valid node type
+ */
+enum {
+ UBIFS_LPT_PNODE,
+ UBIFS_LPT_NNODE,
+ UBIFS_LPT_LTAB,
+ UBIFS_LPT_LSAVE,
+ UBIFS_LPT_NODE_CNT,
+ UBIFS_LPT_NOT_A_NODE = (1 << UBIFS_LPT_TYPE_BITS) - 1,
+};
+
+/*
+ * UBIFS inode types.
+ *
+ * UBIFS_ITYPE_REG: regular file
+ * UBIFS_ITYPE_DIR: directory
+ * UBIFS_ITYPE_LNK: soft link
+ * UBIFS_ITYPE_BLK: block device node
+ * UBIFS_ITYPE_CHR: character device node
+ * UBIFS_ITYPE_FIFO: fifo
+ * UBIFS_ITYPE_SOCK: socket
+ * UBIFS_ITYPES_CNT: count of supported file types
+ */
+enum {
+ UBIFS_ITYPE_REG,
+ UBIFS_ITYPE_DIR,
+ UBIFS_ITYPE_LNK,
+ UBIFS_ITYPE_BLK,
+ UBIFS_ITYPE_CHR,
+ UBIFS_ITYPE_FIFO,
+ UBIFS_ITYPE_SOCK,
+ UBIFS_ITYPES_CNT,
+};
+
+/*
+ * Supported key hash functions.
+ *
+ * UBIFS_KEY_HASH_R5: R5 hash
+ * UBIFS_KEY_HASH_TEST: test hash which just returns first 4 bytes of the name
+ */
+enum {
+ UBIFS_KEY_HASH_R5,
+ UBIFS_KEY_HASH_TEST,
+};
+
+/*
+ * Supported key formats.
+ *
+ * UBIFS_SIMPLE_KEY_FMT: simple key format
+ */
+enum {
+ UBIFS_SIMPLE_KEY_FMT,
+};
+
+/*
+ * The simple key format uses 29 bits for storing UBIFS block number and hash
+ * value.
+ */
+#define UBIFS_S_KEY_BLOCK_BITS 29
+#define UBIFS_S_KEY_BLOCK_MASK 0x1FFFFFFF
+#define UBIFS_S_KEY_HASH_BITS UBIFS_S_KEY_BLOCK_BITS
+#define UBIFS_S_KEY_HASH_MASK UBIFS_S_KEY_BLOCK_MASK
+
+/*
+ * Key types.
+ *
+ * UBIFS_INO_KEY: inode node key
+ * UBIFS_DATA_KEY: data node key
+ * UBIFS_DENT_KEY: directory entry node key
+ * UBIFS_XENT_KEY: extended attribute entry key
+ * UBIFS_KEY_TYPES_CNT: number of supported key types
+ */
+enum {
+ UBIFS_INO_KEY,
+ UBIFS_DATA_KEY,
+ UBIFS_DENT_KEY,
+ UBIFS_XENT_KEY,
+ UBIFS_KEY_TYPES_CNT,
+};
+
+/* Count of LEBs reserved for the superblock area */
+#define UBIFS_SB_LEBS 1
+/* Count of LEBs reserved for the master area */
+#define UBIFS_MST_LEBS 2
+
+/* First LEB of the superblock area */
+#define UBIFS_SB_LNUM 0
+/* First LEB of the master area */
+#define UBIFS_MST_LNUM (UBIFS_SB_LNUM + UBIFS_SB_LEBS)
+/* First LEB of the log area */
+#define UBIFS_LOG_LNUM (UBIFS_MST_LNUM + UBIFS_MST_LEBS)
+
+/*
+ * The below constants define the absolute minimum values for various UBIFS
+ * media areas. Many of them actually depend of flash geometry and the FS
+ * configuration (number of journal heads, orphan LEBs, etc). This means that
+ * the smallest volume size which can be used for UBIFS cannot be pre-defined
+ * by these constants. The file-system that meets the below limitation will not
+ * necessarily mount. UBIFS does run-time calculations and validates the FS
+ * size.
+ */
+
+/* Minimum number of logical eraseblocks in the log */
+#define UBIFS_MIN_LOG_LEBS 2
+/* Minimum number of bud logical eraseblocks (one for each head) */
+#define UBIFS_MIN_BUD_LEBS 3
+/* Minimum number of journal logical eraseblocks */
+#define UBIFS_MIN_JNL_LEBS (UBIFS_MIN_LOG_LEBS + UBIFS_MIN_BUD_LEBS)
+/* Minimum number of LPT area logical eraseblocks */
+#define UBIFS_MIN_LPT_LEBS 2
+/* Minimum number of orphan area logical eraseblocks */
+#define UBIFS_MIN_ORPH_LEBS 1
+/*
+ * Minimum number of main area logical eraseblocks (buds, 3 for the index, 1
+ * for GC, 1 for deletions, and at least 1 for committed data).
+ */
+#define UBIFS_MIN_MAIN_LEBS (UBIFS_MIN_BUD_LEBS + 6)
+
+/* Minimum number of logical eraseblocks */
+#define UBIFS_MIN_LEB_CNT (UBIFS_SB_LEBS + UBIFS_MST_LEBS + \
+ UBIFS_MIN_LOG_LEBS + UBIFS_MIN_LPT_LEBS + \
+ UBIFS_MIN_ORPH_LEBS + UBIFS_MIN_MAIN_LEBS)
+
+/* Node sizes (N.B. these are guaranteed to be multiples of 8) */
+#define UBIFS_CH_SZ sizeof(struct ubifs_ch)
+#define UBIFS_INO_NODE_SZ sizeof(struct ubifs_ino_node)
+#define UBIFS_DATA_NODE_SZ sizeof(struct ubifs_data_node)
+#define UBIFS_DENT_NODE_SZ sizeof(struct ubifs_dent_node)
+#define UBIFS_TRUN_NODE_SZ sizeof(struct ubifs_trun_node)
+#define UBIFS_PAD_NODE_SZ sizeof(struct ubifs_pad_node)
+#define UBIFS_SB_NODE_SZ sizeof(struct ubifs_sb_node)
+#define UBIFS_MST_NODE_SZ sizeof(struct ubifs_mst_node)
+#define UBIFS_REF_NODE_SZ sizeof(struct ubifs_ref_node)
+#define UBIFS_IDX_NODE_SZ sizeof(struct ubifs_idx_node)
+#define UBIFS_CS_NODE_SZ sizeof(struct ubifs_cs_node)
+#define UBIFS_ORPH_NODE_SZ sizeof(struct ubifs_orph_node)
+/* Extended attribute entry nodes are identical to directory entry nodes */
+#define UBIFS_XENT_NODE_SZ UBIFS_DENT_NODE_SZ
+/* Only this does not have to be multiple of 8 bytes */
+#define UBIFS_BRANCH_SZ sizeof(struct ubifs_branch)
+
+/* Maximum node sizes (N.B. these are guaranteed to be multiples of 8) */
+#define UBIFS_MAX_DATA_NODE_SZ (UBIFS_DATA_NODE_SZ + UBIFS_BLOCK_SIZE)
+#define UBIFS_MAX_INO_NODE_SZ (UBIFS_INO_NODE_SZ + UBIFS_MAX_INO_DATA)
+#define UBIFS_MAX_DENT_NODE_SZ (UBIFS_DENT_NODE_SZ + UBIFS_MAX_NLEN + 1)
+#define UBIFS_MAX_XENT_NODE_SZ UBIFS_MAX_DENT_NODE_SZ
+
+/* The largest UBIFS node */
+#define UBIFS_MAX_NODE_SZ UBIFS_MAX_INO_NODE_SZ
+
+/*
+ * On-flash inode flags.
+ *
+ * UBIFS_COMPR_FL: use compression for this inode
+ * UBIFS_SYNC_FL: I/O on this inode has to be synchronous
+ * UBIFS_IMMUTABLE_FL: inode is immutable
+ * UBIFS_APPEND_FL: writes to the inode may only append data
+ * UBIFS_DIRSYNC_FL: I/O on this directory inode has to be synchronous
+ * UBIFS_XATTR_FL: this inode is the inode for an extended attribute value
+ *
+ * Note, these are on-flash flags which correspond to ioctl flags
+ * (@FS_COMPR_FL, etc). They have the same values now, but generally, do not
+ * have to be the same.
+ */
+enum {
+ UBIFS_COMPR_FL = 0x01,
+ UBIFS_SYNC_FL = 0x02,
+ UBIFS_IMMUTABLE_FL = 0x04,
+ UBIFS_APPEND_FL = 0x08,
+ UBIFS_DIRSYNC_FL = 0x10,
+ UBIFS_XATTR_FL = 0x20,
+};
+
+/* Inode flag bits used by UBIFS */
+#define UBIFS_FL_MASK 0x0000001F
+
+/*
+ * UBIFS compression algorithms.
+ *
+ * UBIFS_COMPR_NONE: no compression
+ * UBIFS_COMPR_LZO: LZO compression
+ * UBIFS_COMPR_ZLIB: ZLIB compression
+ * UBIFS_COMPR_TYPES_CNT: count of supported compression types
+ */
+enum {
+ UBIFS_COMPR_NONE,
+ UBIFS_COMPR_LZO,
+ UBIFS_COMPR_ZLIB,
+ UBIFS_COMPR_TYPES_CNT,
+};
+
+/*
+ * UBIFS node types.
+ *
+ * UBIFS_INO_NODE: inode node
+ * UBIFS_DATA_NODE: data node
+ * UBIFS_DENT_NODE: directory entry node
+ * UBIFS_XENT_NODE: extended attribute node
+ * UBIFS_TRUN_NODE: truncation node
+ * UBIFS_PAD_NODE: padding node
+ * UBIFS_SB_NODE: superblock node
+ * UBIFS_MST_NODE: master node
+ * UBIFS_REF_NODE: LEB reference node
+ * UBIFS_IDX_NODE: index node
+ * UBIFS_CS_NODE: commit start node
+ * UBIFS_ORPH_NODE: orphan node
+ * UBIFS_NODE_TYPES_CNT: count of supported node types
+ *
+ * Note, we index arrays by these numbers, so keep them low and contiguous.
+ * Node type constants for inodes, direntries and so on have to be the same as
+ * corresponding key type constants.
+ */
+enum {
+ UBIFS_INO_NODE,
+ UBIFS_DATA_NODE,
+ UBIFS_DENT_NODE,
+ UBIFS_XENT_NODE,
+ UBIFS_TRUN_NODE,
+ UBIFS_PAD_NODE,
+ UBIFS_SB_NODE,
+ UBIFS_MST_NODE,
+ UBIFS_REF_NODE,
+ UBIFS_IDX_NODE,
+ UBIFS_CS_NODE,
+ UBIFS_ORPH_NODE,
+ UBIFS_NODE_TYPES_CNT,
+};
+
+/*
+ * Master node flags.
+ *
+ * UBIFS_MST_DIRTY: rebooted uncleanly - master node is dirty
+ * UBIFS_MST_NO_ORPHS: no orphan inodes present
+ * UBIFS_MST_RCVRY: written by recovery
+ */
+enum {
+ UBIFS_MST_DIRTY = 1,
+ UBIFS_MST_NO_ORPHS = 2,
+ UBIFS_MST_RCVRY = 4,
+};
+
+/*
+ * Node group type (used by recovery to recover whole group or none).
+ *
+ * UBIFS_NO_NODE_GROUP: this node is not part of a group
+ * UBIFS_IN_NODE_GROUP: this node is a part of a group
+ * UBIFS_LAST_OF_NODE_GROUP: this node is the last in a group
+ */
+enum {
+ UBIFS_NO_NODE_GROUP = 0,
+ UBIFS_IN_NODE_GROUP,
+ UBIFS_LAST_OF_NODE_GROUP,
+};
+
+/*
+ * Superblock flags.
+ *
+ * UBIFS_FLG_BIGLPT: if "big" LPT model is used if set
+ */
+enum {
+ UBIFS_FLG_BIGLPT = 0x02,
+};
+
+/**
+ * struct ubifs_ch - common header node.
+ * @magic: UBIFS node magic number (%UBIFS_NODE_MAGIC)
+ * @crc: CRC-32 checksum of the node header
+ * @sqnum: sequence number
+ * @len: full node length
+ * @node_type: node type
+ * @group_type: node group type
+ * @padding: reserved for future, zeroes
+ *
+ * Every UBIFS node starts with this common part. If the node has a key, the
+ * key always goes next.
+ */
+struct ubifs_ch {
+ __le32 magic;
+ __le32 crc;
+ __le64 sqnum;
+ __le32 len;
+ __u8 node_type;
+ __u8 group_type;
+ __u8 padding[2];
+} __attribute__ ((packed));
+
+/**
+ * union ubifs_dev_desc - device node descriptor.
+ * @new: new type device descriptor
+ * @huge: huge type device descriptor
+ *
+ * This data structure describes major/minor numbers of a device node. In an
+ * inode is a device node then its data contains an object of this type. UBIFS
+ * uses standard Linux "new" and "huge" device node encodings.
+ */
+union ubifs_dev_desc {
+ __le32 new;
+ __le64 huge;
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_ino_node - inode node.
+ * @ch: common header
+ * @key: node key
+ * @creat_sqnum: sequence number at time of creation
+ * @size: inode size in bytes (amount of uncompressed data)
+ * @atime_sec: access time seconds
+ * @ctime_sec: creation time seconds
+ * @mtime_sec: modification time seconds
+ * @atime_nsec: access time nanoseconds
+ * @ctime_nsec: creation time nanoseconds
+ * @mtime_nsec: modification time nanoseconds
+ * @nlink: number of hard links
+ * @uid: owner ID
+ * @gid: group ID
+ * @mode: access flags
+ * @flags: per-inode flags (%UBIFS_COMPR_FL, %UBIFS_SYNC_FL, etc)
+ * @data_len: inode data length
+ * @xattr_cnt: count of extended attributes this inode has
+ * @xattr_size: summarized size of all extended attributes in bytes
+ * @padding1: reserved for future, zeroes
+ * @xattr_names: sum of lengths of all extended attribute names belonging to
+ * this inode
+ * @compr_type: compression type used for this inode
+ * @padding2: reserved for future, zeroes
+ * @data: data attached to the inode
+ *
+ * Note, even though inode compression type is defined by @compr_type, some
+ * nodes of this inode may be compressed with different compressor - this
+ * happens if compression type is changed while the inode already has data
+ * nodes. But @compr_type will be use for further writes to the inode.
+ *
+ * Note, do not forget to amend 'zero_ino_node_unused()' function when changing
+ * the padding fields.
+ */
+struct ubifs_ino_node {
+ struct ubifs_ch ch;
+ __u8 key[UBIFS_MAX_KEY_LEN];
+ __le64 creat_sqnum;
+ __le64 size;
+ __le64 atime_sec;
+ __le64 ctime_sec;
+ __le64 mtime_sec;
+ __le32 atime_nsec;
+ __le32 ctime_nsec;
+ __le32 mtime_nsec;
+ __le32 nlink;
+ __le32 uid;
+ __le32 gid;
+ __le32 mode;
+ __le32 flags;
+ __le32 data_len;
+ __le32 xattr_cnt;
+ __le32 xattr_size;
+ __u8 padding1[4]; /* Watch 'zero_ino_node_unused()' if changing! */
+ __le32 xattr_names;
+ __le16 compr_type;
+ __u8 padding2[26]; /* Watch 'zero_ino_node_unused()' if changing! */
+ __u8 data[];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_dent_node - directory entry node.
+ * @ch: common header
+ * @key: node key
+ * @inum: target inode number
+ * @padding1: reserved for future, zeroes
+ * @type: type of the target inode (%UBIFS_ITYPE_REG, %UBIFS_ITYPE_DIR, etc)
+ * @nlen: name length
+ * @padding2: reserved for future, zeroes
+ * @name: zero-terminated name
+ *
+ * Note, do not forget to amend 'zero_dent_node_unused()' function when
+ * changing the padding fields.
+ */
+struct ubifs_dent_node {
+ struct ubifs_ch ch;
+ __u8 key[UBIFS_MAX_KEY_LEN];
+ __le64 inum;
+ __u8 padding1;
+ __u8 type;
+ __le16 nlen;
+ __u8 padding2[4]; /* Watch 'zero_dent_node_unused()' if changing! */
+ __u8 name[];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_data_node - data node.
+ * @ch: common header
+ * @key: node key
+ * @size: uncompressed data size in bytes
+ * @compr_type: compression type (%UBIFS_COMPR_NONE, %UBIFS_COMPR_LZO, etc)
+ * @padding: reserved for future, zeroes
+ * @data: data
+ *
+ * Note, do not forget to amend 'zero_data_node_unused()' function when
+ * changing the padding fields.
+ */
+struct ubifs_data_node {
+ struct ubifs_ch ch;
+ __u8 key[UBIFS_MAX_KEY_LEN];
+ __le32 size;
+ __le16 compr_type;
+ __u8 padding[2]; /* Watch 'zero_data_node_unused()' if changing! */
+ __u8 data[];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_trun_node - truncation node.
+ * @ch: common header
+ * @inum: truncated inode number
+ * @padding: reserved for future, zeroes
+ * @old_size: size before truncation
+ * @new_size: size after truncation
+ *
+ * This node exists only in the journal and never goes to the main area. Note,
+ * do not forget to amend 'zero_trun_node_unused()' function when changing the
+ * padding fields.
+ */
+struct ubifs_trun_node {
+ struct ubifs_ch ch;
+ __le32 inum;
+ __u8 padding[12]; /* Watch 'zero_trun_node_unused()' if changing! */
+ __le64 old_size;
+ __le64 new_size;
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_pad_node - padding node.
+ * @ch: common header
+ * @pad_len: how many bytes after this node are unused (because padded)
+ * @padding: reserved for future, zeroes
+ */
+struct ubifs_pad_node {
+ struct ubifs_ch ch;
+ __le32 pad_len;
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_sb_node - superblock node.
+ * @ch: common header
+ * @padding: reserved for future, zeroes
+ * @key_hash: type of hash function used in keys
+ * @key_fmt: format of the key
+ * @flags: file-system flags (%UBIFS_FLG_BIGLPT, etc)
+ * @min_io_size: minimal input/output unit size
+ * @leb_size: logical eraseblock size in bytes
+ * @leb_cnt: count of LEBs used by file-system
+ * @max_leb_cnt: maximum count of LEBs used by file-system
+ * @max_bud_bytes: maximum amount of data stored in buds
+ * @log_lebs: log size in logical eraseblocks
+ * @lpt_lebs: number of LEBs used for lprops table
+ * @orph_lebs: number of LEBs used for recording orphans
+ * @jhead_cnt: count of journal heads
+ * @fanout: tree fanout (max. number of links per indexing node)
+ * @lsave_cnt: number of LEB numbers in LPT's save table
+ * @fmt_version: UBIFS on-flash format version
+ * @default_compr: default compression algorithm (%UBIFS_COMPR_LZO, etc)
+ * @padding1: reserved for future, zeroes
+ * @rp_uid: reserve pool UID
+ * @rp_gid: reserve pool GID
+ * @rp_size: size of the reserved pool in bytes
+ * @padding2: reserved for future, zeroes
+ * @time_gran: time granularity in nanoseconds
+ * @uuid: UUID generated when the file system image was created
+ */
+struct ubifs_sb_node {
+ struct ubifs_ch ch;
+ __u8 padding[2];
+ __u8 key_hash;
+ __u8 key_fmt;
+ __le32 flags;
+ __le32 min_io_size;
+ __le32 leb_size;
+ __le32 leb_cnt;
+ __le32 max_leb_cnt;
+ __le64 max_bud_bytes;
+ __le32 log_lebs;
+ __le32 lpt_lebs;
+ __le32 orph_lebs;
+ __le32 jhead_cnt;
+ __le32 fanout;
+ __le32 lsave_cnt;
+ __le32 fmt_version;
+ __le16 default_compr;
+ __u8 padding1[2];
+ __le32 rp_uid;
+ __le32 rp_gid;
+ __le64 rp_size;
+ __le32 time_gran;
+ __u8 uuid[16];
+ __u8 padding2[3972];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_mst_node - master node.
+ * @ch: common header
+ * @highest_inum: highest inode number in the committed index
+ * @cmt_no: commit number
+ * @flags: various flags (%UBIFS_MST_DIRTY, etc)
+ * @log_lnum: start of the log
+ * @root_lnum: LEB number of the root indexing node
+ * @root_offs: offset within @root_lnum
+ * @root_len: root indexing node length
+ * @gc_lnum: LEB reserved for garbage collection (%-1 value means the LEB was
+ * not reserved and should be reserved on mount)
+ * @ihead_lnum: LEB number of index head
+ * @ihead_offs: offset of index head
+ * @index_size: size of index on flash
+ * @total_free: total free space in bytes
+ * @total_dirty: total dirty space in bytes
+ * @total_used: total used space in bytes (includes only data LEBs)
+ * @total_dead: total dead space in bytes (includes only data LEBs)
+ * @total_dark: total dark space in bytes (includes only data LEBs)
+ * @lpt_lnum: LEB number of LPT root nnode
+ * @lpt_offs: offset of LPT root nnode
+ * @nhead_lnum: LEB number of LPT head
+ * @nhead_offs: offset of LPT head
+ * @ltab_lnum: LEB number of LPT's own lprops table
+ * @ltab_offs: offset of LPT's own lprops table
+ * @lsave_lnum: LEB number of LPT's save table (big model only)
+ * @lsave_offs: offset of LPT's save table (big model only)
+ * @lscan_lnum: LEB number of last LPT scan
+ * @empty_lebs: number of empty logical eraseblocks
+ * @idx_lebs: number of indexing logical eraseblocks
+ * @leb_cnt: count of LEBs used by file-system
+ * @padding: reserved for future, zeroes
+ */
+struct ubifs_mst_node {
+ struct ubifs_ch ch;
+ __le64 highest_inum;
+ __le64 cmt_no;
+ __le32 flags;
+ __le32 log_lnum;
+ __le32 root_lnum;
+ __le32 root_offs;
+ __le32 root_len;
+ __le32 gc_lnum;
+ __le32 ihead_lnum;
+ __le32 ihead_offs;
+ __le64 index_size;
+ __le64 total_free;
+ __le64 total_dirty;
+ __le64 total_used;
+ __le64 total_dead;
+ __le64 total_dark;
+ __le32 lpt_lnum;
+ __le32 lpt_offs;
+ __le32 nhead_lnum;
+ __le32 nhead_offs;
+ __le32 ltab_lnum;
+ __le32 ltab_offs;
+ __le32 lsave_lnum;
+ __le32 lsave_offs;
+ __le32 lscan_lnum;
+ __le32 empty_lebs;
+ __le32 idx_lebs;
+ __le32 leb_cnt;
+ __u8 padding[344];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_ref_node - logical eraseblock reference node.
+ * @ch: common header
+ * @lnum: the referred logical eraseblock number
+ * @offs: start offset in the referred LEB
+ * @jhead: journal head number
+ * @padding: reserved for future, zeroes
+ */
+struct ubifs_ref_node {
+ struct ubifs_ch ch;
+ __le32 lnum;
+ __le32 offs;
+ __le32 jhead;
+ __u8 padding[28];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_branch - key/reference/length branch
+ * @lnum: LEB number of the target node
+ * @offs: offset within @lnum
+ * @len: target node length
+ * @key: key
+ */
+struct ubifs_branch {
+ __le32 lnum;
+ __le32 offs;
+ __le32 len;
+ __u8 key[];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_idx_node - indexing node.
+ * @ch: common header
+ * @child_cnt: number of child index nodes
+ * @level: tree level
+ * @branches: LEB number / offset / length / key branches
+ */
+struct ubifs_idx_node {
+ struct ubifs_ch ch;
+ __le16 child_cnt;
+ __le16 level;
+ __u8 branches[];
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_cs_node - commit start node.
+ * @ch: common header
+ * @cmt_no: commit number
+ */
+struct ubifs_cs_node {
+ struct ubifs_ch ch;
+ __le64 cmt_no;
+} __attribute__ ((packed));
+
+/**
+ * struct ubifs_orph_node - orphan node.
+ * @ch: common header
+ * @cmt_no: commit number (also top bit is set on the last node of the commit)
+ * @inos: inode numbers of orphans
+ */
+struct ubifs_orph_node {
+ struct ubifs_ch ch;
+ __le64 cmt_no;
+ __le64 inos[];
+} __attribute__ ((packed));
+
+#endif /* __UBIFS_MEDIA_H__ */
diff --git a/fs/ubifs/ubifs.c b/fs/ubifs/ubifs.c
new file mode 100644
index 0000000..d917863
--- /dev/null
+++ b/fs/ubifs/ubifs.c
@@ -0,0 +1,684 @@
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * (C) Copyright 2008-2009
+ * Stefan Roese, DENX Software Engineering, sr@denx.de.
+ *
+ * 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
+ */
+
+#include "ubifs.h"
+
+#if !defined(CONFIG_SYS_64BIT_VSPRINTF)
+#warning Please define CONFIG_SYS_64BIT_VSPRINTF for correct output!
+#endif
+
+DECLARE_GLOBAL_DATA_PTR;
+
+/* compress.c */
+
+/*
+ * We need a wrapper for gunzip() because the parameters are
+ * incompatible with the lzo decompressor.
+ */
+static int gzip_decompress(const unsigned char *in, size_t in_len,
+ unsigned char *out, size_t *out_len)
+{
+ unsigned long len = in_len;
+ return gunzip(out, *out_len, (unsigned char *)in, &len);
+}
+
+/* Fake description object for the "none" compressor */
+static struct ubifs_compressor none_compr = {
+ .compr_type = UBIFS_COMPR_NONE,
+ .name = "no compression",
+ .capi_name = "",
+ .decompress = NULL,
+};
+
+static struct ubifs_compressor lzo_compr = {
+ .compr_type = UBIFS_COMPR_LZO,
+ .name = "LZO",
+ .capi_name = "lzo",
+ .decompress = lzo1x_decompress_safe,
+};
+
+static struct ubifs_compressor zlib_compr = {
+ .compr_type = UBIFS_COMPR_ZLIB,
+ .name = "zlib",
+ .capi_name = "deflate",
+ .decompress = gzip_decompress,
+};
+
+/* All UBIFS compressors */
+struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT];
+
+/**
+ * ubifs_decompress - decompress data.
+ * @in_buf: data to decompress
+ * @in_len: length of the data to decompress
+ * @out_buf: output buffer where decompressed data should
+ * @out_len: output length is returned here
+ * @compr_type: type of compression
+ *
+ * This function decompresses data from buffer @in_buf into buffer @out_buf.
+ * The length of the uncompressed data is returned in @out_len. This functions
+ * returns %0 on success or a negative error code on failure.
+ */
+int ubifs_decompress(const void *in_buf, int in_len, void *out_buf,
+ int *out_len, int compr_type)
+{
+ int err;
+ struct ubifs_compressor *compr;
+
+ if (unlikely(compr_type < 0 || compr_type >= UBIFS_COMPR_TYPES_CNT)) {
+ ubifs_err("invalid compression type %d", compr_type);
+ return -EINVAL;
+ }
+
+ compr = ubifs_compressors[compr_type];
+
+ if (unlikely(!compr->capi_name)) {
+ ubifs_err("%s compression is not compiled in", compr->name);
+ return -EINVAL;
+ }
+
+ if (compr_type == UBIFS_COMPR_NONE) {
+ memcpy(out_buf, in_buf, in_len);
+ *out_len = in_len;
+ return 0;
+ }
+
+ err = compr->decompress(in_buf, in_len, out_buf, (size_t *)out_len);
+ if (err)
+ ubifs_err("cannot decompress %d bytes, compressor %s, "
+ "error %d", in_len, compr->name, err);
+
+ return err;
+}
+
+/**
+ * compr_init - initialize a compressor.
+ * @compr: compressor description object
+ *
+ * This function initializes the requested compressor and returns zero in case
+ * of success or a negative error code in case of failure.
+ */
+static int __init compr_init(struct ubifs_compressor *compr)
+{
+ ubifs_compressors[compr->compr_type] = compr;
+ ubifs_compressors[compr->compr_type]->name += gd->reloc_off;
+ ubifs_compressors[compr->compr_type]->capi_name += gd->reloc_off;
+ ubifs_compressors[compr->compr_type]->decompress += gd->reloc_off;
+ return 0;
+}
+
+/**
+ * ubifs_compressors_init - initialize UBIFS compressors.
+ *
+ * This function initializes the compressor which were compiled in. Returns
+ * zero in case of success and a negative error code in case of failure.
+ */
+int __init ubifs_compressors_init(void)
+{
+ int err;
+
+ err = compr_init(&lzo_compr);
+ if (err)
+ return err;
+
+ err = compr_init(&zlib_compr);
+ if (err)
+ return err;
+
+ ubifs_compressors[UBIFS_COMPR_NONE] = &none_compr;
+ return 0;
+}
+
+/*
+ * ubifsls...
+ */
+
+static int filldir(struct ubifs_info *c, const char *name, int namlen,
+ u64 ino, unsigned int d_type)
+{
+ struct inode *inode;
+ char filetime[32];
+
+ switch (d_type) {
+ case UBIFS_ITYPE_REG:
+ printf("\t");
+ break;
+ case UBIFS_ITYPE_DIR:
+ printf("<DIR>\t");
+ break;
+ case UBIFS_ITYPE_LNK:
+ printf("<LNK>\t");
+ break;
+ default:
+ printf("other\t");
+ break;
+ }
+
+ inode = ubifs_iget(c->vfs_sb, ino);
+ if (IS_ERR(inode)) {
+ printf("%s: Error in ubifs_iget(), ino=%lld ret=%p!\n",
+ __func__, ino, inode);
+ return -1;
+ }
+ ctime_r((time_t *)&inode->i_mtime, filetime);
+ printf("%9lld %24.24s ", inode->i_size, filetime);
+ ubifs_iput(inode);
+
+ printf("%s\n", name);
+
+ return 0;
+}
+
+static int ubifs_printdir(struct file *file, void *dirent)
+{
+ int err, over = 0;
+ struct qstr nm;
+ union ubifs_key key;
+ struct ubifs_dent_node *dent;
+ struct inode *dir = file->f_path.dentry->d_inode;
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+
+ dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, file->f_pos);
+
+ if (file->f_pos > UBIFS_S_KEY_HASH_MASK || file->f_pos == 2)
+ /*
+ * The directory was seek'ed to a senseless position or there
+ * are no more entries.
+ */
+ return 0;
+
+ if (file->f_pos == 1) {
+ /* Find the first entry in TNC and save it */
+ lowest_dent_key(c, &key, dir->i_ino);
+ nm.name = NULL;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ file->f_pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ }
+
+ dent = file->private_data;
+ if (!dent) {
+ /*
+ * The directory was seek'ed to and is now readdir'ed.
+ * Find the entry corresponding to @file->f_pos or the
+ * closest one.
+ */
+ dent_key_init_hash(c, &key, dir->i_ino, file->f_pos);
+ nm.name = NULL;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+ file->f_pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ }
+
+ while (1) {
+ dbg_gen("feed '%s', ino %llu, new f_pos %#x",
+ dent->name, (unsigned long long)le64_to_cpu(dent->inum),
+ key_hash_flash(c, &dent->key));
+ ubifs_assert(le64_to_cpu(dent->ch.sqnum) > ubifs_inode(dir)->creat_sqnum);
+
+ nm.len = le16_to_cpu(dent->nlen);
+ over = filldir(c, (char *)dent->name, nm.len,
+ le64_to_cpu(dent->inum), dent->type);
+ if (over)
+ return 0;
+
+ /* Switch to the next entry */
+ key_read(c, &dent->key, &key);
+ nm.name = (char *)dent->name;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ kfree(file->private_data);
+ file->f_pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ cond_resched();
+ }
+
+out:
+ if (err != -ENOENT) {
+ ubifs_err("cannot find next direntry, error %d", err);
+ return err;
+ }
+
+ kfree(file->private_data);
+ file->private_data = NULL;
+ file->f_pos = 2;
+ return 0;
+}
+
+static int ubifs_finddir(struct super_block *sb, char *dirname,
+ unsigned long root_inum, unsigned long *inum)
+{
+ int err;
+ struct qstr nm;
+ union ubifs_key key;
+ struct ubifs_dent_node *dent;
+ struct ubifs_info *c;
+ struct file *file;
+ struct dentry *dentry;
+ struct inode *dir;
+
+ file = kzalloc(sizeof(struct file), 0);
+ dentry = kzalloc(sizeof(struct dentry), 0);
+ dir = kzalloc(sizeof(struct inode), 0);
+ if (!file || !dentry || !dir) {
+ printf("%s: Error, no memory for malloc!\n", __func__);
+ err = -ENOMEM;
+ goto out;
+ }
+
+ dir->i_sb = sb;
+ file->f_path.dentry = dentry;
+ file->f_path.dentry->d_parent = dentry;
+ file->f_path.dentry->d_inode = dir;
+ file->f_path.dentry->d_inode->i_ino = root_inum;
+ c = sb->s_fs_info;
+
+ dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, file->f_pos);
+
+ /* Find the first entry in TNC and save it */
+ lowest_dent_key(c, &key, dir->i_ino);
+ nm.name = NULL;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ file->f_pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+
+ while (1) {
+ dbg_gen("feed '%s', ino %llu, new f_pos %#x",
+ dent->name, (unsigned long long)le64_to_cpu(dent->inum),
+ key_hash_flash(c, &dent->key));
+ ubifs_assert(le64_to_cpu(dent->ch.sqnum) > ubifs_inode(dir)->creat_sqnum);
+
+ nm.len = le16_to_cpu(dent->nlen);
+ if ((strncmp(dirname, (char *)dent->name, nm.len) == 0) &&
+ (strlen(dirname) == nm.len)) {
+ *inum = le64_to_cpu(dent->inum);
+ return 1;
+ }
+
+ /* Switch to the next entry */
+ key_read(c, &dent->key, &key);
+ nm.name = (char *)dent->name;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ kfree(file->private_data);
+ file->f_pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ cond_resched();
+ }
+
+out:
+ if (err != -ENOENT) {
+ ubifs_err("cannot find next direntry, error %d", err);
+ return err;
+ }
+
+ if (file)
+ free(file);
+ if (dentry)
+ free(dentry);
+ if (dir)
+ free(dir);
+
+ if (file->private_data)
+ kfree(file->private_data);
+ file->private_data = NULL;
+ file->f_pos = 2;
+ return 0;
+}
+
+static unsigned long ubifs_findfile(struct super_block *sb, char *filename)
+{
+ int ret;
+ char *next;
+ char fpath[128];
+ char *name = fpath;
+ unsigned long root_inum = 1;
+ unsigned long inum;
+
+ strcpy(fpath, filename);
+
+ /* Remove all leading slashes */
+ while (*name == '/')
+ name++;
+
+ /*
+ * Handle root-direcoty ('/')
+ */
+ inum = root_inum;
+ if (!name || *name == '\0')
+ return inum;
+
+ for (;;) {
+ /* Extract the actual part from the pathname. */
+ next = strchr(name, '/');
+ if (next) {
+ /* Remove all leading slashes. */
+ while (*next == '/')
+ *(next++) = '\0';
+ }
+
+ ret = ubifs_finddir(sb, name, root_inum, &inum);
+
+ /*
+ * Check if directory with this name exists
+ */
+
+ /* Found the node! */
+ if (!next || *next == '\0') {
+ if (ret)
+ return inum;
+
+ break;
+ }
+
+ root_inum = inum;
+ name = next;
+ }
+
+ return 0;
+}
+
+int ubifs_ls(char *filename)
+{
+ struct ubifs_info *c = ubifs_sb->s_fs_info;
+ struct file *file;
+ struct dentry *dentry;
+ struct inode *dir;
+ void *dirent = NULL;
+ unsigned long inum;
+ int ret = 0;
+
+ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READONLY);
+ inum = ubifs_findfile(ubifs_sb, filename);
+ if (!inum) {
+ ret = -1;
+ goto out;
+ }
+
+ file = kzalloc(sizeof(struct file), 0);
+ dentry = kzalloc(sizeof(struct dentry), 0);
+ dir = kzalloc(sizeof(struct inode), 0);
+ if (!file || !dentry || !dir) {
+ printf("%s: Error, no memory for malloc!\n", __func__);
+ ret = -ENOMEM;
+ goto out_mem;
+ }
+
+ dir->i_sb = ubifs_sb;
+ file->f_path.dentry = dentry;
+ file->f_path.dentry->d_parent = dentry;
+ file->f_path.dentry->d_inode = dir;
+ file->f_path.dentry->d_inode->i_ino = inum;
+ file->f_pos = 1;
+ file->private_data = NULL;
+ ubifs_printdir(file, dirent);
+
+out_mem:
+ if (file)
+ free(file);
+ if (dentry)
+ free(dentry);
+ if (dir)
+ free(dir);
+
+out:
+ ubi_close_volume(c->ubi);
+ return ret;
+}
+
+/*
+ * ubifsload...
+ */
+
+/* file.c */
+
+static inline void *kmap(struct page *page)
+{
+ return page->addr;
+}
+
+static int read_block(struct inode *inode, void *addr, unsigned int block,
+ struct ubifs_data_node *dn)
+{
+ struct ubifs_info *c = inode->i_sb->s_fs_info;
+ int err, len, out_len;
+ union ubifs_key key;
+ unsigned int dlen;
+
+ data_key_init(c, &key, inode->i_ino, block);
+ err = ubifs_tnc_lookup(c, &key, dn);
+ if (err) {
+ if (err == -ENOENT)
+ /* Not found, so it must be a hole */
+ memset(addr, 0, UBIFS_BLOCK_SIZE);
+ return err;
+ }
+
+ ubifs_assert(le64_to_cpu(dn->ch.sqnum) > ubifs_inode(inode)->creat_sqnum);
+
+ len = le32_to_cpu(dn->size);
+ if (len <= 0 || len > UBIFS_BLOCK_SIZE)
+ goto dump;
+
+ dlen = le32_to_cpu(dn->ch.len) - UBIFS_DATA_NODE_SZ;
+ out_len = UBIFS_BLOCK_SIZE;
+ err = ubifs_decompress(&dn->data, dlen, addr, &out_len,
+ le16_to_cpu(dn->compr_type));
+ if (err || len != out_len)
+ goto dump;
+
+ /*
+ * Data length can be less than a full block, even for blocks that are
+ * not the last in the file (e.g., as a result of making a hole and
+ * appending data). Ensure that the remainder is zeroed out.
+ */
+ if (len < UBIFS_BLOCK_SIZE)
+ memset(addr + len, 0, UBIFS_BLOCK_SIZE - len);
+
+ return 0;
+
+dump:
+ ubifs_err("bad data node (block %u, inode %lu)",
+ block, inode->i_ino);
+ dbg_dump_node(c, dn);
+ return -EINVAL;
+}
+
+static int do_readpage(struct ubifs_info *c, struct inode *inode, struct page *page)
+{
+ void *addr;
+ int err = 0, i;
+ unsigned int block, beyond;
+ struct ubifs_data_node *dn;
+ loff_t i_size = inode->i_size;
+
+ dbg_gen("ino %lu, pg %lu, i_size %lld",
+ inode->i_ino, page->index, i_size);
+
+ addr = kmap(page);
+
+ block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT;
+ beyond = (i_size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT;
+ if (block >= beyond) {
+ /* Reading beyond inode */
+ memset(addr, 0, PAGE_CACHE_SIZE);
+ goto out;
+ }
+
+ dn = kmalloc(UBIFS_MAX_DATA_NODE_SZ, GFP_NOFS);
+ if (!dn) {
+ err = -ENOMEM;
+ goto error;
+ }
+
+ i = 0;
+ while (1) {
+ int ret;
+
+ if (block >= beyond) {
+ /* Reading beyond inode */
+ err = -ENOENT;
+ memset(addr, 0, UBIFS_BLOCK_SIZE);
+ } else {
+ ret = read_block(inode, addr, block, dn);
+ if (ret) {
+ err = ret;
+ if (err != -ENOENT)
+ break;
+ } else if (block + 1 == beyond) {
+ int dlen = le32_to_cpu(dn->size);
+ int ilen = i_size & (UBIFS_BLOCK_SIZE - 1);
+
+ if (ilen && ilen < dlen)
+ memset(addr + ilen, 0, dlen - ilen);
+ }
+ }
+ if (++i >= UBIFS_BLOCKS_PER_PAGE)
+ break;
+ block += 1;
+ addr += UBIFS_BLOCK_SIZE;
+ }
+ if (err) {
+ if (err == -ENOENT) {
+ /* Not found, so it must be a hole */
+ dbg_gen("hole");
+ goto out_free;
+ }
+ ubifs_err("cannot read page %lu of inode %lu, error %d",
+ page->index, inode->i_ino, err);
+ goto error;
+ }
+
+out_free:
+ kfree(dn);
+out:
+ return 0;
+
+error:
+ kfree(dn);
+ return err;
+}
+
+int ubifs_load(char *filename, u32 addr, u32 size)
+{
+ struct ubifs_info *c = ubifs_sb->s_fs_info;
+ unsigned long inum;
+ struct inode *inode;
+ struct page page;
+ int err = 0;
+ int i;
+ int count;
+ char link_name[64];
+ struct ubifs_inode *ui;
+
+ c->ubi = ubi_open_volume(c->vi.ubi_num, c->vi.vol_id, UBI_READONLY);
+ inum = ubifs_findfile(ubifs_sb, filename);
+ if (!inum) {
+ err = -1;
+ goto out;
+ }
+
+ /*
+ * Read file inode
+ */
+ inode = ubifs_iget(ubifs_sb, inum);
+ if (IS_ERR(inode)) {
+ printf("%s: Error reading inode %ld!\n", __func__, inum);
+ err = PTR_ERR(inode);
+ goto out;
+ }
+
+ /*
+ * Check for symbolic link
+ */
+ ui = ubifs_inode(inode);
+ if (((inode->i_mode & S_IFMT) == S_IFLNK) && ui->data_len) {
+ memcpy(link_name, ui->data, ui->data_len);
+ printf("%s is linked to %s!\n", filename, link_name);
+ ubifs_iput(inode);
+
+ /*
+ * Now we have the "real" filename, call ubifs_load()
+ * again (recursive call) to load this file instead
+ */
+ return ubifs_load(link_name, addr, size);
+ }
+
+ /*
+ * If no size was specified or if size bigger than filesize
+ * set size to filesize
+ */
+ if ((size == 0) || (size > inode->i_size))
+ size = inode->i_size;
+
+ count = (size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT;
+ printf("Loading file '%s' to addr 0x%08x with size %d (0x%08x)...\n",
+ filename, addr, size, size);
+
+ page.addr = (void *)addr;
+ page.index = 0;
+ page.inode = inode;
+ for (i = 0; i < count; i++) {
+ err = do_readpage(c, inode, &page);
+ if (err)
+ break;
+
+ page.addr += PAGE_SIZE;
+ page.index++;
+ }
+
+ if (err)
+ printf("Error reading file '%s'\n", filename);
+ else
+ printf("Done\n");
+
+ ubifs_iput(inode);
+
+out:
+ ubi_close_volume(c->ubi);
+ return err;
+}
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
new file mode 100644
index 0000000..f342dd8
--- /dev/null
+++ b/fs/ubifs/ubifs.h
@@ -0,0 +1,2173 @@
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation
+ *
+ * (C) Copyright 2008-2009
+ * Stefan Roese, DENX Software Engineering, sr@denx.de.
+ *
+ * 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
+ */
+
+#ifndef __UBIFS_H__
+#define __UBIFS_H__
+
+#if 0 /* Enable for debugging output */
+#define CONFIG_UBIFS_FS_DEBUG
+#define CONFIG_UBIFS_FS_DEBUG_MSG_LVL 3
+#endif
+
+#include <ubi_uboot.h>
+#include <linux/ctype.h>
+#include <linux/time.h>
+#include <linux/math64.h>
+#include "ubifs-media.h"
+
+struct dentry;
+struct file;
+struct iattr;
+struct kstat;
+struct vfsmount;
+
+extern struct super_block *ubifs_sb;
+
+extern unsigned int ubifs_msg_flags;
+extern unsigned int ubifs_chk_flags;
+extern unsigned int ubifs_tst_flags;
+
+#define pgoff_t unsigned long
+
+/*
+ * We "simulate" the Linux page struct much simpler here
+ */
+struct page {
+ pgoff_t index;
+ void *addr;
+ struct inode *inode;
+};
+
+void iput(struct inode *inode);
+
+/*
+ * The atomic operations are used for budgeting etc which is not
+ * needed for the read-only U-Boot implementation:
+ */
+#define atomic_long_inc(a)
+#define atomic_long_dec(a)
+#define atomic_long_sub(a, b)
+
+/* linux/include/time.h */
+
+struct timespec {
+ time_t tv_sec; /* seconds */
+ long tv_nsec; /* nanoseconds */
+};
+
+/* linux/include/dcache.h */
+
+/*
+ * "quick string" -- eases parameter passing, but more importantly
+ * saves "metadata" about the string (ie length and the hash).
+ *
+ * hash comes first so it snuggles against d_parent in the
+ * dentry.
+ */
+struct qstr {
+ unsigned int hash;
+ unsigned int len;
+ const char *name;
+};
+
+struct inode {
+ struct hlist_node i_hash;
+ struct list_head i_list;
+ struct list_head i_sb_list;
+ struct list_head i_dentry;
+ unsigned long i_ino;
+ unsigned int i_nlink;
+ uid_t i_uid;
+ gid_t i_gid;
+ dev_t i_rdev;
+ u64 i_version;
+ loff_t i_size;
+#ifdef __NEED_I_SIZE_ORDERED
+ seqcount_t i_size_seqcount;
+#endif
+ struct timespec i_atime;
+ struct timespec i_mtime;
+ struct timespec i_ctime;
+ unsigned int i_blkbits;
+ unsigned short i_bytes;
+ umode_t i_mode;
+ spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */
+ struct mutex i_mutex;
+ struct rw_semaphore i_alloc_sem;
+ const struct inode_operations *i_op;
+ const struct file_operations *i_fop; /* former ->i_op->default_file_ops */
+ struct super_block *i_sb;
+ struct file_lock *i_flock;
+#ifdef CONFIG_QUOTA
+ struct dquot *i_dquot[MAXQUOTAS];
+#endif
+ struct list_head i_devices;
+ int i_cindex;
+
+ __u32 i_generation;
+
+#ifdef CONFIG_DNOTIFY
+ unsigned long i_dnotify_mask; /* Directory notify events */
+ struct dnotify_struct *i_dnotify; /* for directory notifications */
+#endif
+
+#ifdef CONFIG_INOTIFY
+ struct list_head inotify_watches; /* watches on this inode */
+ struct mutex inotify_mutex; /* protects the watches list */
+#endif
+
+ unsigned long i_state;
+ unsigned long dirtied_when; /* jiffies of first dirtying */
+
+ unsigned int i_flags;
+
+#ifdef CONFIG_SECURITY
+ void *i_security;
+#endif
+ void *i_private; /* fs or device private pointer */
+};
+
+struct super_block {
+ struct list_head s_list; /* Keep this first */
+ dev_t s_dev; /* search index; _not_ kdev_t */
+ unsigned long s_blocksize;
+ unsigned char s_blocksize_bits;
+ unsigned char s_dirt;
+ unsigned long long s_maxbytes; /* Max file size */
+ struct file_system_type *s_type;
+ const struct super_operations *s_op;
+ struct dquot_operations *dq_op;
+ struct quotactl_ops *s_qcop;
+ const struct export_operations *s_export_op;
+ unsigned long s_flags;
+ unsigned long s_magic;
+ struct dentry *s_root;
+ struct rw_semaphore s_umount;
+ struct mutex s_lock;
+ int s_count;
+ int s_syncing;
+ int s_need_sync_fs;
+#ifdef CONFIG_SECURITY
+ void *s_security;
+#endif
+ struct xattr_handler **s_xattr;
+
+ struct list_head s_inodes; /* all inodes */
+ struct list_head s_dirty; /* dirty inodes */
+ struct list_head s_io; /* parked for writeback */
+ struct list_head s_more_io; /* parked for more writeback */
+ struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */
+ struct list_head s_files;
+ /* s_dentry_lru and s_nr_dentry_unused are protected by dcache_lock */
+ struct list_head s_dentry_lru; /* unused dentry lru */
+ int s_nr_dentry_unused; /* # of dentry on lru */
+
+ struct block_device *s_bdev;
+ struct mtd_info *s_mtd;
+ struct list_head s_instances;
+
+ int s_frozen;
+ wait_queue_head_t s_wait_unfrozen;
+
+ char s_id[32]; /* Informational name */
+
+ void *s_fs_info; /* Filesystem private info */
+
+ /*
+ * The next field is for VFS *only*. No filesystems have any business
+ * even looking at it. You had been warned.
+ */
+ struct mutex s_vfs_rename_mutex; /* Kludge */
+
+ /* Granularity of c/m/atime in ns.
+ Cannot be worse than a second */
+ u32 s_time_gran;
+
+ /*
+ * Filesystem subtype. If non-empty the filesystem type field
+ * in /proc/mounts will be "type.subtype"
+ */
+ char *s_subtype;
+
+ /*
+ * Saved mount options for lazy filesystems using
+ * generic_show_options()
+ */
+ char *s_options;
+};
+
+struct file_system_type {
+ const char *name;
+ int fs_flags;
+ int (*get_sb) (struct file_system_type *, int,
+ const char *, void *, struct vfsmount *);
+ void (*kill_sb) (struct super_block *);
+ struct module *owner;
+ struct file_system_type * next;
+ struct list_head fs_supers;
+};
+
+struct vfsmount {
+ struct list_head mnt_hash;
+ struct vfsmount *mnt_parent; /* fs we are mounted on */
+ struct dentry *mnt_mountpoint; /* dentry of mountpoint */
+ struct dentry *mnt_root; /* root of the mounted tree */
+ struct super_block *mnt_sb; /* pointer to superblock */
+ struct list_head mnt_mounts; /* list of children, anchored here */
+ struct list_head mnt_child; /* and going through their mnt_child */
+ int mnt_flags;
+ /* 4 bytes hole on 64bits arches */
+ const char *mnt_devname; /* Name of device e.g. /dev/dsk/hda1 */
+ struct list_head mnt_list;
+ struct list_head mnt_expire; /* link in fs-specific expiry list */
+ struct list_head mnt_share; /* circular list of shared mounts */
+ struct list_head mnt_slave_list;/* list of slave mounts */
+ struct list_head mnt_slave; /* slave list entry */
+ struct vfsmount *mnt_master; /* slave is on master->mnt_slave_list */
+ struct mnt_namespace *mnt_ns; /* containing namespace */
+ int mnt_id; /* mount identifier */
+ int mnt_group_id; /* peer group identifier */
+ /*
+ * We put mnt_count & mnt_expiry_mark at the end of struct vfsmount
+ * to let these frequently modified fields in a separate cache line
+ * (so that reads of mnt_flags wont ping-pong on SMP machines)
+ */
+ int mnt_expiry_mark; /* true if marked for expiry */
+ int mnt_pinned;
+ int mnt_ghosts;
+ /*
+ * This value is not stable unless all of the mnt_writers[] spinlocks
+ * are held, and all mnt_writer[]s on this mount have 0 as their ->count
+ */
+};
+
+struct path {
+ struct vfsmount *mnt;
+ struct dentry *dentry;
+};
+
+struct file {
+ struct path f_path;
+#define f_dentry f_path.dentry
+#define f_vfsmnt f_path.mnt
+ const struct file_operations *f_op;
+ unsigned int f_flags;
+ loff_t f_pos;
+ unsigned int f_uid, f_gid;
+
+ u64 f_version;
+#ifdef CONFIG_SECURITY
+ void *f_security;
+#endif
+ /* needed for tty driver, and maybe others */
+ void *private_data;
+
+#ifdef CONFIG_EPOLL
+ /* Used by fs/eventpoll.c to link all the hooks to this file */
+ struct list_head f_ep_links;
+ spinlock_t f_ep_lock;
+#endif /* #ifdef CONFIG_EPOLL */
+#ifdef CONFIG_DEBUG_WRITECOUNT
+ unsigned long f_mnt_write_state;
+#endif
+};
+
+/*
+ * get_seconds() not really needed in the read-only implmentation
+ */
+#define get_seconds() 0
+
+/* 4k page size */
+#define PAGE_CACHE_SHIFT 12
+#define PAGE_CACHE_SIZE (1 << PAGE_CACHE_SHIFT)
+
+/* Page cache limit. The filesystems should put that into their s_maxbytes
+ limits, otherwise bad things can happen in VM. */
+#if BITS_PER_LONG==32
+#define MAX_LFS_FILESIZE (((u64)PAGE_CACHE_SIZE << (BITS_PER_LONG-1))-1)
+#elif BITS_PER_LONG==64
+#define MAX_LFS_FILESIZE 0x7fffffffffffffffUL
+#endif
+
+#define INT_MAX ((int)(~0U>>1))
+#define INT_MIN (-INT_MAX - 1)
+#define LLONG_MAX ((long long)(~0ULL>>1))
+
+/*
+ * These are the fs-independent mount-flags: up to 32 flags are supported
+ */
+#define MS_RDONLY 1 /* Mount read-only */
+#define MS_NOSUID 2 /* Ignore suid and sgid bits */
+#define MS_NODEV 4 /* Disallow access to device special files */
+#define MS_NOEXEC 8 /* Disallow program execution */
+#define MS_SYNCHRONOUS 16 /* Writes are synced at once */
+#define MS_REMOUNT 32 /* Alter flags of a mounted FS */
+#define MS_MANDLOCK 64 /* Allow mandatory locks on an FS */
+#define MS_DIRSYNC 128 /* Directory modifications are synchronous */
+#define MS_NOATIME 1024 /* Do not update access times. */
+#define MS_NODIRATIME 2048 /* Do not update directory access times */
+#define MS_BIND 4096
+#define MS_MOVE 8192
+#define MS_REC 16384
+#define MS_VERBOSE 32768 /* War is peace. Verbosity is silence.
+ MS_VERBOSE is deprecated. */
+#define MS_SILENT 32768
+#define MS_POSIXACL (1<<16) /* VFS does not apply the umask */
+#define MS_UNBINDABLE (1<<17) /* change to unbindable */
+#define MS_PRIVATE (1<<18) /* change to private */
+#define MS_SLAVE (1<<19) /* change to slave */
+#define MS_SHARED (1<<20) /* change to shared */
+#define MS_RELATIME (1<<21) /* Update atime relative to mtime/ctime. */
+#define MS_KERNMOUNT (1<<22) /* this is a kern_mount call */
+#define MS_I_VERSION (1<<23) /* Update inode I_version field */
+#define MS_ACTIVE (1<<30)
+#define MS_NOUSER (1<<31)
+
+#define I_NEW 8
+
+/* Inode flags - they have nothing to superblock flags now */
+
+#define S_SYNC 1 /* Writes are synced at once */
+#define S_NOATIME 2 /* Do not update access times */
+#define S_APPEND 4 /* Append-only file */
+#define S_IMMUTABLE 8 /* Immutable file */
+#define S_DEAD 16 /* removed, but still open directory */
+#define S_NOQUOTA 32 /* Inode is not counted to quota */
+#define S_DIRSYNC 64 /* Directory modifications are synchronous */
+#define S_NOCMTIME 128 /* Do not update file c/mtime */
+#define S_SWAPFILE 256 /* Do not truncate: swapon got its bmaps */
+#define S_PRIVATE 512 /* Inode is fs-internal */
+
+/* include/linux/stat.h */
+
+#define S_IFMT 00170000
+#define S_IFSOCK 0140000
+#define S_IFLNK 0120000
+#define S_IFREG 0100000
+#define S_IFBLK 0060000
+#define S_IFDIR 0040000
+#define S_IFCHR 0020000
+#define S_IFIFO 0010000
+#define S_ISUID 0004000
+#define S_ISGID 0002000
+#define S_ISVTX 0001000
+
+/* include/linux/fs.h */
+
+/*
+ * File types
+ *
+ * NOTE! These match bits 12..15 of stat.st_mode
+ * (ie "(i_mode >> 12) & 15").
+ */
+#define DT_UNKNOWN 0
+#define DT_FIFO 1
+#define DT_CHR 2
+#define DT_DIR 4
+#define DT_BLK 6
+#define DT_REG 8
+#define DT_LNK 10
+#define DT_SOCK 12
+#define DT_WHT 14
+
+#define I_DIRTY_SYNC 1
+#define I_DIRTY_DATASYNC 2
+#define I_DIRTY_PAGES 4
+#define I_NEW 8
+#define I_WILL_FREE 16
+#define I_FREEING 32
+#define I_CLEAR 64
+#define __I_LOCK 7
+#define I_LOCK (1 << __I_LOCK)
+#define __I_SYNC 8
+#define I_SYNC (1 << __I_SYNC)
+
+#define I_DIRTY (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)
+
+/* linux/include/dcache.h */
+
+#define DNAME_INLINE_LEN_MIN 36
+
+struct dentry {
+ unsigned int d_flags; /* protected by d_lock */
+ spinlock_t d_lock; /* per dentry lock */
+ struct inode *d_inode; /* Where the name belongs to - NULL is
+ * negative */
+ /*
+ * The next three fields are touched by __d_lookup. Place them here
+ * so they all fit in a cache line.
+ */
+ struct hlist_node d_hash; /* lookup hash list */
+ struct dentry *d_parent; /* parent directory */
+ struct qstr d_name;
+
+ struct list_head d_lru; /* LRU list */
+ /*
+ * d_child and d_rcu can share memory
+ */
+ struct list_head d_subdirs; /* our children */
+ struct list_head d_alias; /* inode alias list */
+ unsigned long d_time; /* used by d_revalidate */
+ struct super_block *d_sb; /* The root of the dentry tree */
+ void *d_fsdata; /* fs-specific data */
+#ifdef CONFIG_PROFILING
+ struct dcookie_struct *d_cookie; /* cookie, if any */
+#endif
+ int d_mounted;
+ unsigned char d_iname[DNAME_INLINE_LEN_MIN]; /* small names */
+};
+
+static inline ino_t parent_ino(struct dentry *dentry)
+{
+ ino_t res;
+
+ spin_lock(&dentry->d_lock);
+ res = dentry->d_parent->d_inode->i_ino;
+ spin_unlock(&dentry->d_lock);
+ return res;
+}
+
+/* linux/include/linux/bitops.h */
+
+#define BIT_MASK(nr) (1UL << ((nr) % BITS_PER_LONG))
+#define BIT_WORD(nr) ((nr) / BITS_PER_LONG)
+
+/* linux/include/asm-generic/bitops/non-atomic.h */
+
+/**
+ * __set_bit - Set a bit in memory
+ * @nr: the bit to set
+ * @addr: the address to start counting from
+ *
+ * Unlike set_bit(), this function is non-atomic and may be reordered.
+ * If it's called on the same region of memory simultaneously, the effect
+ * may be that only one operation succeeds.
+ */
+static inline void __set_bit(int nr, volatile unsigned long *addr)
+{
+ unsigned long mask = BIT_MASK(nr);
+ unsigned long *p = ((unsigned long *)addr) + BIT_WORD(nr);
+
+ *p |= mask;
+}
+
+static inline void __clear_bit(int nr, volatile unsigned long *addr)
+{
+ unsigned long mask = BIT_MASK(nr);
+ unsigned long *p = ((unsigned long *)addr) + BIT_WORD(nr);
+
+ *p &= ~mask;
+}
+
+/* debug.c */
+
+#define DEFINE_SPINLOCK(...)
+#define module_param_named(...)
+
+/* misc.h */
+#define mutex_lock_nested(...)
+#define mutex_unlock_nested(...)
+#define mutex_is_locked(...) 0
+
+/* Version of this UBIFS implementation */
+#define UBIFS_VERSION 1
+
+/* Normal UBIFS messages */
+#define ubifs_msg(fmt, ...) \
+ printk(KERN_NOTICE "UBIFS: " fmt "\n", ##__VA_ARGS__)
+/* UBIFS error messages */
+#define ubifs_err(fmt, ...) \
+ printk(KERN_ERR "UBIFS error (pid %d): %s: " fmt "\n", 0, \
+ __func__, ##__VA_ARGS__)
+/* UBIFS warning messages */
+#define ubifs_warn(fmt, ...) \
+ printk(KERN_WARNING "UBIFS warning (pid %d): %s: " fmt "\n", \
+ 0, __func__, ##__VA_ARGS__)
+
+/* UBIFS file system VFS magic number */
+#define UBIFS_SUPER_MAGIC 0x24051905
+
+/* Number of UBIFS blocks per VFS page */
+#define UBIFS_BLOCKS_PER_PAGE (PAGE_CACHE_SIZE / UBIFS_BLOCK_SIZE)
+#define UBIFS_BLOCKS_PER_PAGE_SHIFT (PAGE_CACHE_SHIFT - UBIFS_BLOCK_SHIFT)
+
+/* "File system end of life" sequence number watermark */
+#define SQNUM_WARN_WATERMARK 0xFFFFFFFF00000000ULL
+#define SQNUM_WATERMARK 0xFFFFFFFFFF000000ULL
+
+/*
+ * Minimum amount of LEBs reserved for the index. At present the index needs at
+ * least 2 LEBs: one for the index head and one for in-the-gaps method (which
+ * currently does not cater for the index head and so excludes it from
+ * consideration).
+ */
+#define MIN_INDEX_LEBS 2
+
+/* Minimum amount of data UBIFS writes to the flash */
+#define MIN_WRITE_SZ (UBIFS_DATA_NODE_SZ + 8)
+
+/*
+ * Currently we do not support inode number overlapping and re-using, so this
+ * watermark defines dangerous inode number level. This should be fixed later,
+ * although it is difficult to exceed current limit. Another option is to use
+ * 64-bit inode numbers, but this means more overhead.
+ */
+#define INUM_WARN_WATERMARK 0xFFF00000
+#define INUM_WATERMARK 0xFFFFFF00
+
+/* Largest key size supported in this implementation */
+#define CUR_MAX_KEY_LEN UBIFS_SK_LEN
+
+/* Maximum number of entries in each LPT (LEB category) heap */
+#define LPT_HEAP_SZ 256
+
+/*
+ * Background thread name pattern. The numbers are UBI device and volume
+ * numbers.
+ */
+#define BGT_NAME_PATTERN "ubifs_bgt%d_%d"
+
+/* Default write-buffer synchronization timeout (5 secs) */
+#define DEFAULT_WBUF_TIMEOUT (5 * HZ)
+
+/* Maximum possible inode number (only 32-bit inodes are supported now) */
+#define MAX_INUM 0xFFFFFFFF
+
+/* Number of non-data journal heads */
+#define NONDATA_JHEADS_CNT 2
+
+/* Garbage collector head */
+#define GCHD 0
+/* Base journal head number */
+#define BASEHD 1
+/* First "general purpose" journal head */
+#define DATAHD 2
+
+/* 'No change' value for 'ubifs_change_lp()' */
+#define LPROPS_NC 0x80000001
+
+/*
+ * There is no notion of truncation key because truncation nodes do not exist
+ * in TNC. However, when replaying, it is handy to introduce fake "truncation"
+ * keys for truncation nodes because the code becomes simpler. So we define
+ * %UBIFS_TRUN_KEY type.
+ */
+#define UBIFS_TRUN_KEY UBIFS_KEY_TYPES_CNT
+
+/*
+ * How much a directory entry/extended attribute entry adds to the parent/host
+ * inode.
+ */
+#define CALC_DENT_SIZE(name_len) ALIGN(UBIFS_DENT_NODE_SZ + (name_len) + 1, 8)
+
+/* How much an extended attribute adds to the host inode */
+#define CALC_XATTR_BYTES(data_len) ALIGN(UBIFS_INO_NODE_SZ + (data_len) + 1, 8)
+
+/*
+ * Znodes which were not touched for 'OLD_ZNODE_AGE' seconds are considered
+ * "old", and znode which were touched last 'YOUNG_ZNODE_AGE' seconds ago are
+ * considered "young". This is used by shrinker when selecting znode to trim
+ * off.
+ */
+#define OLD_ZNODE_AGE 20
+#define YOUNG_ZNODE_AGE 5
+
+/*
+ * Some compressors, like LZO, may end up with more data then the input buffer.
+ * So UBIFS always allocates larger output buffer, to be sure the compressor
+ * will not corrupt memory in case of worst case compression.
+ */
+#define WORST_COMPR_FACTOR 2
+
+/* Maximum expected tree height for use by bottom_up_buf */
+#define BOTTOM_UP_HEIGHT 64
+
+/* Maximum number of data nodes to bulk-read */
+#define UBIFS_MAX_BULK_READ 32
+
+/*
+ * Lockdep classes for UBIFS inode @ui_mutex.
+ */
+enum {
+ WB_MUTEX_1 = 0,
+ WB_MUTEX_2 = 1,
+ WB_MUTEX_3 = 2,
+};
+
+/*
+ * Znode flags (actually, bit numbers which store the flags).
+ *
+ * DIRTY_ZNODE: znode is dirty
+ * COW_ZNODE: znode is being committed and a new instance of this znode has to
+ * be created before changing this znode
+ * OBSOLETE_ZNODE: znode is obsolete, which means it was deleted, but it is
+ * still in the commit list and the ongoing commit operation
+ * will commit it, and delete this znode after it is done
+ */
+enum {
+ DIRTY_ZNODE = 0,
+ COW_ZNODE = 1,
+ OBSOLETE_ZNODE = 2,
+};
+
+/*
+ * Commit states.
+ *
+ * COMMIT_RESTING: commit is not wanted
+ * COMMIT_BACKGROUND: background commit has been requested
+ * COMMIT_REQUIRED: commit is required
+ * COMMIT_RUNNING_BACKGROUND: background commit is running
+ * COMMIT_RUNNING_REQUIRED: commit is running and it is required
+ * COMMIT_BROKEN: commit failed
+ */
+enum {
+ COMMIT_RESTING = 0,
+ COMMIT_BACKGROUND,
+ COMMIT_REQUIRED,
+ COMMIT_RUNNING_BACKGROUND,
+ COMMIT_RUNNING_REQUIRED,
+ COMMIT_BROKEN,
+};
+
+/*
+ * 'ubifs_scan_a_node()' return values.
+ *
+ * SCANNED_GARBAGE: scanned garbage
+ * SCANNED_EMPTY_SPACE: scanned empty space
+ * SCANNED_A_NODE: scanned a valid node
+ * SCANNED_A_CORRUPT_NODE: scanned a corrupted node
+ * SCANNED_A_BAD_PAD_NODE: scanned a padding node with invalid pad length
+ *
+ * Greater than zero means: 'scanned that number of padding bytes'
+ */
+enum {
+ SCANNED_GARBAGE = 0,
+ SCANNED_EMPTY_SPACE = -1,
+ SCANNED_A_NODE = -2,
+ SCANNED_A_CORRUPT_NODE = -3,
+ SCANNED_A_BAD_PAD_NODE = -4,
+};
+
+/*
+ * LPT cnode flag bits.
+ *
+ * DIRTY_CNODE: cnode is dirty
+ * COW_CNODE: cnode is being committed and must be copied before writing
+ * OBSOLETE_CNODE: cnode is being committed and has been copied (or deleted),
+ * so it can (and must) be freed when the commit is finished
+ */
+enum {
+ DIRTY_CNODE = 0,
+ COW_CNODE = 1,
+ OBSOLETE_CNODE = 2,
+};
+
+/*
+ * Dirty flag bits (lpt_drty_flgs) for LPT special nodes.
+ *
+ * LTAB_DIRTY: ltab node is dirty
+ * LSAVE_DIRTY: lsave node is dirty
+ */
+enum {
+ LTAB_DIRTY = 1,
+ LSAVE_DIRTY = 2,
+};
+
+/*
+ * Return codes used by the garbage collector.
+ * @LEB_FREED: the logical eraseblock was freed and is ready to use
+ * @LEB_FREED_IDX: indexing LEB was freed and can be used only after the commit
+ * @LEB_RETAINED: the logical eraseblock was freed and retained for GC purposes
+ */
+enum {
+ LEB_FREED,
+ LEB_FREED_IDX,
+ LEB_RETAINED,
+};
+
+/**
+ * struct ubifs_old_idx - index node obsoleted since last commit start.
+ * @rb: rb-tree node
+ * @lnum: LEB number of obsoleted index node
+ * @offs: offset of obsoleted index node
+ */
+struct ubifs_old_idx {
+ struct rb_node rb;
+ int lnum;
+ int offs;
+};
+
+/* The below union makes it easier to deal with keys */
+union ubifs_key {
+ uint8_t u8[CUR_MAX_KEY_LEN];
+ uint32_t u32[CUR_MAX_KEY_LEN/4];
+ uint64_t u64[CUR_MAX_KEY_LEN/8];
+ __le32 j32[CUR_MAX_KEY_LEN/4];
+};
+
+/**
+ * struct ubifs_scan_node - UBIFS scanned node information.
+ * @list: list of scanned nodes
+ * @key: key of node scanned (if it has one)
+ * @sqnum: sequence number
+ * @type: type of node scanned
+ * @offs: offset with LEB of node scanned
+ * @len: length of node scanned
+ * @node: raw node
+ */
+struct ubifs_scan_node {
+ struct list_head list;
+ union ubifs_key key;
+ unsigned long long sqnum;
+ int type;
+ int offs;
+ int len;
+ void *node;
+};
+
+/**
+ * struct ubifs_scan_leb - UBIFS scanned LEB information.
+ * @lnum: logical eraseblock number
+ * @nodes_cnt: number of nodes scanned
+ * @nodes: list of struct ubifs_scan_node
+ * @endpt: end point (and therefore the start of empty space)
+ * @ecc: read returned -EBADMSG
+ * @buf: buffer containing entire LEB scanned
+ */
+struct ubifs_scan_leb {
+ int lnum;
+ int nodes_cnt;
+ struct list_head nodes;
+ int endpt;
+ int ecc;
+ void *buf;
+};
+
+/**
+ * struct ubifs_gced_idx_leb - garbage-collected indexing LEB.
+ * @list: list
+ * @lnum: LEB number
+ * @unmap: OK to unmap this LEB
+ *
+ * This data structure is used to temporary store garbage-collected indexing
+ * LEBs - they are not released immediately, but only after the next commit.
+ * This is needed to guarantee recoverability.
+ */
+struct ubifs_gced_idx_leb {
+ struct list_head list;
+ int lnum;
+ int unmap;
+};
+
+/**
+ * struct ubifs_inode - UBIFS in-memory inode description.
+ * @vfs_inode: VFS inode description object
+ * @creat_sqnum: sequence number at time of creation
+ * @del_cmtno: commit number corresponding to the time the inode was deleted,
+ * protected by @c->commit_sem;
+ * @xattr_size: summarized size of all extended attributes in bytes
+ * @xattr_cnt: count of extended attributes this inode has
+ * @xattr_names: sum of lengths of all extended attribute names belonging to
+ * this inode
+ * @dirty: non-zero if the inode is dirty
+ * @xattr: non-zero if this is an extended attribute inode
+ * @bulk_read: non-zero if bulk-read should be used
+ * @ui_mutex: serializes inode write-back with the rest of VFS operations,
+ * serializes "clean <-> dirty" state changes, serializes bulk-read,
+ * protects @dirty, @bulk_read, @ui_size, and @xattr_size
+ * @ui_lock: protects @synced_i_size
+ * @synced_i_size: synchronized size of inode, i.e. the value of inode size
+ * currently stored on the flash; used only for regular file
+ * inodes
+ * @ui_size: inode size used by UBIFS when writing to flash
+ * @flags: inode flags (@UBIFS_COMPR_FL, etc)
+ * @compr_type: default compression type used for this inode
+ * @last_page_read: page number of last page read (for bulk read)
+ * @read_in_a_row: number of consecutive pages read in a row (for bulk read)
+ * @data_len: length of the data attached to the inode
+ * @data: inode's data
+ *
+ * @ui_mutex exists for two main reasons. At first it prevents inodes from
+ * being written back while UBIFS changing them, being in the middle of an VFS
+ * operation. This way UBIFS makes sure the inode fields are consistent. For
+ * example, in 'ubifs_rename()' we change 3 inodes simultaneously, and
+ * write-back must not write any of them before we have finished.
+ *
+ * The second reason is budgeting - UBIFS has to budget all operations. If an
+ * operation is going to mark an inode dirty, it has to allocate budget for
+ * this. It cannot just mark it dirty because there is no guarantee there will
+ * be enough flash space to write the inode back later. This means UBIFS has
+ * to have full control over inode "clean <-> dirty" transitions (and pages
+ * actually). But unfortunately, VFS marks inodes dirty in many places, and it
+ * does not ask the file-system if it is allowed to do so (there is a notifier,
+ * but it is not enough), i.e., there is no mechanism to synchronize with this.
+ * So UBIFS has its own inode dirty flag and its own mutex to serialize
+ * "clean <-> dirty" transitions.
+ *
+ * The @synced_i_size field is used to make sure we never write pages which are
+ * beyond last synchronized inode size. See 'ubifs_writepage()' for more
+ * information.
+ *
+ * The @ui_size is a "shadow" variable for @inode->i_size and UBIFS uses
+ * @ui_size instead of @inode->i_size. The reason for this is that UBIFS cannot
+ * make sure @inode->i_size is always changed under @ui_mutex, because it
+ * cannot call 'vmtruncate()' with @ui_mutex locked, because it would deadlock
+ * with 'ubifs_writepage()' (see file.c). All the other inode fields are
+ * changed under @ui_mutex, so they do not need "shadow" fields. Note, one
+ * could consider to rework locking and base it on "shadow" fields.
+ */
+struct ubifs_inode {
+ struct inode vfs_inode;
+ unsigned long long creat_sqnum;
+ unsigned long long del_cmtno;
+ unsigned int xattr_size;
+ unsigned int xattr_cnt;
+ unsigned int xattr_names;
+ unsigned int dirty:1;
+ unsigned int xattr:1;
+ unsigned int bulk_read:1;
+ unsigned int compr_type:2;
+ struct mutex ui_mutex;
+ spinlock_t ui_lock;
+ loff_t synced_i_size;
+ loff_t ui_size;
+ int flags;
+ pgoff_t last_page_read;
+ pgoff_t read_in_a_row;
+ int data_len;
+ void *data;
+};
+
+/**
+ * struct ubifs_unclean_leb - records a LEB recovered under read-only mode.
+ * @list: list
+ * @lnum: LEB number of recovered LEB
+ * @endpt: offset where recovery ended
+ *
+ * This structure records a LEB identified during recovery that needs to be
+ * cleaned but was not because UBIFS was mounted read-only. The information
+ * is used to clean the LEB when remounting to read-write mode.
+ */
+struct ubifs_unclean_leb {
+ struct list_head list;
+ int lnum;
+ int endpt;
+};
+
+/*
+ * LEB properties flags.
+ *
+ * LPROPS_UNCAT: not categorized
+ * LPROPS_DIRTY: dirty > free, dirty >= @c->dead_wm, not index
+ * LPROPS_DIRTY_IDX: dirty + free > @c->min_idx_node_sze and index
+ * LPROPS_FREE: free > 0, dirty < @c->dead_wm, not empty, not index
+ * LPROPS_HEAP_CNT: number of heaps used for storing categorized LEBs
+ * LPROPS_EMPTY: LEB is empty, not taken
+ * LPROPS_FREEABLE: free + dirty == leb_size, not index, not taken
+ * LPROPS_FRDI_IDX: free + dirty == leb_size and index, may be taken
+ * LPROPS_CAT_MASK: mask for the LEB categories above
+ * LPROPS_TAKEN: LEB was taken (this flag is not saved on the media)
+ * LPROPS_INDEX: LEB contains indexing nodes (this flag also exists on flash)
+ */
+enum {
+ LPROPS_UNCAT = 0,
+ LPROPS_DIRTY = 1,
+ LPROPS_DIRTY_IDX = 2,
+ LPROPS_FREE = 3,
+ LPROPS_HEAP_CNT = 3,
+ LPROPS_EMPTY = 4,
+ LPROPS_FREEABLE = 5,
+ LPROPS_FRDI_IDX = 6,
+ LPROPS_CAT_MASK = 15,
+ LPROPS_TAKEN = 16,
+ LPROPS_INDEX = 32,
+};
+
+/**
+ * struct ubifs_lprops - logical eraseblock properties.
+ * @free: amount of free space in bytes
+ * @dirty: amount of dirty space in bytes
+ * @flags: LEB properties flags (see above)
+ * @lnum: LEB number
+ * @list: list of same-category lprops (for LPROPS_EMPTY and LPROPS_FREEABLE)
+ * @hpos: heap position in heap of same-category lprops (other categories)
+ */
+struct ubifs_lprops {
+ int free;
+ int dirty;
+ int flags;
+ int lnum;
+ union {
+ struct list_head list;
+ int hpos;
+ };
+};
+
+/**
+ * struct ubifs_lpt_lprops - LPT logical eraseblock properties.
+ * @free: amount of free space in bytes
+ * @dirty: amount of dirty space in bytes
+ * @tgc: trivial GC flag (1 => unmap after commit end)
+ * @cmt: commit flag (1 => reserved for commit)
+ */
+struct ubifs_lpt_lprops {
+ int free;
+ int dirty;
+ unsigned tgc:1;
+ unsigned cmt:1;
+};
+
+/**
+ * struct ubifs_lp_stats - statistics of eraseblocks in the main area.
+ * @empty_lebs: number of empty LEBs
+ * @taken_empty_lebs: number of taken LEBs
+ * @idx_lebs: number of indexing LEBs
+ * @total_free: total free space in bytes (includes all LEBs)
+ * @total_dirty: total dirty space in bytes (includes all LEBs)
+ * @total_used: total used space in bytes (does not include index LEBs)
+ * @total_dead: total dead space in bytes (does not include index LEBs)
+ * @total_dark: total dark space in bytes (does not include index LEBs)
+ *
+ * The @taken_empty_lebs field counts the LEBs that are in the transient state
+ * of having been "taken" for use but not yet written to. @taken_empty_lebs is
+ * needed to account correctly for @gc_lnum, otherwise @empty_lebs could be
+ * used by itself (in which case 'unused_lebs' would be a better name). In the
+ * case of @gc_lnum, it is "taken" at mount time or whenever a LEB is retained
+ * by GC, but unlike other empty LEBs that are "taken", it may not be written
+ * straight away (i.e. before the next commit start or unmount), so either
+ * @gc_lnum must be specially accounted for, or the current approach followed
+ * i.e. count it under @taken_empty_lebs.
+ *
+ * @empty_lebs includes @taken_empty_lebs.
+ *
+ * @total_used, @total_dead and @total_dark fields do not account indexing
+ * LEBs.
+ */
+struct ubifs_lp_stats {
+ int empty_lebs;
+ int taken_empty_lebs;
+ int idx_lebs;
+ long long total_free;
+ long long total_dirty;
+ long long total_used;
+ long long total_dead;
+ long long total_dark;
+};
+
+struct ubifs_nnode;
+
+/**
+ * struct ubifs_cnode - LEB Properties Tree common node.
+ * @parent: parent nnode
+ * @cnext: next cnode to commit
+ * @flags: flags (%DIRTY_LPT_NODE or %OBSOLETE_LPT_NODE)
+ * @iip: index in parent
+ * @level: level in the tree (zero for pnodes, greater than zero for nnodes)
+ * @num: node number
+ */
+struct ubifs_cnode {
+ struct ubifs_nnode *parent;
+ struct ubifs_cnode *cnext;
+ unsigned long flags;
+ int iip;
+ int level;
+ int num;
+};
+
+/**
+ * struct ubifs_pnode - LEB Properties Tree leaf node.
+ * @parent: parent nnode
+ * @cnext: next cnode to commit
+ * @flags: flags (%DIRTY_LPT_NODE or %OBSOLETE_LPT_NODE)
+ * @iip: index in parent
+ * @level: level in the tree (always zero for pnodes)
+ * @num: node number
+ * @lprops: LEB properties array
+ */
+struct ubifs_pnode {
+ struct ubifs_nnode *parent;
+ struct ubifs_cnode *cnext;
+ unsigned long flags;
+ int iip;
+ int level;
+ int num;
+ struct ubifs_lprops lprops[UBIFS_LPT_FANOUT];
+};
+
+/**
+ * struct ubifs_nbranch - LEB Properties Tree internal node branch.
+ * @lnum: LEB number of child
+ * @offs: offset of child
+ * @nnode: nnode child
+ * @pnode: pnode child
+ * @cnode: cnode child
+ */
+struct ubifs_nbranch {
+ int lnum;
+ int offs;
+ union {
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+ struct ubifs_cnode *cnode;
+ };
+};
+
+/**
+ * struct ubifs_nnode - LEB Properties Tree internal node.
+ * @parent: parent nnode
+ * @cnext: next cnode to commit
+ * @flags: flags (%DIRTY_LPT_NODE or %OBSOLETE_LPT_NODE)
+ * @iip: index in parent
+ * @level: level in the tree (always greater than zero for nnodes)
+ * @num: node number
+ * @nbranch: branches to child nodes
+ */
+struct ubifs_nnode {
+ struct ubifs_nnode *parent;
+ struct ubifs_cnode *cnext;
+ unsigned long flags;
+ int iip;
+ int level;
+ int num;
+ struct ubifs_nbranch nbranch[UBIFS_LPT_FANOUT];
+};
+
+/**
+ * struct ubifs_lpt_heap - heap of categorized lprops.
+ * @arr: heap array
+ * @cnt: number in heap
+ * @max_cnt: maximum number allowed in heap
+ *
+ * There are %LPROPS_HEAP_CNT heaps.
+ */
+struct ubifs_lpt_heap {
+ struct ubifs_lprops **arr;
+ int cnt;
+ int max_cnt;
+};
+
+/*
+ * Return codes for LPT scan callback function.
+ *
+ * LPT_SCAN_CONTINUE: continue scanning
+ * LPT_SCAN_ADD: add the LEB properties scanned to the tree in memory
+ * LPT_SCAN_STOP: stop scanning
+ */
+enum {
+ LPT_SCAN_CONTINUE = 0,
+ LPT_SCAN_ADD = 1,
+ LPT_SCAN_STOP = 2,
+};
+
+struct ubifs_info;
+
+/* Callback used by the 'ubifs_lpt_scan_nolock()' function */
+typedef int (*ubifs_lpt_scan_callback)(struct ubifs_info *c,
+ const struct ubifs_lprops *lprops,
+ int in_tree, void *data);
+
+/**
+ * struct ubifs_wbuf - UBIFS write-buffer.
+ * @c: UBIFS file-system description object
+ * @buf: write-buffer (of min. flash I/O unit size)
+ * @lnum: logical eraseblock number the write-buffer points to
+ * @offs: write-buffer offset in this logical eraseblock
+ * @avail: number of bytes available in the write-buffer
+ * @used: number of used bytes in the write-buffer
+ * @dtype: type of data stored in this LEB (%UBI_LONGTERM, %UBI_SHORTTERM,
+ * %UBI_UNKNOWN)
+ * @jhead: journal head the mutex belongs to (note, needed only to shut lockdep
+ * up by 'mutex_lock_nested()).
+ * @sync_callback: write-buffer synchronization callback
+ * @io_mutex: serializes write-buffer I/O
+ * @lock: serializes @buf, @lnum, @offs, @avail, @used, @next_ino and @inodes
+ * fields
+ * @timer: write-buffer timer
+ * @timeout: timer expire interval in jiffies
+ * @need_sync: it is set if its timer expired and needs sync
+ * @next_ino: points to the next position of the following inode number
+ * @inodes: stores the inode numbers of the nodes which are in wbuf
+ *
+ * The write-buffer synchronization callback is called when the write-buffer is
+ * synchronized in order to notify how much space was wasted due to
+ * write-buffer padding and how much free space is left in the LEB.
+ *
+ * Note: the fields @buf, @lnum, @offs, @avail and @used can be read under
+ * spin-lock or mutex because they are written under both mutex and spin-lock.
+ * @buf is appended to under mutex but overwritten under both mutex and
+ * spin-lock. Thus the data between @buf and @buf + @used can be read under
+ * spinlock.
+ */
+struct ubifs_wbuf {
+ struct ubifs_info *c;
+ void *buf;
+ int lnum;
+ int offs;
+ int avail;
+ int used;
+ int dtype;
+ int jhead;
+ int (*sync_callback)(struct ubifs_info *c, int lnum, int free, int pad);
+ struct mutex io_mutex;
+ spinlock_t lock;
+ int timeout;
+ int need_sync;
+ int next_ino;
+ ino_t *inodes;
+};
+
+/**
+ * struct ubifs_bud - bud logical eraseblock.
+ * @lnum: logical eraseblock number
+ * @start: where the (uncommitted) bud data starts
+ * @jhead: journal head number this bud belongs to
+ * @list: link in the list buds belonging to the same journal head
+ * @rb: link in the tree of all buds
+ */
+struct ubifs_bud {
+ int lnum;
+ int start;
+ int jhead;
+ struct list_head list;
+ struct rb_node rb;
+};
+
+/**
+ * struct ubifs_jhead - journal head.
+ * @wbuf: head's write-buffer
+ * @buds_list: list of bud LEBs belonging to this journal head
+ *
+ * Note, the @buds list is protected by the @c->buds_lock.
+ */
+struct ubifs_jhead {
+ struct ubifs_wbuf wbuf;
+ struct list_head buds_list;
+};
+
+/**
+ * struct ubifs_zbranch - key/coordinate/length branch stored in znodes.
+ * @key: key
+ * @znode: znode address in memory
+ * @lnum: LEB number of the target node (indexing node or data node)
+ * @offs: target node offset within @lnum
+ * @len: target node length
+ */
+struct ubifs_zbranch {
+ union ubifs_key key;
+ union {
+ struct ubifs_znode *znode;
+ void *leaf;
+ };
+ int lnum;
+ int offs;
+ int len;
+};
+
+/**
+ * struct ubifs_znode - in-memory representation of an indexing node.
+ * @parent: parent znode or NULL if it is the root
+ * @cnext: next znode to commit
+ * @flags: znode flags (%DIRTY_ZNODE, %COW_ZNODE or %OBSOLETE_ZNODE)
+ * @time: last access time (seconds)
+ * @level: level of the entry in the TNC tree
+ * @child_cnt: count of child znodes
+ * @iip: index in parent's zbranch array
+ * @alt: lower bound of key range has altered i.e. child inserted at slot 0
+ * @lnum: LEB number of the corresponding indexing node
+ * @offs: offset of the corresponding indexing node
+ * @len: length of the corresponding indexing node
+ * @zbranch: array of znode branches (@c->fanout elements)
+ */
+struct ubifs_znode {
+ struct ubifs_znode *parent;
+ struct ubifs_znode *cnext;
+ unsigned long flags;
+ unsigned long time;
+ int level;
+ int child_cnt;
+ int iip;
+ int alt;
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ int lnum, offs, len;
+#endif
+ struct ubifs_zbranch zbranch[];
+};
+
+/**
+ * struct bu_info - bulk-read information.
+ * @key: first data node key
+ * @zbranch: zbranches of data nodes to bulk read
+ * @buf: buffer to read into
+ * @buf_len: buffer length
+ * @gc_seq: GC sequence number to detect races with GC
+ * @cnt: number of data nodes for bulk read
+ * @blk_cnt: number of data blocks including holes
+ * @oef: end of file reached
+ */
+struct bu_info {
+ union ubifs_key key;
+ struct ubifs_zbranch zbranch[UBIFS_MAX_BULK_READ];
+ void *buf;
+ int buf_len;
+ int gc_seq;
+ int cnt;
+ int blk_cnt;
+ int eof;
+};
+
+/**
+ * struct ubifs_node_range - node length range description data structure.
+ * @len: fixed node length
+ * @min_len: minimum possible node length
+ * @max_len: maximum possible node length
+ *
+ * If @max_len is %0, the node has fixed length @len.
+ */
+struct ubifs_node_range {
+ union {
+ int len;
+ int min_len;
+ };
+ int max_len;
+};
+
+/**
+ * struct ubifs_compressor - UBIFS compressor description structure.
+ * @compr_type: compressor type (%UBIFS_COMPR_LZO, etc)
+ * @cc: cryptoapi compressor handle
+ * @comp_mutex: mutex used during compression
+ * @decomp_mutex: mutex used during decompression
+ * @name: compressor name
+ * @capi_name: cryptoapi compressor name
+ */
+struct ubifs_compressor {
+ int compr_type;
+ char *name;
+ char *capi_name;
+ int (*decompress)(const unsigned char *in, size_t in_len,
+ unsigned char *out, size_t *out_len);
+};
+
+/**
+ * struct ubifs_budget_req - budget requirements of an operation.
+ *
+ * @fast: non-zero if the budgeting should try to acquire budget quickly and
+ * should not try to call write-back
+ * @recalculate: non-zero if @idx_growth, @data_growth, and @dd_growth fields
+ * have to be re-calculated
+ * @new_page: non-zero if the operation adds a new page
+ * @dirtied_page: non-zero if the operation makes a page dirty
+ * @new_dent: non-zero if the operation adds a new directory entry
+ * @mod_dent: non-zero if the operation removes or modifies an existing
+ * directory entry
+ * @new_ino: non-zero if the operation adds a new inode
+ * @new_ino_d: now much data newly created inode contains
+ * @dirtied_ino: how many inodes the operation makes dirty
+ * @dirtied_ino_d: now much data dirtied inode contains
+ * @idx_growth: how much the index will supposedly grow
+ * @data_growth: how much new data the operation will supposedly add
+ * @dd_growth: how much data that makes other data dirty the operation will
+ * supposedly add
+ *
+ * @idx_growth, @data_growth and @dd_growth are not used in budget request. The
+ * budgeting subsystem caches index and data growth values there to avoid
+ * re-calculating them when the budget is released. However, if @idx_growth is
+ * %-1, it is calculated by the release function using other fields.
+ *
+ * An inode may contain 4KiB of data at max., thus the widths of @new_ino_d
+ * is 13 bits, and @dirtied_ino_d - 15, because up to 4 inodes may be made
+ * dirty by the re-name operation.
+ *
+ * Note, UBIFS aligns node lengths to 8-bytes boundary, so the requester has to
+ * make sure the amount of inode data which contribute to @new_ino_d and
+ * @dirtied_ino_d fields are aligned.
+ */
+struct ubifs_budget_req {
+ unsigned int fast:1;
+ unsigned int recalculate:1;
+#ifndef UBIFS_DEBUG
+ unsigned int new_page:1;
+ unsigned int dirtied_page:1;
+ unsigned int new_dent:1;
+ unsigned int mod_dent:1;
+ unsigned int new_ino:1;
+ unsigned int new_ino_d:13;
+ unsigned int dirtied_ino:4;
+ unsigned int dirtied_ino_d:15;
+#else
+ /* Not bit-fields to check for overflows */
+ unsigned int new_page;
+ unsigned int dirtied_page;
+ unsigned int new_dent;
+ unsigned int mod_dent;
+ unsigned int new_ino;
+ unsigned int new_ino_d;
+ unsigned int dirtied_ino;
+ unsigned int dirtied_ino_d;
+#endif
+ int idx_growth;
+ int data_growth;
+ int dd_growth;
+};
+
+/**
+ * struct ubifs_orphan - stores the inode number of an orphan.
+ * @rb: rb-tree node of rb-tree of orphans sorted by inode number
+ * @list: list head of list of orphans in order added
+ * @new_list: list head of list of orphans added since the last commit
+ * @cnext: next orphan to commit
+ * @dnext: next orphan to delete
+ * @inum: inode number
+ * @new: %1 => added since the last commit, otherwise %0
+ */
+struct ubifs_orphan {
+ struct rb_node rb;
+ struct list_head list;
+ struct list_head new_list;
+ struct ubifs_orphan *cnext;
+ struct ubifs_orphan *dnext;
+ ino_t inum;
+ int new;
+};
+
+/**
+ * struct ubifs_mount_opts - UBIFS-specific mount options information.
+ * @unmount_mode: selected unmount mode (%0 default, %1 normal, %2 fast)
+ * @bulk_read: enable/disable bulk-reads (%0 default, %1 disabe, %2 enable)
+ * @chk_data_crc: enable/disable CRC data checking when reading data nodes
+ * (%0 default, %1 disabe, %2 enable)
+ * @override_compr: override default compressor (%0 - do not override and use
+ * superblock compressor, %1 - override and use compressor
+ * specified in @compr_type)
+ * @compr_type: compressor type to override the superblock compressor with
+ * (%UBIFS_COMPR_NONE, etc)
+ */
+struct ubifs_mount_opts {
+ unsigned int unmount_mode:2;
+ unsigned int bulk_read:2;
+ unsigned int chk_data_crc:2;
+ unsigned int override_compr:1;
+ unsigned int compr_type:2;
+};
+
+struct ubifs_debug_info;
+
+/**
+ * struct ubifs_info - UBIFS file-system description data structure
+ * (per-superblock).
+ * @vfs_sb: VFS @struct super_block object
+ * @bdi: backing device info object to make VFS happy and disable read-ahead
+ *
+ * @highest_inum: highest used inode number
+ * @max_sqnum: current global sequence number
+ * @cmt_no: commit number of the last successfully completed commit, protected
+ * by @commit_sem
+ * @cnt_lock: protects @highest_inum and @max_sqnum counters
+ * @fmt_version: UBIFS on-flash format version
+ * @uuid: UUID from super block
+ *
+ * @lhead_lnum: log head logical eraseblock number
+ * @lhead_offs: log head offset
+ * @ltail_lnum: log tail logical eraseblock number (offset is always 0)
+ * @log_mutex: protects the log, @lhead_lnum, @lhead_offs, @ltail_lnum, and
+ * @bud_bytes
+ * @min_log_bytes: minimum required number of bytes in the log
+ * @cmt_bud_bytes: used during commit to temporarily amount of bytes in
+ * committed buds
+ *
+ * @buds: tree of all buds indexed by bud LEB number
+ * @bud_bytes: how many bytes of flash is used by buds
+ * @buds_lock: protects the @buds tree, @bud_bytes, and per-journal head bud
+ * lists
+ * @jhead_cnt: count of journal heads
+ * @jheads: journal heads (head zero is base head)
+ * @max_bud_bytes: maximum number of bytes allowed in buds
+ * @bg_bud_bytes: number of bud bytes when background commit is initiated
+ * @old_buds: buds to be released after commit ends
+ * @max_bud_cnt: maximum number of buds
+ *
+ * @commit_sem: synchronizes committer with other processes
+ * @cmt_state: commit state
+ * @cs_lock: commit state lock
+ * @cmt_wq: wait queue to sleep on if the log is full and a commit is running
+ *
+ * @big_lpt: flag that LPT is too big to write whole during commit
+ * @no_chk_data_crc: do not check CRCs when reading data nodes (except during
+ * recovery)
+ * @bulk_read: enable bulk-reads
+ * @default_compr: default compression algorithm (%UBIFS_COMPR_LZO, etc)
+ *
+ * @tnc_mutex: protects the Tree Node Cache (TNC), @zroot, @cnext, @enext, and
+ * @calc_idx_sz
+ * @zroot: zbranch which points to the root index node and znode
+ * @cnext: next znode to commit
+ * @enext: next znode to commit to empty space
+ * @gap_lebs: array of LEBs used by the in-gaps commit method
+ * @cbuf: commit buffer
+ * @ileb_buf: buffer for commit in-the-gaps method
+ * @ileb_len: length of data in ileb_buf
+ * @ihead_lnum: LEB number of index head
+ * @ihead_offs: offset of index head
+ * @ilebs: pre-allocated index LEBs
+ * @ileb_cnt: number of pre-allocated index LEBs
+ * @ileb_nxt: next pre-allocated index LEBs
+ * @old_idx: tree of index nodes obsoleted since the last commit start
+ * @bottom_up_buf: a buffer which is used by 'dirty_cow_bottom_up()' in tnc.c
+ *
+ * @mst_node: master node
+ * @mst_offs: offset of valid master node
+ * @mst_mutex: protects the master node area, @mst_node, and @mst_offs
+ *
+ * @max_bu_buf_len: maximum bulk-read buffer length
+ * @bu_mutex: protects the pre-allocated bulk-read buffer and @c->bu
+ * @bu: pre-allocated bulk-read information
+ *
+ * @log_lebs: number of logical eraseblocks in the log
+ * @log_bytes: log size in bytes
+ * @log_last: last LEB of the log
+ * @lpt_lebs: number of LEBs used for lprops table
+ * @lpt_first: first LEB of the lprops table area
+ * @lpt_last: last LEB of the lprops table area
+ * @orph_lebs: number of LEBs used for the orphan area
+ * @orph_first: first LEB of the orphan area
+ * @orph_last: last LEB of the orphan area
+ * @main_lebs: count of LEBs in the main area
+ * @main_first: first LEB of the main area
+ * @main_bytes: main area size in bytes
+ *
+ * @key_hash_type: type of the key hash
+ * @key_hash: direntry key hash function
+ * @key_fmt: key format
+ * @key_len: key length
+ * @fanout: fanout of the index tree (number of links per indexing node)
+ *
+ * @min_io_size: minimal input/output unit size
+ * @min_io_shift: number of bits in @min_io_size minus one
+ * @leb_size: logical eraseblock size in bytes
+ * @half_leb_size: half LEB size
+ * @leb_cnt: count of logical eraseblocks
+ * @max_leb_cnt: maximum count of logical eraseblocks
+ * @old_leb_cnt: count of logical eraseblocks before re-size
+ * @ro_media: the underlying UBI volume is read-only
+ *
+ * @dirty_pg_cnt: number of dirty pages (not used)
+ * @dirty_zn_cnt: number of dirty znodes
+ * @clean_zn_cnt: number of clean znodes
+ *
+ * @budg_idx_growth: amount of bytes budgeted for index growth
+ * @budg_data_growth: amount of bytes budgeted for cached data
+ * @budg_dd_growth: amount of bytes budgeted for cached data that will make
+ * other data dirty
+ * @budg_uncommitted_idx: amount of bytes were budgeted for growth of the index,
+ * but which still have to be taken into account because
+ * the index has not been committed so far
+ * @space_lock: protects @budg_idx_growth, @budg_data_growth, @budg_dd_growth,
+ * @budg_uncommited_idx, @min_idx_lebs, @old_idx_sz, @lst,
+ * @nospace, and @nospace_rp;
+ * @min_idx_lebs: minimum number of LEBs required for the index
+ * @old_idx_sz: size of index on flash
+ * @calc_idx_sz: temporary variable which is used to calculate new index size
+ * (contains accurate new index size at end of TNC commit start)
+ * @lst: lprops statistics
+ * @nospace: non-zero if the file-system does not have flash space (used as
+ * optimization)
+ * @nospace_rp: the same as @nospace, but additionally means that even reserved
+ * pool is full
+ *
+ * @page_budget: budget for a page
+ * @inode_budget: budget for an inode
+ * @dent_budget: budget for a directory entry
+ *
+ * @ref_node_alsz: size of the LEB reference node aligned to the min. flash
+ * I/O unit
+ * @mst_node_alsz: master node aligned size
+ * @min_idx_node_sz: minimum indexing node aligned on 8-bytes boundary
+ * @max_idx_node_sz: maximum indexing node aligned on 8-bytes boundary
+ * @max_inode_sz: maximum possible inode size in bytes
+ * @max_znode_sz: size of znode in bytes
+ *
+ * @leb_overhead: how many bytes are wasted in an LEB when it is filled with
+ * data nodes of maximum size - used in free space reporting
+ * @dead_wm: LEB dead space watermark
+ * @dark_wm: LEB dark space watermark
+ * @block_cnt: count of 4KiB blocks on the FS
+ *
+ * @ranges: UBIFS node length ranges
+ * @ubi: UBI volume descriptor
+ * @di: UBI device information
+ * @vi: UBI volume information
+ *
+ * @orph_tree: rb-tree of orphan inode numbers
+ * @orph_list: list of orphan inode numbers in order added
+ * @orph_new: list of orphan inode numbers added since last commit
+ * @orph_cnext: next orphan to commit
+ * @orph_dnext: next orphan to delete
+ * @orphan_lock: lock for orph_tree and orph_new
+ * @orph_buf: buffer for orphan nodes
+ * @new_orphans: number of orphans since last commit
+ * @cmt_orphans: number of orphans being committed
+ * @tot_orphans: number of orphans in the rb_tree
+ * @max_orphans: maximum number of orphans allowed
+ * @ohead_lnum: orphan head LEB number
+ * @ohead_offs: orphan head offset
+ * @no_orphs: non-zero if there are no orphans
+ *
+ * @bgt: UBIFS background thread
+ * @bgt_name: background thread name
+ * @need_bgt: if background thread should run
+ * @need_wbuf_sync: if write-buffers have to be synchronized
+ *
+ * @gc_lnum: LEB number used for garbage collection
+ * @sbuf: a buffer of LEB size used by GC and replay for scanning
+ * @idx_gc: list of index LEBs that have been garbage collected
+ * @idx_gc_cnt: number of elements on the idx_gc list
+ * @gc_seq: incremented for every non-index LEB garbage collected
+ * @gced_lnum: last non-index LEB that was garbage collected
+ *
+ * @infos_list: links all 'ubifs_info' objects
+ * @umount_mutex: serializes shrinker and un-mount
+ * @shrinker_run_no: shrinker run number
+ *
+ * @space_bits: number of bits needed to record free or dirty space
+ * @lpt_lnum_bits: number of bits needed to record a LEB number in the LPT
+ * @lpt_offs_bits: number of bits needed to record an offset in the LPT
+ * @lpt_spc_bits: number of bits needed to space in the LPT
+ * @pcnt_bits: number of bits needed to record pnode or nnode number
+ * @lnum_bits: number of bits needed to record LEB number
+ * @nnode_sz: size of on-flash nnode
+ * @pnode_sz: size of on-flash pnode
+ * @ltab_sz: size of on-flash LPT lprops table
+ * @lsave_sz: size of on-flash LPT save table
+ * @pnode_cnt: number of pnodes
+ * @nnode_cnt: number of nnodes
+ * @lpt_hght: height of the LPT
+ * @pnodes_have: number of pnodes in memory
+ *
+ * @lp_mutex: protects lprops table and all the other lprops-related fields
+ * @lpt_lnum: LEB number of the root nnode of the LPT
+ * @lpt_offs: offset of the root nnode of the LPT
+ * @nhead_lnum: LEB number of LPT head
+ * @nhead_offs: offset of LPT head
+ * @lpt_drty_flgs: dirty flags for LPT special nodes e.g. ltab
+ * @dirty_nn_cnt: number of dirty nnodes
+ * @dirty_pn_cnt: number of dirty pnodes
+ * @check_lpt_free: flag that indicates LPT GC may be needed
+ * @lpt_sz: LPT size
+ * @lpt_nod_buf: buffer for an on-flash nnode or pnode
+ * @lpt_buf: buffer of LEB size used by LPT
+ * @nroot: address in memory of the root nnode of the LPT
+ * @lpt_cnext: next LPT node to commit
+ * @lpt_heap: array of heaps of categorized lprops
+ * @dirty_idx: a (reverse sorted) copy of the LPROPS_DIRTY_IDX heap as at
+ * previous commit start
+ * @uncat_list: list of un-categorized LEBs
+ * @empty_list: list of empty LEBs
+ * @freeable_list: list of freeable non-index LEBs (free + dirty == leb_size)
+ * @frdi_idx_list: list of freeable index LEBs (free + dirty == leb_size)
+ * @freeable_cnt: number of freeable LEBs in @freeable_list
+ *
+ * @ltab_lnum: LEB number of LPT's own lprops table
+ * @ltab_offs: offset of LPT's own lprops table
+ * @ltab: LPT's own lprops table
+ * @ltab_cmt: LPT's own lprops table (commit copy)
+ * @lsave_cnt: number of LEB numbers in LPT's save table
+ * @lsave_lnum: LEB number of LPT's save table
+ * @lsave_offs: offset of LPT's save table
+ * @lsave: LPT's save table
+ * @lscan_lnum: LEB number of last LPT scan
+ *
+ * @rp_size: size of the reserved pool in bytes
+ * @report_rp_size: size of the reserved pool reported to user-space
+ * @rp_uid: reserved pool user ID
+ * @rp_gid: reserved pool group ID
+ *
+ * @empty: if the UBI device is empty
+ * @replay_tree: temporary tree used during journal replay
+ * @replay_list: temporary list used during journal replay
+ * @replay_buds: list of buds to replay
+ * @cs_sqnum: sequence number of first node in the log (commit start node)
+ * @replay_sqnum: sequence number of node currently being replayed
+ * @need_recovery: file-system needs recovery
+ * @replaying: set to %1 during journal replay
+ * @unclean_leb_list: LEBs to recover when mounting ro to rw
+ * @rcvrd_mst_node: recovered master node to write when mounting ro to rw
+ * @size_tree: inode size information for recovery
+ * @remounting_rw: set while remounting from ro to rw (sb flags have MS_RDONLY)
+ * @always_chk_crc: always check CRCs (while mounting and remounting rw)
+ * @mount_opts: UBIFS-specific mount options
+ *
+ * @dbg: debugging-related information
+ */
+struct ubifs_info {
+ struct super_block *vfs_sb;
+
+ ino_t highest_inum;
+ unsigned long long max_sqnum;
+ unsigned long long cmt_no;
+ spinlock_t cnt_lock;
+ int fmt_version;
+ unsigned char uuid[16];
+
+ int lhead_lnum;
+ int lhead_offs;
+ int ltail_lnum;
+ struct mutex log_mutex;
+ int min_log_bytes;
+ long long cmt_bud_bytes;
+
+ struct rb_root buds;
+ long long bud_bytes;
+ spinlock_t buds_lock;
+ int jhead_cnt;
+ struct ubifs_jhead *jheads;
+ long long max_bud_bytes;
+ long long bg_bud_bytes;
+ struct list_head old_buds;
+ int max_bud_cnt;
+
+ struct rw_semaphore commit_sem;
+ int cmt_state;
+ spinlock_t cs_lock;
+ wait_queue_head_t cmt_wq;
+
+ unsigned int big_lpt:1;
+ unsigned int no_chk_data_crc:1;
+ unsigned int bulk_read:1;
+ unsigned int default_compr:2;
+
+ struct mutex tnc_mutex;
+ struct ubifs_zbranch zroot;
+ struct ubifs_znode *cnext;
+ struct ubifs_znode *enext;
+ int *gap_lebs;
+ void *cbuf;
+ void *ileb_buf;
+ int ileb_len;
+ int ihead_lnum;
+ int ihead_offs;
+ int *ilebs;
+ int ileb_cnt;
+ int ileb_nxt;
+ struct rb_root old_idx;
+ int *bottom_up_buf;
+
+ struct ubifs_mst_node *mst_node;
+ int mst_offs;
+ struct mutex mst_mutex;
+
+ int max_bu_buf_len;
+ struct mutex bu_mutex;
+ struct bu_info bu;
+
+ int log_lebs;
+ long long log_bytes;
+ int log_last;
+ int lpt_lebs;
+ int lpt_first;
+ int lpt_last;
+ int orph_lebs;
+ int orph_first;
+ int orph_last;
+ int main_lebs;
+ int main_first;
+ long long main_bytes;
+
+ uint8_t key_hash_type;
+ uint32_t (*key_hash)(const char *str, int len);
+ int key_fmt;
+ int key_len;
+ int fanout;
+
+ int min_io_size;
+ int min_io_shift;
+ int leb_size;
+ int half_leb_size;
+ int leb_cnt;
+ int max_leb_cnt;
+ int old_leb_cnt;
+ int ro_media;
+
+ long long budg_idx_growth;
+ long long budg_data_growth;
+ long long budg_dd_growth;
+ long long budg_uncommitted_idx;
+ spinlock_t space_lock;
+ int min_idx_lebs;
+ unsigned long long old_idx_sz;
+ unsigned long long calc_idx_sz;
+ struct ubifs_lp_stats lst;
+ unsigned int nospace:1;
+ unsigned int nospace_rp:1;
+
+ int page_budget;
+ int inode_budget;
+ int dent_budget;
+
+ int ref_node_alsz;
+ int mst_node_alsz;
+ int min_idx_node_sz;
+ int max_idx_node_sz;
+ long long max_inode_sz;
+ int max_znode_sz;
+
+ int leb_overhead;
+ int dead_wm;
+ int dark_wm;
+ int block_cnt;
+
+ struct ubifs_node_range ranges[UBIFS_NODE_TYPES_CNT];
+ struct ubi_volume_desc *ubi;
+ struct ubi_device_info di;
+ struct ubi_volume_info vi;
+
+ struct rb_root orph_tree;
+ struct list_head orph_list;
+ struct list_head orph_new;
+ struct ubifs_orphan *orph_cnext;
+ struct ubifs_orphan *orph_dnext;
+ spinlock_t orphan_lock;
+ void *orph_buf;
+ int new_orphans;
+ int cmt_orphans;
+ int tot_orphans;
+ int max_orphans;
+ int ohead_lnum;
+ int ohead_offs;
+ int no_orphs;
+
+ struct task_struct *bgt;
+ char bgt_name[sizeof(BGT_NAME_PATTERN) + 9];
+ int need_bgt;
+ int need_wbuf_sync;
+
+ int gc_lnum;
+ void *sbuf;
+ struct list_head idx_gc;
+ int idx_gc_cnt;
+ int gc_seq;
+ int gced_lnum;
+
+ struct list_head infos_list;
+ struct mutex umount_mutex;
+ unsigned int shrinker_run_no;
+
+ int space_bits;
+ int lpt_lnum_bits;
+ int lpt_offs_bits;
+ int lpt_spc_bits;
+ int pcnt_bits;
+ int lnum_bits;
+ int nnode_sz;
+ int pnode_sz;
+ int ltab_sz;
+ int lsave_sz;
+ int pnode_cnt;
+ int nnode_cnt;
+ int lpt_hght;
+ int pnodes_have;
+
+ struct mutex lp_mutex;
+ int lpt_lnum;
+ int lpt_offs;
+ int nhead_lnum;
+ int nhead_offs;
+ int lpt_drty_flgs;
+ int dirty_nn_cnt;
+ int dirty_pn_cnt;
+ int check_lpt_free;
+ long long lpt_sz;
+ void *lpt_nod_buf;
+ void *lpt_buf;
+ struct ubifs_nnode *nroot;
+ struct ubifs_cnode *lpt_cnext;
+ struct ubifs_lpt_heap lpt_heap[LPROPS_HEAP_CNT];
+ struct ubifs_lpt_heap dirty_idx;
+ struct list_head uncat_list;
+ struct list_head empty_list;
+ struct list_head freeable_list;
+ struct list_head frdi_idx_list;
+ int freeable_cnt;
+
+ int ltab_lnum;
+ int ltab_offs;
+ struct ubifs_lpt_lprops *ltab;
+ struct ubifs_lpt_lprops *ltab_cmt;
+ int lsave_cnt;
+ int lsave_lnum;
+ int lsave_offs;
+ int *lsave;
+ int lscan_lnum;
+
+ long long rp_size;
+ long long report_rp_size;
+ uid_t rp_uid;
+ gid_t rp_gid;
+
+ /* The below fields are used only during mounting and re-mounting */
+ int empty;
+ struct rb_root replay_tree;
+ struct list_head replay_list;
+ struct list_head replay_buds;
+ unsigned long long cs_sqnum;
+ unsigned long long replay_sqnum;
+ int need_recovery;
+ int replaying;
+ struct list_head unclean_leb_list;
+ struct ubifs_mst_node *rcvrd_mst_node;
+ struct rb_root size_tree;
+ int remounting_rw;
+ int always_chk_crc;
+ struct ubifs_mount_opts mount_opts;
+
+#ifdef CONFIG_UBIFS_FS_DEBUG
+ struct ubifs_debug_info *dbg;
+#endif
+};
+
+extern spinlock_t ubifs_infos_lock;
+extern struct kmem_cache *ubifs_inode_slab;
+extern const struct super_operations ubifs_super_operations;
+extern const struct address_space_operations ubifs_file_address_operations;
+extern const struct file_operations ubifs_file_operations;
+extern const struct inode_operations ubifs_file_inode_operations;
+extern const struct file_operations ubifs_dir_operations;
+extern const struct inode_operations ubifs_dir_inode_operations;
+extern const struct inode_operations ubifs_symlink_inode_operations;
+extern struct backing_dev_info ubifs_backing_dev_info;
+extern struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT];
+
+/* io.c */
+void ubifs_ro_mode(struct ubifs_info *c, int err);
+int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len);
+int ubifs_wbuf_seek_nolock(struct ubifs_wbuf *wbuf, int lnum, int offs,
+ int dtype);
+int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf);
+int ubifs_read_node(const struct ubifs_info *c, void *buf, int type, int len,
+ int lnum, int offs);
+int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
+ int lnum, int offs);
+int ubifs_write_node(struct ubifs_info *c, void *node, int len, int lnum,
+ int offs, int dtype);
+int ubifs_check_node(const struct ubifs_info *c, const void *buf, int lnum,
+ int offs, int quiet, int must_chk_crc);
+void ubifs_prepare_node(struct ubifs_info *c, void *buf, int len, int pad);
+void ubifs_prep_grp_node(struct ubifs_info *c, void *node, int len, int last);
+int ubifs_io_init(struct ubifs_info *c);
+void ubifs_pad(const struct ubifs_info *c, void *buf, int pad);
+int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf);
+int ubifs_bg_wbufs_sync(struct ubifs_info *c);
+void ubifs_wbuf_add_ino_nolock(struct ubifs_wbuf *wbuf, ino_t inum);
+int ubifs_sync_wbufs_by_inode(struct ubifs_info *c, struct inode *inode);
+
+/* scan.c */
+struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
+ int offs, void *sbuf);
+void ubifs_scan_destroy(struct ubifs_scan_leb *sleb);
+int ubifs_scan_a_node(const struct ubifs_info *c, void *buf, int len, int lnum,
+ int offs, int quiet);
+struct ubifs_scan_leb *ubifs_start_scan(const struct ubifs_info *c, int lnum,
+ int offs, void *sbuf);
+void ubifs_end_scan(const struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ int lnum, int offs);
+int ubifs_add_snod(const struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ void *buf, int offs);
+void ubifs_scanned_corruption(const struct ubifs_info *c, int lnum, int offs,
+ void *buf);
+
+/* log.c */
+void ubifs_add_bud(struct ubifs_info *c, struct ubifs_bud *bud);
+void ubifs_create_buds_lists(struct ubifs_info *c);
+int ubifs_add_bud_to_log(struct ubifs_info *c, int jhead, int lnum, int offs);
+struct ubifs_bud *ubifs_search_bud(struct ubifs_info *c, int lnum);
+struct ubifs_wbuf *ubifs_get_wbuf(struct ubifs_info *c, int lnum);
+int ubifs_log_start_commit(struct ubifs_info *c, int *ltail_lnum);
+int ubifs_log_end_commit(struct ubifs_info *c, int new_ltail_lnum);
+int ubifs_log_post_commit(struct ubifs_info *c, int old_ltail_lnum);
+int ubifs_consolidate_log(struct ubifs_info *c);
+
+/* journal.c */
+int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
+ const struct qstr *nm, const struct inode *inode,
+ int deletion, int xent);
+int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
+ const union ubifs_key *key, const void *buf, int len);
+int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode);
+int ubifs_jnl_delete_inode(struct ubifs_info *c, const struct inode *inode);
+int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
+ const struct dentry *old_dentry,
+ const struct inode *new_dir,
+ const struct dentry *new_dentry, int sync);
+int ubifs_jnl_truncate(struct ubifs_info *c, const struct inode *inode,
+ loff_t old_size, loff_t new_size);
+int ubifs_jnl_delete_xattr(struct ubifs_info *c, const struct inode *host,
+ const struct inode *inode, const struct qstr *nm);
+int ubifs_jnl_change_xattr(struct ubifs_info *c, const struct inode *inode1,
+ const struct inode *inode2);
+
+/* budget.c */
+int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req);
+void ubifs_release_budget(struct ubifs_info *c, struct ubifs_budget_req *req);
+void ubifs_release_dirty_inode_budget(struct ubifs_info *c,
+ struct ubifs_inode *ui);
+int ubifs_budget_inode_op(struct ubifs_info *c, struct inode *inode,
+ struct ubifs_budget_req *req);
+void ubifs_release_ino_dirty(struct ubifs_info *c, struct inode *inode,
+ struct ubifs_budget_req *req);
+void ubifs_cancel_ino_op(struct ubifs_info *c, struct inode *inode,
+ struct ubifs_budget_req *req);
+long long ubifs_get_free_space(struct ubifs_info *c);
+long long ubifs_get_free_space_nolock(struct ubifs_info *c);
+int ubifs_calc_min_idx_lebs(struct ubifs_info *c);
+void ubifs_convert_page_budget(struct ubifs_info *c);
+long long ubifs_reported_space(const struct ubifs_info *c, long long free);
+long long ubifs_calc_available(const struct ubifs_info *c, int min_idx_lebs);
+
+/* find.c */
+int ubifs_find_free_space(struct ubifs_info *c, int min_space, int *free,
+ int squeeze);
+int ubifs_find_free_leb_for_idx(struct ubifs_info *c);
+int ubifs_find_dirty_leb(struct ubifs_info *c, struct ubifs_lprops *ret_lp,
+ int min_space, int pick_free);
+int ubifs_find_dirty_idx_leb(struct ubifs_info *c);
+int ubifs_save_dirty_idx_lnums(struct ubifs_info *c);
+
+/* tnc.c */
+int ubifs_lookup_level0(struct ubifs_info *c, const union ubifs_key *key,
+ struct ubifs_znode **zn, int *n);
+int ubifs_tnc_lookup_nm(struct ubifs_info *c, const union ubifs_key *key,
+ void *node, const struct qstr *nm);
+int ubifs_tnc_locate(struct ubifs_info *c, const union ubifs_key *key,
+ void *node, int *lnum, int *offs);
+int ubifs_tnc_add(struct ubifs_info *c, const union ubifs_key *key, int lnum,
+ int offs, int len);
+int ubifs_tnc_replace(struct ubifs_info *c, const union ubifs_key *key,
+ int old_lnum, int old_offs, int lnum, int offs, int len);
+int ubifs_tnc_add_nm(struct ubifs_info *c, const union ubifs_key *key,
+ int lnum, int offs, int len, const struct qstr *nm);
+int ubifs_tnc_remove(struct ubifs_info *c, const union ubifs_key *key);
+int ubifs_tnc_remove_nm(struct ubifs_info *c, const union ubifs_key *key,
+ const struct qstr *nm);
+int ubifs_tnc_remove_range(struct ubifs_info *c, union ubifs_key *from_key,
+ union ubifs_key *to_key);
+int ubifs_tnc_remove_ino(struct ubifs_info *c, ino_t inum);
+struct ubifs_dent_node *ubifs_tnc_next_ent(struct ubifs_info *c,
+ union ubifs_key *key,
+ const struct qstr *nm);
+void ubifs_tnc_close(struct ubifs_info *c);
+int ubifs_tnc_has_node(struct ubifs_info *c, union ubifs_key *key, int level,
+ int lnum, int offs, int is_idx);
+int ubifs_dirty_idx_node(struct ubifs_info *c, union ubifs_key *key, int level,
+ int lnum, int offs);
+/* Shared by tnc.c for tnc_commit.c */
+void destroy_old_idx(struct ubifs_info *c);
+int is_idx_node_in_tnc(struct ubifs_info *c, union ubifs_key *key, int level,
+ int lnum, int offs);
+int insert_old_idx_znode(struct ubifs_info *c, struct ubifs_znode *znode);
+int ubifs_tnc_get_bu_keys(struct ubifs_info *c, struct bu_info *bu);
+int ubifs_tnc_bulk_read(struct ubifs_info *c, struct bu_info *bu);
+
+/* tnc_misc.c */
+struct ubifs_znode *ubifs_tnc_levelorder_next(struct ubifs_znode *zr,
+ struct ubifs_znode *znode);
+int ubifs_search_zbranch(const struct ubifs_info *c,
+ const struct ubifs_znode *znode,
+ const union ubifs_key *key, int *n);
+struct ubifs_znode *ubifs_tnc_postorder_first(struct ubifs_znode *znode);
+struct ubifs_znode *ubifs_tnc_postorder_next(struct ubifs_znode *znode);
+long ubifs_destroy_tnc_subtree(struct ubifs_znode *zr);
+struct ubifs_znode *ubifs_load_znode(struct ubifs_info *c,
+ struct ubifs_zbranch *zbr,
+ struct ubifs_znode *parent, int iip);
+int ubifs_tnc_read_node(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ void *node);
+
+/* tnc_commit.c */
+int ubifs_tnc_start_commit(struct ubifs_info *c, struct ubifs_zbranch *zroot);
+int ubifs_tnc_end_commit(struct ubifs_info *c);
+
+/* shrinker.c */
+int ubifs_shrinker(int nr_to_scan, gfp_t gfp_mask);
+
+/* commit.c */
+int ubifs_bg_thread(void *info);
+void ubifs_commit_required(struct ubifs_info *c);
+void ubifs_request_bg_commit(struct ubifs_info *c);
+int ubifs_run_commit(struct ubifs_info *c);
+void ubifs_recovery_commit(struct ubifs_info *c);
+int ubifs_gc_should_commit(struct ubifs_info *c);
+void ubifs_wait_for_commit(struct ubifs_info *c);
+
+/* master.c */
+int ubifs_read_master(struct ubifs_info *c);
+int ubifs_write_master(struct ubifs_info *c);
+
+/* sb.c */
+int ubifs_read_superblock(struct ubifs_info *c);
+struct ubifs_sb_node *ubifs_read_sb_node(struct ubifs_info *c);
+int ubifs_write_sb_node(struct ubifs_info *c, struct ubifs_sb_node *sup);
+
+/* replay.c */
+int ubifs_validate_entry(struct ubifs_info *c,
+ const struct ubifs_dent_node *dent);
+int ubifs_replay_journal(struct ubifs_info *c);
+
+/* gc.c */
+int ubifs_garbage_collect(struct ubifs_info *c, int anyway);
+int ubifs_gc_start_commit(struct ubifs_info *c);
+int ubifs_gc_end_commit(struct ubifs_info *c);
+void ubifs_destroy_idx_gc(struct ubifs_info *c);
+int ubifs_get_idx_gc_leb(struct ubifs_info *c);
+int ubifs_garbage_collect_leb(struct ubifs_info *c, struct ubifs_lprops *lp);
+
+/* orphan.c */
+int ubifs_add_orphan(struct ubifs_info *c, ino_t inum);
+void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum);
+int ubifs_orphan_start_commit(struct ubifs_info *c);
+int ubifs_orphan_end_commit(struct ubifs_info *c);
+int ubifs_mount_orphans(struct ubifs_info *c, int unclean, int read_only);
+int ubifs_clear_orphans(struct ubifs_info *c);
+
+/* lpt.c */
+int ubifs_calc_lpt_geom(struct ubifs_info *c);
+int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first,
+ int *lpt_lebs, int *big_lpt);
+int ubifs_lpt_init(struct ubifs_info *c, int rd, int wr);
+struct ubifs_lprops *ubifs_lpt_lookup(struct ubifs_info *c, int lnum);
+struct ubifs_lprops *ubifs_lpt_lookup_dirty(struct ubifs_info *c, int lnum);
+int ubifs_lpt_scan_nolock(struct ubifs_info *c, int start_lnum, int end_lnum,
+ ubifs_lpt_scan_callback scan_cb, void *data);
+
+/* Shared by lpt.c for lpt_commit.c */
+void ubifs_pack_lsave(struct ubifs_info *c, void *buf, int *lsave);
+void ubifs_pack_ltab(struct ubifs_info *c, void *buf,
+ struct ubifs_lpt_lprops *ltab);
+void ubifs_pack_pnode(struct ubifs_info *c, void *buf,
+ struct ubifs_pnode *pnode);
+void ubifs_pack_nnode(struct ubifs_info *c, void *buf,
+ struct ubifs_nnode *nnode);
+struct ubifs_pnode *ubifs_get_pnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip);
+struct ubifs_nnode *ubifs_get_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip);
+int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip);
+void ubifs_add_lpt_dirt(struct ubifs_info *c, int lnum, int dirty);
+void ubifs_add_nnode_dirt(struct ubifs_info *c, struct ubifs_nnode *nnode);
+uint32_t ubifs_unpack_bits(uint8_t **addr, int *pos, int nrbits);
+struct ubifs_nnode *ubifs_first_nnode(struct ubifs_info *c, int *hght);
+/* Needed only in debugging code in lpt_commit.c */
+int ubifs_unpack_nnode(const struct ubifs_info *c, void *buf,
+ struct ubifs_nnode *nnode);
+
+/* lpt_commit.c */
+int ubifs_lpt_start_commit(struct ubifs_info *c);
+int ubifs_lpt_end_commit(struct ubifs_info *c);
+int ubifs_lpt_post_commit(struct ubifs_info *c);
+void ubifs_lpt_free(struct ubifs_info *c, int wr_only);
+
+/* lprops.c */
+const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
+ const struct ubifs_lprops *lp,
+ int free, int dirty, int flags,
+ int idx_gc_cnt);
+void ubifs_get_lp_stats(struct ubifs_info *c, struct ubifs_lp_stats *lst);
+void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops,
+ int cat);
+void ubifs_replace_cat(struct ubifs_info *c, struct ubifs_lprops *old_lprops,
+ struct ubifs_lprops *new_lprops);
+void ubifs_ensure_cat(struct ubifs_info *c, struct ubifs_lprops *lprops);
+int ubifs_categorize_lprops(const struct ubifs_info *c,
+ const struct ubifs_lprops *lprops);
+int ubifs_change_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean, int idx_gc_cnt);
+int ubifs_update_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean);
+int ubifs_read_one_lp(struct ubifs_info *c, int lnum, struct ubifs_lprops *lp);
+const struct ubifs_lprops *ubifs_fast_find_free(struct ubifs_info *c);
+const struct ubifs_lprops *ubifs_fast_find_empty(struct ubifs_info *c);
+const struct ubifs_lprops *ubifs_fast_find_freeable(struct ubifs_info *c);
+const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c);
+
+/* file.c */
+int ubifs_fsync(struct file *file, struct dentry *dentry, int datasync);
+int ubifs_setattr(struct dentry *dentry, struct iattr *attr);
+
+/* dir.c */
+struct inode *ubifs_new_inode(struct ubifs_info *c, const struct inode *dir,
+ int mode);
+int ubifs_getattr(struct vfsmount *mnt, struct dentry *dentry,
+ struct kstat *stat);
+
+/* xattr.c */
+int ubifs_setxattr(struct dentry *dentry, const char *name,
+ const void *value, size_t size, int flags);
+ssize_t ubifs_getxattr(struct dentry *dentry, const char *name, void *buf,
+ size_t size);
+ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size);
+int ubifs_removexattr(struct dentry *dentry, const char *name);
+
+/* super.c */
+struct inode *ubifs_iget(struct super_block *sb, unsigned long inum);
+int ubifs_iput(struct inode *inode);
+
+/* recovery.c */
+int ubifs_recover_master_node(struct ubifs_info *c);
+int ubifs_write_rcvrd_mst_node(struct ubifs_info *c);
+struct ubifs_scan_leb *ubifs_recover_leb(struct ubifs_info *c, int lnum,
+ int offs, void *sbuf, int grouped);
+struct ubifs_scan_leb *ubifs_recover_log_leb(struct ubifs_info *c, int lnum,
+ int offs, void *sbuf);
+int ubifs_recover_inl_heads(const struct ubifs_info *c, void *sbuf);
+int ubifs_clean_lebs(const struct ubifs_info *c, void *sbuf);
+int ubifs_rcvry_gc_commit(struct ubifs_info *c);
+int ubifs_recover_size_accum(struct ubifs_info *c, union ubifs_key *key,
+ int deletion, loff_t new_size);
+int ubifs_recover_size(struct ubifs_info *c);
+void ubifs_destroy_size_tree(struct ubifs_info *c);
+
+/* ioctl.c */
+long ubifs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
+void ubifs_set_inode_flags(struct inode *inode);
+#ifdef CONFIG_COMPAT
+long ubifs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
+#endif
+
+/* compressor.c */
+int __init ubifs_compressors_init(void);
+void __exit ubifs_compressors_exit(void);
+void ubifs_compress(const void *in_buf, int in_len, void *out_buf, int *out_len,
+ int *compr_type);
+int ubifs_decompress(const void *buf, int len, void *out, int *out_len,
+ int compr_type);
+
+#include "debug.h"
+#include "misc.h"
+#include "key.h"
+
+/* todo: Move these to a common U-Boot header */
+int lzo1x_decompress_safe(const unsigned char *in, size_t in_len,
+ unsigned char *out, size_t *out_len);
+int gunzip(void *dst, int dstlen, unsigned char *src, unsigned long *lenp);
+
+#endif /* !__UBIFS_H__ */
diff --git a/include/linux/math64.h b/include/linux/math64.h
new file mode 100644
index 0000000..6d760d7
--- /dev/null
+++ b/include/linux/math64.h
@@ -0,0 +1,85 @@
+#ifndef _LINUX_MATH64_H
+#define _LINUX_MATH64_H
+
+#include <linux/types.h>
+
+#if BITS_PER_LONG == 64
+
+/**
+ * div_u64_rem - unsigned 64bit divide with 32bit divisor with remainder
+ *
+ * This is commonly provided by 32bit archs to provide an optimized 64bit
+ * divide.
+ */
+static inline u64 div_u64_rem(u64 dividend, u32 divisor, u32 *remainder)
+{
+ *remainder = dividend % divisor;
+ return dividend / divisor;
+}
+
+/**
+ * div_s64_rem - signed 64bit divide with 32bit divisor with remainder
+ */
+static inline s64 div_s64_rem(s64 dividend, s32 divisor, s32 *remainder)
+{
+ *remainder = dividend % divisor;
+ return dividend / divisor;
+}
+
+/**
+ * div64_u64 - unsigned 64bit divide with 64bit divisor
+ */
+static inline u64 div64_u64(u64 dividend, u64 divisor)
+{
+ return dividend / divisor;
+}
+
+#elif BITS_PER_LONG == 32
+
+#ifndef div_u64_rem
+static inline u64 div_u64_rem(u64 dividend, u32 divisor, u32 *remainder)
+{
+ *remainder = do_div(dividend, divisor);
+ return dividend;
+}
+#endif
+
+#ifndef div_s64_rem
+extern s64 div_s64_rem(s64 dividend, s32 divisor, s32 *remainder);
+#endif
+
+#ifndef div64_u64
+extern u64 div64_u64(u64 dividend, u64 divisor);
+#endif
+
+#endif /* BITS_PER_LONG */
+
+/**
+ * div_u64 - unsigned 64bit divide with 32bit divisor
+ *
+ * This is the most common 64bit divide and should be used if possible,
+ * as many 32bit archs can optimize this variant better than a full 64bit
+ * divide.
+ */
+#ifndef div_u64
+static inline u64 div_u64(u64 dividend, u32 divisor)
+{
+ u32 remainder;
+ return div_u64_rem(dividend, divisor, &remainder);
+}
+#endif
+
+/**
+ * div_s64 - signed 64bit divide with 32bit divisor
+ */
+#ifndef div_s64
+static inline s64 div_s64(s64 dividend, s32 divisor)
+{
+ s32 remainder;
+ return div_s64_rem(dividend, divisor, &remainder);
+}
+#endif
+
+u32 iter_div_u64_rem(u64 dividend, u32 divisor, u64 *remainder);
+
+#endif /* _LINUX_MATH64_H */
diff --git a/include/ubi_uboot.h b/include/ubi_uboot.h
index b415219..74312ab 100644
--- a/include/ubi_uboot.h
+++ b/include/ubi_uboot.h
@@ -18,7 +18,12 @@
#include <malloc.h>
#include <div64.h>
#include <linux/crc32.h>
+#include <linux/types.h>
+#include <linux/list.h>
+#include <linux/rbtree.h>
+#include <linux/string.h>
#include <linux/mtd/mtd.h>
+#include <linux/mtd/ubi.h>
#ifdef CONFIG_CMD_ONENAND
#include <onenand_uboot.h>
@@ -193,7 +198,7 @@ static inline long IS_ERR(const void *ptr)
/* module */
#define THIS_MODULE 0
-#define try_module_get(...) 0
+#define try_module_get(...) 1
#define module_put(...) do { } while (0)
#define module_init(...)
#define module_exit(...)
@@ -206,7 +211,9 @@ static inline long IS_ERR(const void *ptr)
#define MODULE_AUTHOR(...)
#define MODULE_LICENSE(...)
+#ifndef __UBIFS_H__
#include "../drivers/mtd/ubi/ubi.h"
+#endif
/* functions */
extern int ubi_mtd_param_parse(const char *val, struct kernel_param *kp);