Resguardo inicial

linux-7.2.3/fs/ubifs/Kconfig

@@ -0,0 +1,101 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config UBIFS_FS
+	tristate "UBIFS file system support"
+	select CRC16
+	select CRC32
+	select CRYPTO if UBIFS_FS_ADVANCED_COMPR
+	select CRYPTO if UBIFS_FS_LZO
+	select CRYPTO if UBIFS_FS_ZLIB
+	select CRYPTO if UBIFS_FS_ZSTD
+	select CRYPTO_LZO if UBIFS_FS_LZO
+	select CRYPTO_DEFLATE if UBIFS_FS_ZLIB
+	select CRYPTO_ZSTD if UBIFS_FS_ZSTD
+	select CRYPTO_HASH_INFO
+	select UBIFS_FS_XATTR if FS_ENCRYPTION
+	select FS_ENCRYPTION_ALGS if FS_ENCRYPTION
+	depends on MTD_UBI
+	help
+	  UBIFS is a file system for flash devices which works on top of UBI.
+
+if UBIFS_FS
+
+config UBIFS_FS_ADVANCED_COMPR
+	bool "Advanced compression options"
+	help
+	  This option allows to explicitly choose which compressions, if any,
+	  are enabled in UBIFS. Removing compressors means inability to read
+	  existing file systems.
+
+	  If unsure, say 'N'.
+
+config UBIFS_FS_LZO
+	bool "LZO compression support" if UBIFS_FS_ADVANCED_COMPR
+	default y
+	help
+	   LZO compressor is generally faster than zlib but compresses worse.
+	   Say 'Y' if unsure.
+
+config UBIFS_FS_ZLIB
+	bool "ZLIB compression support" if UBIFS_FS_ADVANCED_COMPR
+	default y
+	help
+	  Zlib compresses better than LZO but it is slower. Say 'Y' if unsure.
+
+config UBIFS_FS_ZSTD
+	bool "ZSTD compression support" if UBIFS_FS_ADVANCED_COMPR
+	depends on UBIFS_FS
+	default y
+	help
+	  ZSTD compresses is a big win in speed over Zlib and
+	  in compression ratio over LZO. Say 'Y' if unsure.
+
+config UBIFS_ATIME_SUPPORT
+	bool "Access time support"
+	default n
+	help
+	  Originally UBIFS did not support atime, because it looked like a bad idea due
+	  increased flash wear. This option adds atime support and it is disabled by default
+	  to preserve the old behavior. If you enable this option, UBIFS starts updating atime,
+	  which means that file-system read operations will cause writes (inode atime
+	  updates). This may affect file-system performance and increase flash device wear,
+	  so be careful. How often atime is updated depends on the selected strategy:
+	  strictatime is the "heavy", relatime is "lighter", etc.
+
+	  If unsure, say 'N'
+
+config UBIFS_FS_XATTR
+	bool "UBIFS XATTR support"
+	default y
+	help
+	  Saying Y here includes support for extended attributes (xattrs).
+	  Xattrs are name:value pairs associated with inodes by
+	  the kernel or by users (see the attr(5) manual page).
+
+	  If unsure, say Y.
+
+config UBIFS_FS_SECURITY
+	bool "UBIFS Security Labels"
+	depends on UBIFS_FS_XATTR
+	default y
+	help
+	  Security labels provide an access control facility to support Linux
+	  Security Models (LSMs) accepted by AppArmor, SELinux, Smack and TOMOYO
+	  Linux. This option enables an extended attribute handler for file
+	  security labels in the ubifs filesystem, so that it requires enabling
+	  the extended attribute support in advance.
+
+	  If you are not using a security module, say N.
+
+config UBIFS_FS_AUTHENTICATION
+	bool "UBIFS authentication support"
+	select KEYS
+	select CRYPTO_HMAC
+	select SYSTEM_DATA_VERIFICATION
+	help
+	  Enable authentication support for UBIFS. This feature offers protection
+	  against offline changes for both data and metadata of the filesystem.
+	  If you say yes here you should also select a hashing algorithm such as
+	  sha256, these are not selected automatically since there are many
+	  different options.
+
+endif # UBIFS_FS

linux-7.2.3/fs/ubifs/Makefile

@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_UBIFS_FS) += ubifs.o
+
+ubifs-y += shrinker.o journal.o file.o dir.o super.o sb.o io.o
+ubifs-y += tnc.o master.o scan.o replay.o log.o commit.o gc.o orphan.o
+ubifs-y += budget.o find.o tnc_commit.o compress.o lpt.o lprops.o
+ubifs-y += recovery.o ioctl.o lpt_commit.o tnc_misc.o debug.o
+ubifs-y += misc.o sysfs.o
+ubifs-$(CONFIG_FS_ENCRYPTION) += crypto.o
+ubifs-$(CONFIG_UBIFS_FS_XATTR) += xattr.o
+ubifs-$(CONFIG_UBIFS_FS_AUTHENTICATION) += auth.o

linux-7.2.3/fs/ubifs/auth.c

@@ -0,0 +1,530 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2018 Pengutronix, Sascha Hauer <s.hauer@pengutronix.de>
+ */
+
+/*
+ * This file implements various helper functions for UBIFS authentication support
+ */
+
+#include <linux/verification.h>
+#include <crypto/hash.h>
+#include <crypto/utils.h>
+#include <keys/user-type.h>
+#include <keys/asymmetric-type.h>
+
+#include "ubifs.h"
+
+/**
+ * __ubifs_node_calc_hash - calculate the hash of a UBIFS node
+ * @c: UBIFS file-system description object
+ * @node: the node to calculate a hash for
+ * @hash: the returned hash
+ *
+ * Returns 0 for success or a negative error code otherwise.
+ */
+int __ubifs_node_calc_hash(const struct ubifs_info *c, const void *node,
+			    u8 *hash)
+{
+	const struct ubifs_ch *ch = node;
+
+	return crypto_shash_tfm_digest(c->hash_tfm, node, le32_to_cpu(ch->len),
+				       hash);
+}
+
+/**
+ * ubifs_hash_calc_hmac - calculate a HMAC from a hash
+ * @c: UBIFS file-system description object
+ * @hash: the node to calculate a HMAC for
+ * @hmac: the returned HMAC
+ *
+ * Returns 0 for success or a negative error code otherwise.
+ */
+static int ubifs_hash_calc_hmac(const struct ubifs_info *c, const u8 *hash,
+				 u8 *hmac)
+{
+	return crypto_shash_tfm_digest(c->hmac_tfm, hash, c->hash_len, hmac);
+}
+
+/**
+ * ubifs_prepare_auth_node - Prepare an authentication node
+ * @c: UBIFS file-system description object
+ * @node: the node to calculate a hash for
+ * @inhash: input hash of previous nodes
+ *
+ * This function prepares an authentication node for writing onto flash.
+ * It creates a HMAC from the given input hash and writes it to the node.
+ *
+ * Returns 0 for success or a negative error code otherwise.
+ */
+int ubifs_prepare_auth_node(struct ubifs_info *c, void *node,
+			     struct shash_desc *inhash)
+{
+	struct ubifs_auth_node *auth = node;
+	u8 hash[UBIFS_HASH_ARR_SZ];
+	int err;
+
+	{
+		SHASH_DESC_ON_STACK(hash_desc, c->hash_tfm);
+
+		hash_desc->tfm = c->hash_tfm;
+		ubifs_shash_copy_state(c, inhash, hash_desc);
+
+		err = crypto_shash_final(hash_desc, hash);
+		if (err)
+			return err;
+	}
+
+	err = ubifs_hash_calc_hmac(c, hash, auth->hmac);
+	if (err)
+		return err;
+
+	auth->ch.node_type = UBIFS_AUTH_NODE;
+	ubifs_prepare_node(c, auth, ubifs_auth_node_sz(c), 0);
+	return 0;
+}
+
+static struct shash_desc *ubifs_get_desc(const struct ubifs_info *c,
+					 struct crypto_shash *tfm)
+{
+	struct shash_desc *desc;
+	int err;
+
+	if (!ubifs_authenticated(c))
+		return NULL;
+
+	desc = kmalloc(sizeof(*desc) + crypto_shash_descsize(tfm), GFP_KERNEL);
+	if (!desc)
+		return ERR_PTR(-ENOMEM);
+
+	desc->tfm = tfm;
+
+	err = crypto_shash_init(desc);
+	if (err) {
+		kfree(desc);
+		return ERR_PTR(err);
+	}
+
+	return desc;
+}
+
+/**
+ * __ubifs_hash_get_desc - get a descriptor suitable for hashing a node
+ * @c: UBIFS file-system description object
+ *
+ * This function returns a descriptor suitable for hashing a node. Free after use
+ * with kfree.
+ */
+struct shash_desc *__ubifs_hash_get_desc(const struct ubifs_info *c)
+{
+	return ubifs_get_desc(c, c->hash_tfm);
+}
+
+/**
+ * ubifs_bad_hash - Report hash mismatches
+ * @c: UBIFS file-system description object
+ * @node: the node
+ * @hash: the expected hash
+ * @lnum: the LEB @node was read from
+ * @offs: offset in LEB @node was read from
+ *
+ * This function reports a hash mismatch when a node has a different hash than
+ * expected.
+ */
+void ubifs_bad_hash(const struct ubifs_info *c, const void *node, const u8 *hash,
+		    int lnum, int offs)
+{
+	int len = min(c->hash_len, 20);
+	int cropped = len != c->hash_len;
+	const char *cont = cropped ? "..." : "";
+
+	u8 calc[UBIFS_HASH_ARR_SZ];
+
+	__ubifs_node_calc_hash(c, node, calc);
+
+	ubifs_err(c, "hash mismatch on node at LEB %d:%d", lnum, offs);
+	ubifs_err(c, "hash expected:   %*ph%s", len, hash, cont);
+	ubifs_err(c, "hash calculated: %*ph%s", len, calc, cont);
+}
+
+/**
+ * __ubifs_node_check_hash - check the hash of a node against given hash
+ * @c: UBIFS file-system description object
+ * @node: the node
+ * @expected: the expected hash
+ *
+ * This function calculates a hash over a node and compares it to the given hash.
+ * Returns 0 if both hashes are equal or authentication is disabled, otherwise a
+ * negative error code is returned.
+ */
+int __ubifs_node_check_hash(const struct ubifs_info *c, const void *node,
+			    const u8 *expected)
+{
+	u8 calc[UBIFS_HASH_ARR_SZ];
+	int err;
+
+	err = __ubifs_node_calc_hash(c, node, calc);
+	if (err)
+		return err;
+
+	if (ubifs_check_hash(c, expected, calc))
+		return -EPERM;
+
+	return 0;
+}
+
+/**
+ * ubifs_sb_verify_signature - verify the signature of a superblock
+ * @c: UBIFS file-system description object
+ * @sup: The superblock node
+ *
+ * To support offline signed images the superblock can be signed with a
+ * PKCS#7 signature. The signature is placed directly behind the superblock
+ * node in an ubifs_sig_node.
+ *
+ * Returns 0 when the signature can be successfully verified or a negative
+ * error code if not.
+ */
+int ubifs_sb_verify_signature(struct ubifs_info *c,
+			      const struct ubifs_sb_node *sup)
+{
+	int err;
+	struct ubifs_scan_leb *sleb;
+	struct ubifs_scan_node *snod;
+	const struct ubifs_sig_node *signode;
+
+	sleb = ubifs_scan(c, UBIFS_SB_LNUM, UBIFS_SB_NODE_SZ, c->sbuf, 0);
+	if (IS_ERR(sleb)) {
+		err = PTR_ERR(sleb);
+		return err;
+	}
+
+	if (sleb->nodes_cnt == 0) {
+		ubifs_err(c, "Unable to find signature node");
+		err = -EINVAL;
+		goto out_destroy;
+	}
+
+	snod = list_first_entry(&sleb->nodes, struct ubifs_scan_node, list);
+
+	if (snod->type != UBIFS_SIG_NODE) {
+		ubifs_err(c, "Signature node is of wrong type");
+		err = -EINVAL;
+		goto out_destroy;
+	}
+
+	signode = snod->node;
+
+	if (le32_to_cpu(signode->len) > snod->len + sizeof(struct ubifs_sig_node)) {
+		ubifs_err(c, "invalid signature len %d", le32_to_cpu(signode->len));
+		err = -EINVAL;
+		goto out_destroy;
+	}
+
+	if (le32_to_cpu(signode->type) != UBIFS_SIGNATURE_TYPE_PKCS7) {
+		ubifs_err(c, "Signature type %d is not supported\n",
+			  le32_to_cpu(signode->type));
+		err = -EINVAL;
+		goto out_destroy;
+	}
+
+	err = verify_pkcs7_signature(sup, sizeof(struct ubifs_sb_node),
+				     signode->sig, le32_to_cpu(signode->len),
+				     NULL, VERIFYING_UNSPECIFIED_SIGNATURE,
+				     NULL, NULL);
+
+	if (err)
+		ubifs_err(c, "Failed to verify signature");
+	else
+		ubifs_msg(c, "Successfully verified super block signature");
+
+out_destroy:
+	ubifs_scan_destroy(sleb);
+
+	return err;
+}
+
+/**
+ * ubifs_init_authentication - initialize UBIFS authentication support
+ * @c: UBIFS file-system description object
+ *
+ * This function returns 0 for success or a negative error code otherwise.
+ */
+int ubifs_init_authentication(struct ubifs_info *c)
+{
+	struct key *keyring_key;
+	const struct user_key_payload *ukp;
+	int err;
+	char hmac_name[CRYPTO_MAX_ALG_NAME];
+
+	if (!c->auth_hash_name) {
+		ubifs_err(c, "authentication hash name needed with authentication");
+		return -EINVAL;
+	}
+
+	c->auth_hash_algo = match_string(hash_algo_name, HASH_ALGO__LAST,
+					 c->auth_hash_name);
+	if ((int)c->auth_hash_algo < 0) {
+		ubifs_err(c, "Unknown hash algo %s specified",
+			  c->auth_hash_name);
+		return -EINVAL;
+	}
+
+	snprintf(hmac_name, CRYPTO_MAX_ALG_NAME, "hmac(%s)",
+		 c->auth_hash_name);
+
+	keyring_key = request_key(&key_type_logon, c->auth_key_name, NULL);
+
+	if (IS_ERR(keyring_key)) {
+		ubifs_err(c, "Failed to request key: %ld",
+			  PTR_ERR(keyring_key));
+		return PTR_ERR(keyring_key);
+	}
+
+	down_read(&keyring_key->sem);
+
+	if (keyring_key->type != &key_type_logon) {
+		ubifs_err(c, "key type must be logon");
+		err = -ENOKEY;
+		goto out;
+	}
+
+	ukp = user_key_payload_locked(keyring_key);
+	if (!ukp) {
+		/* key was revoked before we acquired its semaphore */
+		err = -EKEYREVOKED;
+		goto out;
+	}
+
+	c->hash_tfm = crypto_alloc_shash(c->auth_hash_name, 0, 0);
+	if (IS_ERR(c->hash_tfm)) {
+		err = PTR_ERR(c->hash_tfm);
+		ubifs_err(c, "Can not allocate %s: %d",
+			  c->auth_hash_name, err);
+		goto out;
+	}
+
+	c->hash_len = crypto_shash_digestsize(c->hash_tfm);
+	if (c->hash_len > UBIFS_HASH_ARR_SZ) {
+		ubifs_err(c, "hash %s is bigger than maximum allowed hash size (%d > %d)",
+			  c->auth_hash_name, c->hash_len, UBIFS_HASH_ARR_SZ);
+		err = -EINVAL;
+		goto out_free_hash;
+	}
+
+	c->hmac_tfm = crypto_alloc_shash(hmac_name, 0, 0);
+	if (IS_ERR(c->hmac_tfm)) {
+		err = PTR_ERR(c->hmac_tfm);
+		ubifs_err(c, "Can not allocate %s: %d", hmac_name, err);
+		goto out_free_hash;
+	}
+
+	c->hmac_desc_len = crypto_shash_digestsize(c->hmac_tfm);
+	if (c->hmac_desc_len > UBIFS_HMAC_ARR_SZ) {
+		ubifs_err(c, "hmac %s is bigger than maximum allowed hmac size (%d > %d)",
+			  hmac_name, c->hmac_desc_len, UBIFS_HMAC_ARR_SZ);
+		err = -EINVAL;
+		goto out_free_hmac;
+	}
+
+	err = crypto_shash_setkey(c->hmac_tfm, ukp->data, ukp->datalen);
+	if (err)
+		goto out_free_hmac;
+
+	c->authenticated = true;
+
+	c->log_hash = ubifs_hash_get_desc(c);
+	if (IS_ERR(c->log_hash)) {
+		err = PTR_ERR(c->log_hash);
+		goto out_free_hmac;
+	}
+
+	err = 0;
+
+out_free_hmac:
+	if (err)
+		crypto_free_shash(c->hmac_tfm);
+out_free_hash:
+	if (err)
+		crypto_free_shash(c->hash_tfm);
+out:
+	up_read(&keyring_key->sem);
+	key_put(keyring_key);
+
+	return err;
+}
+
+/**
+ * __ubifs_exit_authentication - release resource
+ * @c: UBIFS file-system description object
+ *
+ * This function releases the authentication related resources.
+ */
+void __ubifs_exit_authentication(struct ubifs_info *c)
+{
+	if (!ubifs_authenticated(c))
+		return;
+
+	crypto_free_shash(c->hmac_tfm);
+	crypto_free_shash(c->hash_tfm);
+	kfree(c->log_hash);
+}
+
+/**
+ * ubifs_node_calc_hmac - calculate the HMAC of a UBIFS node
+ * @c: UBIFS file-system description object
+ * @node: the node to insert a HMAC into.
+ * @len: the length of the node
+ * @ofs_hmac: the offset in the node where the HMAC is inserted
+ * @hmac: returned HMAC
+ *
+ * This function calculates a HMAC of a UBIFS node. The HMAC is expected to be
+ * embedded into the node, so this area is not covered by the HMAC. Also not
+ * covered is the UBIFS_NODE_MAGIC and the CRC of the node.
+ */
+static int ubifs_node_calc_hmac(const struct ubifs_info *c, const void *node,
+				int len, int ofs_hmac, void *hmac)
+{
+	SHASH_DESC_ON_STACK(shash, c->hmac_tfm);
+	int hmac_len = c->hmac_desc_len;
+	int err;
+
+	ubifs_assert(c, ofs_hmac > 8);
+	ubifs_assert(c, ofs_hmac + hmac_len < len);
+
+	shash->tfm = c->hmac_tfm;
+
+	err = crypto_shash_init(shash);
+	if (err)
+		return err;
+
+	/* behind common node header CRC up to HMAC begin */
+	err = crypto_shash_update(shash, node + 8, ofs_hmac - 8);
+	if (err < 0)
+		return err;
+
+	/* behind HMAC, if any */
+	if (len - ofs_hmac - hmac_len > 0) {
+		err = crypto_shash_update(shash, node + ofs_hmac + hmac_len,
+			    len - ofs_hmac - hmac_len);
+		if (err < 0)
+			return err;
+	}
+
+	return crypto_shash_final(shash, hmac);
+}
+
+/**
+ * __ubifs_node_insert_hmac - insert a HMAC into a UBIFS node
+ * @c: UBIFS file-system description object
+ * @node: the node to insert a HMAC into.
+ * @len: the length of the node
+ * @ofs_hmac: the offset in the node where the HMAC is inserted
+ *
+ * This function inserts a HMAC at offset @ofs_hmac into the node given in
+ * @node.
+ *
+ * This function returns 0 for success or a negative error code otherwise.
+ */
+int __ubifs_node_insert_hmac(const struct ubifs_info *c, void *node, int len,
+			    int ofs_hmac)
+{
+	return ubifs_node_calc_hmac(c, node, len, ofs_hmac, node + ofs_hmac);
+}
+
+/**
+ * __ubifs_node_verify_hmac - verify the HMAC of UBIFS node
+ * @c: UBIFS file-system description object
+ * @node: the node to insert a HMAC into.
+ * @len: the length of the node
+ * @ofs_hmac: the offset in the node where the HMAC is inserted
+ *
+ * This function verifies the HMAC at offset @ofs_hmac of the node given in
+ * @node. Returns 0 if successful or a negative error code otherwise.
+ */
+int __ubifs_node_verify_hmac(const struct ubifs_info *c, const void *node,
+			     int len, int ofs_hmac)
+{
+	int hmac_len = c->hmac_desc_len;
+	u8 *hmac;
+	int err;
+
+	hmac = kmalloc(hmac_len, GFP_NOFS);
+	if (!hmac)
+		return -ENOMEM;
+
+	err = ubifs_node_calc_hmac(c, node, len, ofs_hmac, hmac);
+	if (err) {
+		kfree(hmac);
+		return err;
+	}
+
+	err = crypto_memneq(hmac, node + ofs_hmac, hmac_len);
+
+	kfree(hmac);
+
+	if (!err)
+		return 0;
+
+	return -EPERM;
+}
+
+int __ubifs_shash_copy_state(const struct ubifs_info *c, struct shash_desc *src,
+			     struct shash_desc *target)
+{
+	u8 *state;
+	int err;
+
+	state = kmalloc(crypto_shash_descsize(src->tfm), GFP_NOFS);
+	if (!state)
+		return -ENOMEM;
+
+	err = crypto_shash_export(src, state);
+	if (err)
+		goto out;
+
+	err = crypto_shash_import(target, state);
+
+out:
+	kfree(state);
+
+	return err;
+}
+
+/**
+ * ubifs_hmac_wkm - Create a HMAC of the well known message
+ * @c: UBIFS file-system description object
+ * @hmac: The HMAC of the well known message
+ *
+ * This function creates a HMAC of a well known message. This is used
+ * to check if the provided key is suitable to authenticate a UBIFS
+ * image. This is only a convenience to the user to provide a better
+ * error message when the wrong key is provided.
+ *
+ * This function returns 0 for success or a negative error code otherwise.
+ */
+int ubifs_hmac_wkm(struct ubifs_info *c, u8 *hmac)
+{
+	const char well_known_message[] = "UBIFS";
+
+	if (!ubifs_authenticated(c))
+		return 0;
+
+	return crypto_shash_tfm_digest(c->hmac_tfm, well_known_message,
+				       sizeof(well_known_message) - 1, hmac);
+}
+
+/*
+ * ubifs_hmac_zero - test if a HMAC is zero
+ * @c: UBIFS file-system description object
+ * @hmac: the HMAC to test
+ *
+ * This function tests if a HMAC is zero and returns true if it is
+ * and false otherwise.
+ */
+bool ubifs_hmac_zero(struct ubifs_info *c, const u8 *hmac)
+{
+	return !memchr_inv(hmac, 0, c->hmac_desc_len);
+}

linux-7.2.3/fs/ubifs/budget.c

@@ -0,0 +1,714 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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/writeback.h>
+#include <linux/math64.h>
+
+/*
+ * When pessimistic budget calculations say that there is no enough space,
+ * UBIFS starts writing back dirty inodes and pages, doing garbage collection,
+ * or committing. The below constant defines maximum number of times UBIFS
+ * repeats the operations.
+ */
+#define MAX_MKSPC_RETRIES 3
+
+/*
+ * The below constant defines amount of dirty pages which should be written
+ * back at when trying to shrink the liability.
+ */
+#define NR_TO_WRITE 16
+
+/**
+ * shrink_liability - write-back some dirty pages/inodes.
+ * @c: UBIFS file-system description object
+ * @nr_to_write: how many dirty pages to write-back
+ *
+ * This function shrinks UBIFS liability by means of writing back some amount
+ * of dirty inodes and their pages.
+ *
+ * Note, this function synchronizes even VFS inodes which are locked
+ * (@i_mutex) by the caller of the budgeting function, because write-back does
+ * not touch @i_mutex.
+ */
+static void shrink_liability(struct ubifs_info *c, int nr_to_write)
+{
+	down_read(&c->vfs_sb->s_umount);
+	writeback_inodes_sb_nr(c->vfs_sb, nr_to_write, WB_REASON_FS_FREE_SPACE);
+	up_read(&c->vfs_sb->s_umount);
+}
+
+/**
+ * run_gc - run garbage collector.
+ * @c: UBIFS file-system description object
+ *
+ * This function runs garbage collector to make some more free space. Returns
+ * zero if a free LEB has been produced, %-EAGAIN if commit is required, and a
+ * negative error code in case of failure.
+ */
+static int run_gc(struct ubifs_info *c)
+{
+	int lnum;
+
+	/* Make some free space by garbage-collecting dirty space */
+	down_read(&c->commit_sem);
+	lnum = ubifs_garbage_collect(c, 1);
+	up_read(&c->commit_sem);
+	if (lnum < 0)
+		return lnum;
+
+	/* GC freed one LEB, return it to lprops */
+	dbg_budg("GC freed LEB %d", lnum);
+	return ubifs_return_leb(c, lnum);
+}
+
+/**
+ * get_liability - calculate current liability.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates and returns current UBIFS liability, i.e. the
+ * amount of bytes UBIFS has "promised" to write to the media.
+ */
+static long long get_liability(struct ubifs_info *c)
+{
+	long long liab;
+
+	spin_lock(&c->space_lock);
+	liab = c->bi.idx_growth + c->bi.data_growth + c->bi.dd_growth;
+	spin_unlock(&c->space_lock);
+	return liab;
+}
+
+/**
+ * make_free_space - make more free space on the file-system.
+ * @c: UBIFS file-system description object
+ *
+ * This function is called when an operation cannot be budgeted because there
+ * is supposedly no free space. But in most cases there is some free space:
+ *   o budgeting is pessimistic, so it always budgets more than it is actually
+ *     needed, so shrinking the liability is one way to make free space - the
+ *     cached data will take less space then it was budgeted for;
+ *   o GC may turn some dark space into free space (budgeting treats dark space
+ *     as not available);
+ *   o commit may free some LEB, i.e., turn freeable LEBs into free LEBs.
+ *
+ * So this function tries to do the above. Returns %-EAGAIN if some free space
+ * was presumably made and the caller has to re-try budgeting the operation.
+ * Returns %-ENOSPC if it couldn't do more free space, and other negative error
+ * codes on failures.
+ */
+static int make_free_space(struct ubifs_info *c)
+{
+	int err, retries = 0;
+	long long liab1, liab2;
+
+	do {
+		liab1 = get_liability(c);
+		/*
+		 * We probably have some dirty pages or inodes (liability), try
+		 * to write them back.
+		 */
+		dbg_budg("liability %lld, run write-back", liab1);
+		shrink_liability(c, NR_TO_WRITE);
+
+		liab2 = get_liability(c);
+		if (liab2 < liab1)
+			return -EAGAIN;
+
+		dbg_budg("new liability %lld (not shrunk)", liab2);
+
+		/* Liability did not shrink again, try GC */
+		dbg_budg("Run GC");
+		err = run_gc(c);
+		if (!err)
+			return -EAGAIN;
+
+		if (err != -EAGAIN && err != -ENOSPC)
+			/* Some real error happened */
+			return err;
+
+		dbg_budg("Run commit (retries %d)", retries);
+		err = ubifs_run_commit(c);
+		if (err)
+			return err;
+	} while (retries++ < MAX_MKSPC_RETRIES);
+
+	return -ENOSPC;
+}
+
+/**
+ * ubifs_calc_min_idx_lebs - calculate amount of LEBs for the index.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates and returns the number of LEBs which should be kept
+ * for index usage.
+ */
+int ubifs_calc_min_idx_lebs(struct ubifs_info *c)
+{
+	int idx_lebs;
+	long long idx_size;
+
+	idx_size = c->bi.old_idx_sz + c->bi.idx_growth + c->bi.uncommitted_idx;
+	/* And make sure we have thrice the index size of space reserved */
+	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_lebs = div_u64(idx_size + c->idx_leb_size - 1, c->idx_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_calc_available - calculate available FS space.
+ * @c: UBIFS file-system description object
+ * @min_idx_lebs: minimum number of LEBs reserved for the index
+ *
+ * This function calculates and returns amount of FS space available for use.
+ */
+long long ubifs_calc_available(const struct ubifs_info *c, int min_idx_lebs)
+{
+	int subtract_lebs;
+	long long available;
+
+	available = c->main_bytes - c->lst.total_used;
+
+	/*
+	 * Now 'available' contains theoretically available flash space
+	 * assuming there is no index, so we have to subtract the space which
+	 * is reserved for the index.
+	 */
+	subtract_lebs = min_idx_lebs;
+
+	/* Take into account that GC reserves one LEB for its own needs */
+	subtract_lebs += 1;
+
+	/*
+	 * Since different write types go to different heads, we should
+	 * reserve one leb for each head.
+	 */
+	subtract_lebs += c->jhead_cnt;
+
+	/* We also reserve one LEB for deletions, which bypass budgeting */
+	subtract_lebs += 1;
+
+	available -= (long long)subtract_lebs * c->leb_size;
+
+	/* Subtract the dead space which is not available for use */
+	available -= c->lst.total_dead;
+
+	/*
+	 * Subtract dark space, which might or might not be usable - it depends
+	 * on the data which we have on the media and which will be written. If
+	 * this is a lot of uncompressed or not-compressible data, the dark
+	 * space cannot be used.
+	 */
+	available -= c->lst.total_dark;
+
+	/*
+	 * However, there is more dark space. The index may be bigger than
+	 * @min_idx_lebs. Those extra LEBs are assumed to be available, but
+	 * their dark space is not included in total_dark, so it is subtracted
+	 * here.
+	 */
+	if (c->lst.idx_lebs > min_idx_lebs) {
+		subtract_lebs = c->lst.idx_lebs - min_idx_lebs;
+		available -= subtract_lebs * c->dark_wm;
+	}
+
+	/* The calculations are rough and may end up with a negative number */
+	return available > 0 ? available : 0;
+}
+
+/**
+ * can_use_rp - check whether the user is allowed to use reserved pool.
+ * @c: UBIFS file-system description object
+ *
+ * UBIFS has so-called "reserved pool" which is flash space reserved
+ * for the superuser and for uses whose UID/GID is recorded in UBIFS superblock.
+ * This function checks whether current user is allowed to use reserved pool.
+ * Returns %1  current user is allowed to use reserved pool and %0 otherwise.
+ */
+static int can_use_rp(struct ubifs_info *c)
+{
+	if (uid_eq(current_fsuid(), c->rp_uid) || capable(CAP_SYS_RESOURCE) ||
+	    (!gid_eq(c->rp_gid, GLOBAL_ROOT_GID) && in_group_p(c->rp_gid)))
+		return 1;
+	return 0;
+}
+
+/**
+ * do_budget_space - reserve flash space for index and data growth.
+ * @c: UBIFS file-system description object
+ *
+ * This function makes sure UBIFS has enough free LEBs for index growth and
+ * data.
+ *
+ * When budgeting index space, UBIFS reserves thrice as many LEBs as the index
+ * would take if it was consolidated and written to the flash. This guarantees
+ * that the "in-the-gaps" commit method always succeeds and UBIFS will always
+ * be able to commit dirty index. So this function basically adds amount of
+ * budgeted index space to the size of the current index, multiplies this by 3,
+ * and makes sure this does not exceed the amount of free LEBs.
+ *
+ * Notes about @c->bi.min_idx_lebs and @c->lst.idx_lebs variables:
+ * o @c->lst.idx_lebs is the number of LEBs the index currently uses. It might
+ *    be large, because UBIFS does not do any index consolidation as long as
+ *    there is free space. IOW, the index may take a lot of LEBs, but the LEBs
+ *    will contain a lot of dirt.
+ * o @c->bi.min_idx_lebs is the number of LEBS the index presumably takes. IOW,
+ *    the index may be consolidated to take up to @c->bi.min_idx_lebs LEBs.
+ *
+ * This function returns zero in case of success, and %-ENOSPC in case of
+ * failure.
+ */
+static int do_budget_space(struct ubifs_info *c)
+{
+	long long outstanding, available;
+	int lebs, rsvd_idx_lebs, min_idx_lebs;
+
+	/* First budget index space */
+	min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+
+	/* Now 'min_idx_lebs' contains number of LEBs to reserve */
+	if (min_idx_lebs > c->lst.idx_lebs)
+		rsvd_idx_lebs = min_idx_lebs - c->lst.idx_lebs;
+	else
+		rsvd_idx_lebs = 0;
+
+	/*
+	 * The number of LEBs that are available to be used by the index is:
+	 *
+	 *    @c->lst.empty_lebs + @c->freeable_cnt + @c->idx_gc_cnt -
+	 *    @c->lst.taken_empty_lebs
+	 *
+	 * @c->lst.empty_lebs are available because they are empty.
+	 * @c->freeable_cnt are available because they contain only free and
+	 * dirty space, @c->idx_gc_cnt are available because they are index
+	 * LEBs that have been garbage collected and are awaiting the commit
+	 * before they can be used. And the in-the-gaps method will grab these
+	 * if it needs them. @c->lst.taken_empty_lebs are empty LEBs that have
+	 * already been allocated for some purpose.
+	 *
+	 * Note, @c->idx_gc_cnt is included to both @c->lst.empty_lebs (because
+	 * these LEBs are empty) and to @c->lst.taken_empty_lebs (because they
+	 * are taken until after the commit).
+	 *
+	 * Note, @c->lst.taken_empty_lebs may temporarily be higher by one
+	 * because of the way we serialize LEB allocations and budgeting. See a
+	 * comment in 'ubifs_find_free_space()'.
+	 */
+	lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt -
+	       c->lst.taken_empty_lebs;
+	if (unlikely(rsvd_idx_lebs > lebs)) {
+		dbg_budg("out of indexing space: min_idx_lebs %d (old %d), rsvd_idx_lebs %d",
+			 min_idx_lebs, c->bi.min_idx_lebs, rsvd_idx_lebs);
+		return -ENOSPC;
+	}
+
+	available = ubifs_calc_available(c, min_idx_lebs);
+	outstanding = c->bi.data_growth + c->bi.dd_growth;
+
+	if (unlikely(available < outstanding)) {
+		dbg_budg("out of data space: available %lld, outstanding %lld",
+			 available, outstanding);
+		return -ENOSPC;
+	}
+
+	if (available - outstanding <= c->rp_size && !can_use_rp(c))
+		return -ENOSPC;
+
+	c->bi.min_idx_lebs = min_idx_lebs;
+	return 0;
+}
+
+/**
+ * calc_idx_growth - calculate approximate index growth from budgeting request.
+ * @c: UBIFS file-system description object
+ * @req: budgeting request
+ *
+ * For now we assume each new node adds one znode. But this is rather poor
+ * approximation, though.
+ */
+static int calc_idx_growth(const struct ubifs_info *c,
+			   const struct ubifs_budget_req *req)
+{
+	int znodes;
+
+	znodes = req->new_ino + (req->new_page << UBIFS_BLOCKS_PER_PAGE_SHIFT) +
+		 req->new_dent;
+	return znodes * c->max_idx_node_sz;
+}
+
+/**
+ * calc_data_growth - calculate approximate amount of new data from budgeting
+ * request.
+ * @c: UBIFS file-system description object
+ * @req: budgeting request
+ */
+static int calc_data_growth(const struct ubifs_info *c,
+			    const struct ubifs_budget_req *req)
+{
+	int data_growth;
+
+	data_growth = req->new_ino  ? c->bi.inode_budget : 0;
+	if (req->new_page)
+		data_growth += c->bi.page_budget;
+	if (req->new_dent)
+		data_growth += c->bi.dent_budget;
+	data_growth += req->new_ino_d;
+	return data_growth;
+}
+
+/**
+ * calc_dd_growth - calculate approximate amount of data which makes other data
+ * dirty from budgeting request.
+ * @c: UBIFS file-system description object
+ * @req: budgeting request
+ */
+static int calc_dd_growth(const struct ubifs_info *c,
+			  const struct ubifs_budget_req *req)
+{
+	int dd_growth;
+
+	dd_growth = req->dirtied_page ? c->bi.page_budget : 0;
+
+	if (req->dirtied_ino)
+		dd_growth += c->bi.inode_budget * req->dirtied_ino;
+	if (req->mod_dent)
+		dd_growth += c->bi.dent_budget;
+	dd_growth += req->dirtied_ino_d;
+	return dd_growth;
+}
+
+/**
+ * ubifs_budget_space - ensure there is enough space to complete an operation.
+ * @c: UBIFS file-system description object
+ * @req: budget request
+ *
+ * This function allocates budget for an operation. It uses pessimistic
+ * approximation of how much flash space the operation needs. The goal of this
+ * function is to make sure UBIFS always has flash space to flush all dirty
+ * pages, dirty inodes, and dirty znodes (liability). This function may force
+ * commit, garbage-collection or write-back. Returns zero in case of success,
+ * %-ENOSPC if there is no free space and other negative error codes in case of
+ * failures.
+ */
+int ubifs_budget_space(struct ubifs_info *c, struct ubifs_budget_req *req)
+{
+	int err, idx_growth, data_growth, dd_growth, retried = 0;
+
+	ubifs_assert(c, req->new_page <= 1);
+	ubifs_assert(c, req->dirtied_page <= 1);
+	ubifs_assert(c, req->new_dent <= 1);
+	ubifs_assert(c, req->mod_dent <= 1);
+	ubifs_assert(c, req->new_ino <= 1);
+	ubifs_assert(c, req->new_ino_d <= UBIFS_MAX_INO_DATA);
+	ubifs_assert(c, req->dirtied_ino <= 4);
+	ubifs_assert(c, req->dirtied_ino_d <= UBIFS_MAX_INO_DATA * 4);
+	ubifs_assert(c, !(req->new_ino_d & 7));
+	ubifs_assert(c, !(req->dirtied_ino_d & 7));
+
+	data_growth = calc_data_growth(c, req);
+	dd_growth = calc_dd_growth(c, req);
+	if (!data_growth && !dd_growth)
+		return 0;
+	idx_growth = calc_idx_growth(c, req);
+
+again:
+	spin_lock(&c->space_lock);
+	ubifs_assert(c, c->bi.idx_growth >= 0);
+	ubifs_assert(c, c->bi.data_growth >= 0);
+	ubifs_assert(c, c->bi.dd_growth >= 0);
+
+	if (unlikely(c->bi.nospace) && (c->bi.nospace_rp || !can_use_rp(c))) {
+		dbg_budg("no space");
+		spin_unlock(&c->space_lock);
+		return -ENOSPC;
+	}
+
+	c->bi.idx_growth += idx_growth;
+	c->bi.data_growth += data_growth;
+	c->bi.dd_growth += dd_growth;
+
+	err = do_budget_space(c);
+	if (likely(!err)) {
+		req->idx_growth = idx_growth;
+		req->data_growth = data_growth;
+		req->dd_growth = dd_growth;
+		spin_unlock(&c->space_lock);
+		return 0;
+	}
+
+	/* Restore the old values */
+	c->bi.idx_growth -= idx_growth;
+	c->bi.data_growth -= data_growth;
+	c->bi.dd_growth -= dd_growth;
+	spin_unlock(&c->space_lock);
+
+	if (req->fast) {
+		dbg_budg("no space for fast budgeting");
+		return err;
+	}
+
+	err = make_free_space(c);
+	cond_resched();
+	if (err == -EAGAIN) {
+		dbg_budg("try again");
+		goto again;
+	} else if (err == -ENOSPC) {
+		if (!retried) {
+			retried = 1;
+			dbg_budg("-ENOSPC, but anyway try once again");
+			goto again;
+		}
+		dbg_budg("FS is full, -ENOSPC");
+		c->bi.nospace = 1;
+		if (can_use_rp(c) || c->rp_size == 0)
+			c->bi.nospace_rp = 1;
+		smp_wmb();
+	} else
+		ubifs_err(c, "cannot budget space, error %d", err);
+	return err;
+}
+
+/**
+ * ubifs_release_budget - release budgeted free space.
+ * @c: UBIFS file-system description object
+ * @req: budget request
+ *
+ * This function releases the space budgeted by 'ubifs_budget_space()'. Note,
+ * since the index changes (which were budgeted for in @req->idx_growth) will
+ * only be written to the media on commit, this function moves the index budget
+ * from @c->bi.idx_growth to @c->bi.uncommitted_idx. The latter will be zeroed
+ * by the commit operation.
+ */
+void ubifs_release_budget(struct ubifs_info *c, struct ubifs_budget_req *req)
+{
+	ubifs_assert(c, req->new_page <= 1);
+	ubifs_assert(c, req->dirtied_page <= 1);
+	ubifs_assert(c, req->new_dent <= 1);
+	ubifs_assert(c, req->mod_dent <= 1);
+	ubifs_assert(c, req->new_ino <= 1);
+	ubifs_assert(c, req->new_ino_d <= UBIFS_MAX_INO_DATA);
+	ubifs_assert(c, req->dirtied_ino <= 4);
+	ubifs_assert(c, req->dirtied_ino_d <= UBIFS_MAX_INO_DATA * 4);
+	ubifs_assert(c, !(req->new_ino_d & 7));
+	ubifs_assert(c, !(req->dirtied_ino_d & 7));
+	if (!req->recalculate) {
+		ubifs_assert(c, req->idx_growth >= 0);
+		ubifs_assert(c, req->data_growth >= 0);
+		ubifs_assert(c, req->dd_growth >= 0);
+	}
+
+	if (req->recalculate) {
+		req->data_growth = calc_data_growth(c, req);
+		req->dd_growth = calc_dd_growth(c, req);
+		req->idx_growth = calc_idx_growth(c, req);
+	}
+
+	if (!req->data_growth && !req->dd_growth)
+		return;
+
+	c->bi.nospace = c->bi.nospace_rp = 0;
+	smp_wmb();
+
+	spin_lock(&c->space_lock);
+	c->bi.idx_growth -= req->idx_growth;
+	c->bi.uncommitted_idx += req->idx_growth;
+	c->bi.data_growth -= req->data_growth;
+	c->bi.dd_growth -= req->dd_growth;
+	c->bi.min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+
+	ubifs_assert(c, c->bi.idx_growth >= 0);
+	ubifs_assert(c, c->bi.data_growth >= 0);
+	ubifs_assert(c, c->bi.dd_growth >= 0);
+	ubifs_assert(c, c->bi.min_idx_lebs < c->main_lebs);
+	ubifs_assert(c, !(c->bi.idx_growth & 7));
+	ubifs_assert(c, !(c->bi.data_growth & 7));
+	ubifs_assert(c, !(c->bi.dd_growth & 7));
+	spin_unlock(&c->space_lock);
+}
+
+/**
+ * ubifs_convert_page_budget - convert budget of a new page.
+ * @c: UBIFS file-system description object
+ *
+ * This function converts budget which was allocated for a new page of data to
+ * the budget of changing an existing page of data. The latter is smaller than
+ * the former, so this function only does simple re-calculation and does not
+ * involve any write-back.
+ */
+void ubifs_convert_page_budget(struct ubifs_info *c)
+{
+	spin_lock(&c->space_lock);
+	/* Release the index growth reservation */
+	c->bi.idx_growth -= c->max_idx_node_sz << UBIFS_BLOCKS_PER_PAGE_SHIFT;
+	/* Release the data growth reservation */
+	c->bi.data_growth -= c->bi.page_budget;
+	/* Increase the dirty data growth reservation instead */
+	c->bi.dd_growth += c->bi.page_budget;
+	/* And re-calculate the indexing space reservation */
+	c->bi.min_idx_lebs = ubifs_calc_min_idx_lebs(c);
+	spin_unlock(&c->space_lock);
+}
+
+/**
+ * ubifs_release_dirty_inode_budget - release dirty inode budget.
+ * @c: UBIFS file-system description object
+ * @ui: UBIFS inode to release the budget for
+ *
+ * This function releases budget corresponding to a dirty inode. It is usually
+ * called when after the inode has been written to the media and marked as
+ * clean. It also causes the "no space" flags to be cleared.
+ */
+void ubifs_release_dirty_inode_budget(struct ubifs_info *c,
+				      struct ubifs_inode *ui)
+{
+	struct ubifs_budget_req req;
+
+	memset(&req, 0, sizeof(struct ubifs_budget_req));
+	/* The "no space" flags will be cleared because dd_growth is > 0 */
+	req.dd_growth = c->bi.inode_budget + ALIGN(ui->data_len, 8);
+	ubifs_release_budget(c, &req);
+}
+
+/**
+ * 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);
+}
+
+/**
+ * ubifs_get_free_space_nolock - return amount of free space.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates amount of free space to report to user-space.
+ *
+ * Because UBIFS may introduce substantial overhead (the index, node headers,
+ * alignment, wastage at the end of LEBs, etc), it cannot report real amount of
+ * free flash space it has (well, because not all dirty space is reclaimable,
+ * UBIFS does not actually know the real amount). If UBIFS did so, it would
+ * bread user expectations about what free space is. Users seem to accustomed
+ * to assume that if the file-system reports N bytes of free space, they would
+ * be able to fit a file of N bytes to the FS. This almost works for
+ * traditional file-systems, because they have way less overhead than UBIFS.
+ * So, to keep users happy, UBIFS tries to take the overhead into account.
+ */
+long long ubifs_get_free_space_nolock(struct ubifs_info *c)
+{
+	int rsvd_idx_lebs, lebs;
+	long long available, outstanding, free;
+
+	ubifs_assert(c, c->bi.min_idx_lebs == ubifs_calc_min_idx_lebs(c));
+	outstanding = c->bi.data_growth + c->bi.dd_growth;
+	available = ubifs_calc_available(c, c->bi.min_idx_lebs);
+
+	/*
+	 * When reporting free space to user-space, UBIFS guarantees that it is
+	 * possible to write a file of free space size. This means that for
+	 * empty LEBs we may use more precise calculations than
+	 * 'ubifs_calc_available()' is using. Namely, we know that in empty
+	 * LEBs we would waste only @c->leb_overhead bytes, not @c->dark_wm.
+	 * Thus, amend the available space.
+	 *
+	 * Note, the calculations below are similar to what we have in
+	 * 'do_budget_space()', so refer there for comments.
+	 */
+	if (c->bi.min_idx_lebs > c->lst.idx_lebs)
+		rsvd_idx_lebs = c->bi.min_idx_lebs - c->lst.idx_lebs;
+	else
+		rsvd_idx_lebs = 0;
+	lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt -
+	       c->lst.taken_empty_lebs;
+	lebs -= rsvd_idx_lebs;
+	available += lebs * (c->dark_wm - c->leb_overhead);
+
+	if (available > outstanding)
+		free = ubifs_reported_space(c, available - outstanding);
+	else
+		free = 0;
+	return free;
+}
+
+/**
+ * ubifs_get_free_space - return amount of free space.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates and returns amount of free space to report to
+ * user-space.
+ */
+long long ubifs_get_free_space(struct ubifs_info *c)
+{
+	long long free;
+
+	spin_lock(&c->space_lock);
+	free = ubifs_get_free_space_nolock(c);
+	spin_unlock(&c->space_lock);
+
+	return free;
+}

linux-7.2.3/fs/ubifs/commit.c

@@ -0,0 +1,733 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Adrian Hunter
+ *          Artem Bityutskiy (Битюцкий Артём)
+ */
+
+/*
+ * This file implements functions that manage the running of the commit process.
+ * Each affected module has its own functions to accomplish their part in the
+ * commit and those functions are called here.
+ *
+ * The commit is the process whereby all updates to the index and LEB properties
+ * are written out together and the journal becomes empty. This keeps the
+ * file system consistent - at all times the state can be recreated by reading
+ * the index and LEB properties and then replaying the journal.
+ *
+ * The commit is split into two parts named "commit start" and "commit end".
+ * During commit start, the commit process has exclusive access to the journal
+ * by holding the commit semaphore down for writing. As few I/O operations as
+ * possible are performed during commit start, instead the nodes that are to be
+ * written are merely identified. During commit end, the commit semaphore is no
+ * longer held and the journal is again in operation, allowing users to continue
+ * to use the file system while the bulk of the commit I/O is performed. The
+ * purpose of this two-step approach is to prevent the commit from causing any
+ * latency blips. Note that in any case, the commit does not prevent lookups
+ * (as permitted by the TNC mutex), or access to VFS data structures e.g. page
+ * cache.
+ */
+
+#include <linux/freezer.h>
+#include <linux/kthread.h>
+#include <linux/slab.h>
+#include "ubifs.h"
+
+/*
+ * nothing_to_commit - check if there is nothing to commit.
+ * @c: UBIFS file-system description object
+ *
+ * This is a helper function which checks if there is anything to commit. It is
+ * used as an optimization to avoid starting the commit if it is not really
+ * necessary. Indeed, the commit operation always assumes flash I/O (e.g.,
+ * writing the commit start node to the log), and it is better to avoid doing
+ * this unnecessarily. E.g., 'ubifs_sync_fs()' runs the commit, but if there is
+ * nothing to commit, it is more optimal to avoid any flash I/O.
+ *
+ * This function has to be called with @c->commit_sem locked for writing -
+ * this function does not take LPT/TNC locks because the @c->commit_sem
+ * guarantees that we have exclusive access to the TNC and LPT data structures.
+ *
+ * This function returns %1 if there is nothing to commit and %0 otherwise.
+ */
+static int nothing_to_commit(struct ubifs_info *c)
+{
+	/*
+	 * During mounting or remounting from R/O mode to R/W mode we may
+	 * commit for various recovery-related reasons.
+	 */
+	if (c->mounting || c->remounting_rw)
+		return 0;
+
+	/*
+	 * If the root TNC node is dirty, we definitely have something to
+	 * commit.
+	 */
+	if (c->zroot.znode && ubifs_zn_dirty(c->zroot.znode))
+		return 0;
+
+	/*
+	 * Increasing @c->dirty_pn_cnt/@c->dirty_nn_cnt and marking
+	 * nnodes/pnodes as dirty in run_gc() could race with following
+	 * checking, which leads inconsistent states between @c->nroot
+	 * and @c->dirty_pn_cnt/@c->dirty_nn_cnt, holding @c->lp_mutex
+	 * to avoid that.
+	 */
+	mutex_lock(&c->lp_mutex);
+	/*
+	 * Even though the TNC is clean, the LPT tree may have dirty nodes. For
+	 * example, this may happen if the budgeting subsystem invoked GC to
+	 * make some free space, and the GC found an LEB with only dirty and
+	 * free space. In this case GC would just change the lprops of this
+	 * LEB (by turning all space into free space) and unmap it.
+	 */
+	if (c->nroot && test_bit(DIRTY_CNODE, &c->nroot->flags)) {
+		mutex_unlock(&c->lp_mutex);
+		return 0;
+	}
+
+	ubifs_assert(c, atomic_long_read(&c->dirty_zn_cnt) == 0);
+	ubifs_assert(c, c->dirty_pn_cnt == 0);
+	ubifs_assert(c, c->dirty_nn_cnt == 0);
+	mutex_unlock(&c->lp_mutex);
+
+	return 1;
+}
+
+/**
+ * do_commit - commit the journal.
+ * @c: UBIFS file-system description object
+ *
+ * This function implements UBIFS commit. It has to be called with commit lock
+ * locked. Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int do_commit(struct ubifs_info *c)
+{
+	int err, new_ltail_lnum, old_ltail_lnum, i;
+	struct ubifs_zbranch zroot;
+	struct ubifs_lp_stats lst;
+
+	dbg_cmt("start");
+	ubifs_assert(c, !c->ro_media && !c->ro_mount);
+
+	if (c->ro_error) {
+		err = -EROFS;
+		goto out_up;
+	}
+
+	if (nothing_to_commit(c)) {
+		up_write(&c->commit_sem);
+		err = 0;
+		goto out_cancel;
+	}
+
+	/* Sync all write buffers (necessary for recovery) */
+	for (i = 0; i < c->jhead_cnt; i++) {
+		err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
+		if (err)
+			goto out_up;
+	}
+
+	c->cmt_no += 1;
+	err = ubifs_gc_start_commit(c);
+	if (err)
+		goto out_up;
+	err = dbg_check_lprops(c);
+	if (err)
+		goto out_up;
+	err = ubifs_log_start_commit(c, &new_ltail_lnum);
+	if (err)
+		goto out_up;
+	err = ubifs_tnc_start_commit(c, &zroot);
+	if (err)
+		goto out_up;
+	err = ubifs_lpt_start_commit(c);
+	if (err)
+		goto out_up;
+	err = ubifs_orphan_start_commit(c);
+	if (err)
+		goto out_up;
+
+	ubifs_get_lp_stats(c, &lst);
+
+	up_write(&c->commit_sem);
+
+	err = ubifs_tnc_end_commit(c);
+	if (err)
+		goto out;
+	err = ubifs_lpt_end_commit(c);
+	if (err)
+		goto out;
+	err = ubifs_orphan_end_commit(c);
+	if (err)
+		goto out;
+	err = dbg_check_old_index(c, &zroot);
+	if (err)
+		goto out;
+
+	c->mst_node->cmt_no      = cpu_to_le64(c->cmt_no);
+	c->mst_node->log_lnum    = cpu_to_le32(new_ltail_lnum);
+	c->mst_node->root_lnum   = cpu_to_le32(zroot.lnum);
+	c->mst_node->root_offs   = cpu_to_le32(zroot.offs);
+	c->mst_node->root_len    = cpu_to_le32(zroot.len);
+	c->mst_node->ihead_lnum  = cpu_to_le32(c->ihead_lnum);
+	c->mst_node->ihead_offs  = cpu_to_le32(c->ihead_offs);
+	c->mst_node->index_size  = cpu_to_le64(c->bi.old_idx_sz);
+	c->mst_node->lpt_lnum    = cpu_to_le32(c->lpt_lnum);
+	c->mst_node->lpt_offs    = cpu_to_le32(c->lpt_offs);
+	c->mst_node->nhead_lnum  = cpu_to_le32(c->nhead_lnum);
+	c->mst_node->nhead_offs  = cpu_to_le32(c->nhead_offs);
+	c->mst_node->ltab_lnum   = cpu_to_le32(c->ltab_lnum);
+	c->mst_node->ltab_offs   = cpu_to_le32(c->ltab_offs);
+	c->mst_node->lsave_lnum  = cpu_to_le32(c->lsave_lnum);
+	c->mst_node->lsave_offs  = cpu_to_le32(c->lsave_offs);
+	c->mst_node->lscan_lnum  = cpu_to_le32(c->lscan_lnum);
+	c->mst_node->empty_lebs  = cpu_to_le32(lst.empty_lebs);
+	c->mst_node->idx_lebs    = cpu_to_le32(lst.idx_lebs);
+	c->mst_node->total_free  = cpu_to_le64(lst.total_free);
+	c->mst_node->total_dirty = cpu_to_le64(lst.total_dirty);
+	c->mst_node->total_used  = cpu_to_le64(lst.total_used);
+	c->mst_node->total_dead  = cpu_to_le64(lst.total_dead);
+	c->mst_node->total_dark  = cpu_to_le64(lst.total_dark);
+	if (c->no_orphs)
+		c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS);
+	else
+		c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_NO_ORPHS);
+
+	old_ltail_lnum = c->ltail_lnum;
+	err = ubifs_log_end_commit(c, new_ltail_lnum);
+	if (err)
+		goto out;
+
+	err = ubifs_log_post_commit(c, old_ltail_lnum);
+	if (err)
+		goto out;
+	err = ubifs_gc_end_commit(c);
+	if (err)
+		goto out;
+	err = ubifs_lpt_post_commit(c);
+	if (err)
+		goto out;
+
+out_cancel:
+	spin_lock(&c->cs_lock);
+	c->cmt_state = COMMIT_RESTING;
+	wake_up(&c->cmt_wq);
+	dbg_cmt("commit end");
+	spin_unlock(&c->cs_lock);
+	return 0;
+
+out_up:
+	up_write(&c->commit_sem);
+out:
+	ubifs_err(c, "commit failed, error %d", err);
+	spin_lock(&c->cs_lock);
+	c->cmt_state = COMMIT_BROKEN;
+	wake_up(&c->cmt_wq);
+	spin_unlock(&c->cs_lock);
+	ubifs_ro_mode(c, err);
+	return err;
+}
+
+/**
+ * run_bg_commit - run background commit if it is needed.
+ * @c: UBIFS file-system description object
+ *
+ * This function runs background commit if it is needed. Returns zero in case
+ * of success and a negative error code in case of failure.
+ */
+static int run_bg_commit(struct ubifs_info *c)
+{
+	spin_lock(&c->cs_lock);
+	/*
+	 * Run background commit only if background commit was requested or if
+	 * commit is required.
+	 */
+	if (c->cmt_state != COMMIT_BACKGROUND &&
+	    c->cmt_state != COMMIT_REQUIRED)
+		goto out;
+	spin_unlock(&c->cs_lock);
+
+	down_write(&c->commit_sem);
+	spin_lock(&c->cs_lock);
+	if (c->cmt_state == COMMIT_REQUIRED)
+		c->cmt_state = COMMIT_RUNNING_REQUIRED;
+	else if (c->cmt_state == COMMIT_BACKGROUND)
+		c->cmt_state = COMMIT_RUNNING_BACKGROUND;
+	else
+		goto out_cmt_unlock;
+	spin_unlock(&c->cs_lock);
+
+	return do_commit(c);
+
+out_cmt_unlock:
+	up_write(&c->commit_sem);
+out:
+	spin_unlock(&c->cs_lock);
+	return 0;
+}
+
+/**
+ * ubifs_bg_thread - UBIFS background thread function.
+ * @info: points to the file-system description object
+ *
+ * This function implements various file-system background activities:
+ * o when a write-buffer timer expires it synchronizes the appropriate
+ *   write-buffer;
+ * o when the journal is about to be full, it starts in-advance commit.
+ *
+ * Note, other stuff like background garbage collection may be added here in
+ * future.
+ */
+int ubifs_bg_thread(void *info)
+{
+	int err;
+	struct ubifs_info *c = info;
+
+	ubifs_msg(c, "background thread \"%s\" started, PID %d",
+		  c->bgt_name, current->pid);
+	set_freezable();
+
+	while (1) {
+		if (kthread_should_stop())
+			break;
+
+		if (try_to_freeze())
+			continue;
+
+		set_current_state(TASK_INTERRUPTIBLE);
+		/* Check if there is something to do */
+		if (!c->need_bgt) {
+			/*
+			 * Nothing prevents us from going sleep now and
+			 * be never woken up and block the task which
+			 * could wait in 'kthread_stop()' forever.
+			 */
+			if (kthread_should_stop())
+				break;
+			schedule();
+			continue;
+		} else
+			__set_current_state(TASK_RUNNING);
+
+		c->need_bgt = 0;
+		err = ubifs_bg_wbufs_sync(c);
+		if (err)
+			ubifs_ro_mode(c, err);
+
+		run_bg_commit(c);
+		cond_resched();
+	}
+
+	ubifs_msg(c, "background thread \"%s\" stops", c->bgt_name);
+	return 0;
+}
+
+/**
+ * ubifs_commit_required - set commit state to "required".
+ * @c: UBIFS file-system description object
+ *
+ * This function is called if a commit is required but cannot be done from the
+ * calling function, so it is just flagged instead.
+ */
+void ubifs_commit_required(struct ubifs_info *c)
+{
+	spin_lock(&c->cs_lock);
+	switch (c->cmt_state) {
+	case COMMIT_RESTING:
+	case COMMIT_BACKGROUND:
+		dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
+			dbg_cstate(COMMIT_REQUIRED));
+		c->cmt_state = COMMIT_REQUIRED;
+		break;
+	case COMMIT_RUNNING_BACKGROUND:
+		dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
+			dbg_cstate(COMMIT_RUNNING_REQUIRED));
+		c->cmt_state = COMMIT_RUNNING_REQUIRED;
+		break;
+	case COMMIT_REQUIRED:
+	case COMMIT_RUNNING_REQUIRED:
+	case COMMIT_BROKEN:
+		break;
+	}
+	spin_unlock(&c->cs_lock);
+}
+
+/**
+ * ubifs_request_bg_commit - notify the background thread to do a commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function is called if the journal is full enough to make a commit
+ * worthwhile, so background thread is kicked to start it.
+ */
+void ubifs_request_bg_commit(struct ubifs_info *c)
+{
+	spin_lock(&c->cs_lock);
+	if (c->cmt_state == COMMIT_RESTING) {
+		dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
+			dbg_cstate(COMMIT_BACKGROUND));
+		c->cmt_state = COMMIT_BACKGROUND;
+		spin_unlock(&c->cs_lock);
+		ubifs_wake_up_bgt(c);
+	} else
+		spin_unlock(&c->cs_lock);
+}
+
+/**
+ * wait_for_commit - wait for commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function sleeps until the commit operation is no longer running.
+ */
+static int wait_for_commit(struct ubifs_info *c)
+{
+	dbg_cmt("pid %d goes sleep", current->pid);
+
+	/*
+	 * The following sleeps if the condition is false, and will be woken
+	 * when the commit ends. It is possible, although very unlikely, that we
+	 * will wake up and see the subsequent commit running, rather than the
+	 * one we were waiting for, and go back to sleep.  However, we will be
+	 * woken again, so there is no danger of sleeping forever.
+	 */
+	wait_event(c->cmt_wq, c->cmt_state != COMMIT_RUNNING_BACKGROUND &&
+			      c->cmt_state != COMMIT_RUNNING_REQUIRED);
+	dbg_cmt("commit finished, pid %d woke up", current->pid);
+	return 0;
+}
+
+/**
+ * ubifs_run_commit - run or wait for commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function runs commit and returns zero in case of success and a negative
+ * error code in case of failure.
+ */
+int ubifs_run_commit(struct ubifs_info *c)
+{
+	int err = 0;
+
+	spin_lock(&c->cs_lock);
+	if (c->cmt_state == COMMIT_BROKEN) {
+		err = -EROFS;
+		goto out;
+	}
+
+	if (c->cmt_state == COMMIT_RUNNING_BACKGROUND)
+		/*
+		 * We set the commit state to 'running required' to indicate
+		 * that we want it to complete as quickly as possible.
+		 */
+		c->cmt_state = COMMIT_RUNNING_REQUIRED;
+
+	if (c->cmt_state == COMMIT_RUNNING_REQUIRED) {
+		spin_unlock(&c->cs_lock);
+		return wait_for_commit(c);
+	}
+	spin_unlock(&c->cs_lock);
+
+	/* Ok, the commit is indeed needed */
+
+	down_write(&c->commit_sem);
+	spin_lock(&c->cs_lock);
+	/*
+	 * Since we unlocked 'c->cs_lock', the state may have changed, so
+	 * re-check it.
+	 */
+	if (c->cmt_state == COMMIT_BROKEN) {
+		err = -EROFS;
+		goto out_cmt_unlock;
+	}
+
+	if (c->cmt_state == COMMIT_RUNNING_BACKGROUND)
+		c->cmt_state = COMMIT_RUNNING_REQUIRED;
+
+	if (c->cmt_state == COMMIT_RUNNING_REQUIRED) {
+		up_write(&c->commit_sem);
+		spin_unlock(&c->cs_lock);
+		return wait_for_commit(c);
+	}
+	c->cmt_state = COMMIT_RUNNING_REQUIRED;
+	spin_unlock(&c->cs_lock);
+
+	err = do_commit(c);
+	return err;
+
+out_cmt_unlock:
+	up_write(&c->commit_sem);
+out:
+	spin_unlock(&c->cs_lock);
+	return err;
+}
+
+/**
+ * ubifs_gc_should_commit - determine if it is time for GC to run commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function is called by garbage collection to determine if commit should
+ * be run. If commit state is @COMMIT_BACKGROUND, which means that the journal
+ * is full enough to start commit, this function returns true. It is not
+ * absolutely necessary to commit yet, but it feels like this should be better
+ * then to keep doing GC. This function returns %1 if GC has to initiate commit
+ * and %0 if not.
+ */
+int ubifs_gc_should_commit(struct ubifs_info *c)
+{
+	int ret = 0;
+
+	spin_lock(&c->cs_lock);
+	if (c->cmt_state == COMMIT_BACKGROUND) {
+		dbg_cmt("commit required now");
+		c->cmt_state = COMMIT_REQUIRED;
+	} else
+		dbg_cmt("commit not requested");
+	if (c->cmt_state == COMMIT_REQUIRED)
+		ret = 1;
+	spin_unlock(&c->cs_lock);
+	return ret;
+}
+
+/*
+ * Everything below is related to debugging.
+ */
+
+/**
+ * struct idx_node - hold index nodes during index tree traversal.
+ * @list: list
+ * @iip: index in parent (slot number of this indexing node in the parent
+ *       indexing node)
+ * @upper_key: all keys in this indexing node have to be less or equivalent to
+ *             this key
+ * @idx: index node (8-byte aligned because all node structures must be 8-byte
+ *       aligned)
+ */
+struct idx_node {
+	struct list_head list;
+	int iip;
+	union ubifs_key upper_key;
+	struct ubifs_idx_node idx __aligned(8);
+};
+
+/**
+ * dbg_old_index_check_init - get information for the next old index check.
+ * @c: UBIFS file-system description object
+ * @zroot: root of the index
+ *
+ * This function records information about the index that will be needed for the
+ * next old index check i.e. 'dbg_check_old_index()'.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot)
+{
+	struct ubifs_idx_node *idx;
+	int lnum, offs, len, err = 0;
+	struct ubifs_debug_info *d = c->dbg;
+
+	d->old_zroot = *zroot;
+	lnum = d->old_zroot.lnum;
+	offs = d->old_zroot.offs;
+	len = d->old_zroot.len;
+
+	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)
+		goto out;
+
+	d->old_zroot_level = le16_to_cpu(idx->level);
+	d->old_zroot_sqnum = le64_to_cpu(idx->ch.sqnum);
+out:
+	kfree(idx);
+	return err;
+}
+
+/**
+ * dbg_check_old_index - check the old copy of the index.
+ * @c: UBIFS file-system description object
+ * @zroot: root of the new index
+ *
+ * In order to be able to recover from an unclean unmount, a complete copy of
+ * the index must exist on flash. This is the "old" index. The commit process
+ * must write the "new" index to flash without overwriting or destroying any
+ * part of the old index. This function is run at commit end in order to check
+ * that the old index does indeed exist completely intact.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot)
+{
+	int lnum, offs, len, err = 0, last_level, child_cnt;
+	int first = 1, iip;
+	struct ubifs_debug_info *d = c->dbg;
+	union ubifs_key lower_key, upper_key, l_key, u_key;
+	unsigned long long last_sqnum;
+	struct ubifs_idx_node *idx;
+	struct list_head list;
+	struct idx_node *i;
+	size_t sz;
+
+	if (!dbg_is_chk_index(c))
+		return 0;
+
+	INIT_LIST_HEAD(&list);
+
+	sz = sizeof(struct idx_node) + ubifs_idx_node_sz(c, c->fanout) -
+	     UBIFS_IDX_NODE_SZ;
+
+	/* Start at the old zroot */
+	lnum = d->old_zroot.lnum;
+	offs = d->old_zroot.offs;
+	len = d->old_zroot.len;
+	iip = 0;
+
+	/*
+	 * Traverse the index tree preorder depth-first i.e. do a node and then
+	 * its subtrees from left to right.
+	 */
+	while (1) {
+		struct ubifs_branch *br;
+
+		/* Get the next index node */
+		i = kmalloc(sz, GFP_NOFS);
+		if (!i) {
+			err = -ENOMEM;
+			goto out_free;
+		}
+		i->iip = iip;
+		/* Keep the index nodes on our path in a linked list */
+		list_add_tail(&i->list, &list);
+		/* Read the index node */
+		idx = &i->idx;
+		err = ubifs_read_node(c, idx, UBIFS_IDX_NODE, len, lnum, offs);
+		if (err)
+			goto out_free;
+		/* Validate index node */
+		child_cnt = le16_to_cpu(idx->child_cnt);
+		if (child_cnt < 1 || child_cnt > c->fanout) {
+			err = 1;
+			goto out_dump;
+		}
+		if (first) {
+			first = 0;
+			/* Check root level and sqnum */
+			if (le16_to_cpu(idx->level) != d->old_zroot_level) {
+				err = 2;
+				goto out_dump;
+			}
+			if (le64_to_cpu(idx->ch.sqnum) != d->old_zroot_sqnum) {
+				err = 3;
+				goto out_dump;
+			}
+			/* Set last values as though root had a parent */
+			last_level = le16_to_cpu(idx->level) + 1;
+			last_sqnum = le64_to_cpu(idx->ch.sqnum) + 1;
+			key_read(c, ubifs_idx_key(c, idx), &lower_key);
+			highest_ino_key(c, &upper_key, INUM_WATERMARK);
+		}
+		key_copy(c, &upper_key, &i->upper_key);
+		if (le16_to_cpu(idx->level) != last_level - 1) {
+			err = 3;
+			goto out_dump;
+		}
+		/*
+		 * The index is always written bottom up hence a child's sqnum
+		 * is always less than the parents.
+		 */
+		if (le64_to_cpu(idx->ch.sqnum) >= last_sqnum) {
+			err = 4;
+			goto out_dump;
+		}
+		/* Check key range */
+		key_read(c, ubifs_idx_key(c, idx), &l_key);
+		br = ubifs_idx_branch(c, idx, child_cnt - 1);
+		key_read(c, &br->key, &u_key);
+		if (keys_cmp(c, &lower_key, &l_key) > 0) {
+			err = 5;
+			goto out_dump;
+		}
+		if (keys_cmp(c, &upper_key, &u_key) < 0) {
+			err = 6;
+			goto out_dump;
+		}
+		if (keys_cmp(c, &upper_key, &u_key) == 0)
+			if (!is_hash_key(c, &u_key)) {
+				err = 7;
+				goto out_dump;
+			}
+		/* Go to next index node */
+		if (le16_to_cpu(idx->level) == 0) {
+			/* At the bottom, so go up until can go right */
+			while (1) {
+				/* Drop the bottom of the list */
+				list_del(&i->list);
+				kfree(i);
+				/* No more list means we are done */
+				if (list_empty(&list))
+					goto out;
+				/* Look at the new bottom */
+				i = list_entry(list.prev, struct idx_node,
+					       list);
+				idx = &i->idx;
+				/* Can we go right */
+				if (iip + 1 < le16_to_cpu(idx->child_cnt)) {
+					iip = iip + 1;
+					break;
+				} else
+					/* Nope, so go up again */
+					iip = i->iip;
+			}
+		} else
+			/* Go down left */
+			iip = 0;
+		/*
+		 * We have the parent in 'idx' and now we set up for reading the
+		 * child pointed to by slot 'iip'.
+		 */
+		last_level = le16_to_cpu(idx->level);
+		last_sqnum = le64_to_cpu(idx->ch.sqnum);
+		br = ubifs_idx_branch(c, idx, iip);
+		lnum = le32_to_cpu(br->lnum);
+		offs = le32_to_cpu(br->offs);
+		len = le32_to_cpu(br->len);
+		key_read(c, &br->key, &lower_key);
+		if (iip + 1 < le16_to_cpu(idx->child_cnt)) {
+			br = ubifs_idx_branch(c, idx, iip + 1);
+			key_read(c, &br->key, &upper_key);
+		} else
+			key_copy(c, &i->upper_key, &upper_key);
+	}
+out:
+	err = dbg_old_index_check_init(c, zroot);
+	if (err)
+		goto out_free;
+
+	return 0;
+
+out_dump:
+	ubifs_err(c, "dumping index node (iip=%d)", i->iip);
+	ubifs_dump_node(c, idx, ubifs_idx_node_sz(c, c->fanout));
+	list_del(&i->list);
+	kfree(i);
+	if (!list_empty(&list)) {
+		i = list_entry(list.prev, struct idx_node, list);
+		ubifs_err(c, "dumping parent index node");
+		ubifs_dump_node(c, &i->idx, ubifs_idx_node_sz(c, c->fanout));
+	}
+out_free:
+	while (!list_empty(&list)) {
+		i = list_entry(list.next, struct idx_node, list);
+		list_del(&i->list);
+		kfree(i);
+	}
+	ubifs_err(c, "failed, error %d", err);
+	if (err > 0)
+		err = -EINVAL;
+	return err;
+}

linux-7.2.3/fs/ubifs/compress.c

@@ -0,0 +1,369 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ * Copyright (C) 2006, 2007 University of Szeged, Hungary
+ *
+ * Authors: Adrian Hunter
+ *          Artem Bityutskiy (Битюцкий Артём)
+ *          Zoltan Sogor
+ */
+
+/*
+ * This file provides a single place to access to compression and
+ * decompression.
+ */
+
+#include <crypto/acompress.h>
+#include <linux/highmem.h>
+#include "ubifs.h"
+
+union ubifs_in_ptr {
+	const void *buf;
+	struct folio *folio;
+};
+
+/* Fake description object for the "none" compressor */
+static struct ubifs_compressor none_compr = {
+	.compr_type = UBIFS_COMPR_NONE,
+	.name = "none",
+	.capi_name = "",
+};
+
+#ifdef CONFIG_UBIFS_FS_LZO
+static struct ubifs_compressor lzo_compr = {
+	.compr_type = UBIFS_COMPR_LZO,
+	.name = "lzo",
+	.capi_name = "lzo",
+};
+#else
+static struct ubifs_compressor lzo_compr = {
+	.compr_type = UBIFS_COMPR_LZO,
+	.name = "lzo",
+};
+#endif
+
+#ifdef CONFIG_UBIFS_FS_ZLIB
+static struct ubifs_compressor zlib_compr = {
+	.compr_type = UBIFS_COMPR_ZLIB,
+	.name = "zlib",
+	.capi_name = "deflate",
+};
+#else
+static struct ubifs_compressor zlib_compr = {
+	.compr_type = UBIFS_COMPR_ZLIB,
+	.name = "zlib",
+};
+#endif
+
+#ifdef CONFIG_UBIFS_FS_ZSTD
+static struct ubifs_compressor zstd_compr = {
+	.compr_type = UBIFS_COMPR_ZSTD,
+	.name = "zstd",
+	.capi_name = "zstd",
+};
+#else
+static struct ubifs_compressor zstd_compr = {
+	.compr_type = UBIFS_COMPR_ZSTD,
+	.name = "zstd",
+};
+#endif
+
+/* All UBIFS compressors */
+struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT];
+
+static void ubifs_compress_common(int *compr_type, union ubifs_in_ptr in_ptr,
+				  size_t in_offset, int in_len, bool in_folio,
+				  void *out_buf, int *out_len)
+{
+	struct ubifs_compressor *compr = ubifs_compressors[*compr_type];
+	int dlen = *out_len;
+	int err;
+
+	if (*compr_type == UBIFS_COMPR_NONE)
+		goto no_compr;
+
+	/* If the input data is small, do not even try to compress it */
+	if (in_len < UBIFS_MIN_COMPR_LEN)
+		goto no_compr;
+
+	dlen = min(dlen, in_len - UBIFS_MIN_COMPRESS_DIFF);
+
+	do {
+		ACOMP_REQUEST_ON_STACK(req, compr->cc);
+		DECLARE_CRYPTO_WAIT(wait);
+
+		acomp_request_set_callback(req, 0, NULL, NULL);
+		if (in_folio)
+			acomp_request_set_src_folio(req, in_ptr.folio,
+						    in_offset, in_len);
+		else
+			acomp_request_set_src_dma(req, in_ptr.buf, in_len);
+		acomp_request_set_dst_dma(req, out_buf, dlen);
+		err = crypto_acomp_compress(req);
+		dlen = req->dlen;
+		if (err != -EAGAIN)
+			break;
+
+		req = ACOMP_REQUEST_CLONE(req, GFP_NOFS | __GFP_NOWARN);
+		acomp_request_set_callback(req, CRYPTO_TFM_REQ_MAY_BACKLOG,
+					   crypto_req_done, &wait);
+		err = crypto_acomp_compress(req);
+		err = crypto_wait_req(err, &wait);
+		dlen = req->dlen;
+		acomp_request_free(req);
+	} while (0);
+
+	*out_len = dlen;
+	if (err)
+		goto no_compr;
+
+	return;
+
+no_compr:
+	if (in_folio)
+		memcpy_from_folio(out_buf, in_ptr.folio, in_offset, in_len);
+	else
+		memcpy(out_buf, in_ptr.buf, in_len);
+	*out_len = in_len;
+	*compr_type = UBIFS_COMPR_NONE;
+}
+
+/**
+ * ubifs_compress - compress data.
+ * @c: UBIFS file-system description object
+ * @in_buf: data to compress
+ * @in_len: length of the data to compress
+ * @out_buf: output buffer where compressed data should be stored
+ * @out_len: output buffer length is returned here
+ * @compr_type: type of compression to use on enter, actually used compression
+ *              type on exit
+ *
+ * This function compresses input buffer @in_buf of length @in_len and stores
+ * the result in the output buffer @out_buf and the resulting length in
+ * @out_len. If the input buffer does not compress, it is just copied to the
+ * @out_buf. The same happens if @compr_type is %UBIFS_COMPR_NONE or if
+ * compression error occurred.
+ *
+ * Note, if the input buffer was not compressed, it is copied to the output
+ * buffer and %UBIFS_COMPR_NONE is returned in @compr_type.
+ */
+void ubifs_compress(const struct ubifs_info *c, const void *in_buf,
+		    int in_len, void *out_buf, int *out_len, int *compr_type)
+{
+	union ubifs_in_ptr in_ptr = { .buf = in_buf };
+
+	ubifs_compress_common(compr_type, in_ptr, 0, in_len, false,
+			      out_buf, out_len);
+}
+
+/**
+ * ubifs_compress_folio - compress folio.
+ * @c: UBIFS file-system description object
+ * @in_folio: data to compress
+ * @in_offset: offset into @in_folio
+ * @in_len: length of the data to compress
+ * @out_buf: output buffer where compressed data should be stored
+ * @out_len: output buffer length is returned here
+ * @compr_type: type of compression to use on enter, actually used compression
+ *              type on exit
+ *
+ * This function compresses input folio @in_folio of length @in_len and
+ * stores the result in the output buffer @out_buf and the resulting length
+ * in @out_len. If the input buffer does not compress, it is just copied
+ * to the @out_buf. The same happens if @compr_type is %UBIFS_COMPR_NONE
+ * or if compression error occurred.
+ *
+ * Note, if the input buffer was not compressed, it is copied to the output
+ * buffer and %UBIFS_COMPR_NONE is returned in @compr_type.
+ */
+void ubifs_compress_folio(const struct ubifs_info *c, struct folio *in_folio,
+			  size_t in_offset, int in_len, void *out_buf,
+			  int *out_len, int *compr_type)
+{
+	union ubifs_in_ptr in_ptr = { .folio = in_folio };
+
+	ubifs_compress_common(compr_type, in_ptr, in_offset, in_len, true,
+			      out_buf, out_len);
+}
+
+static int ubifs_decompress_common(const struct ubifs_info *c,
+				   const void *in_buf, int in_len,
+				   void *out_ptr, size_t out_offset,
+				   int *out_len, bool out_folio,
+				   int compr_type)
+{
+	struct ubifs_compressor *compr;
+	int dlen = *out_len;
+	int err;
+
+	if (unlikely(compr_type < 0 || compr_type >= UBIFS_COMPR_TYPES_CNT)) {
+		ubifs_err(c, "invalid compression type %d", compr_type);
+		return -EINVAL;
+	}
+
+	compr = ubifs_compressors[compr_type];
+
+	if (unlikely(!compr->capi_name)) {
+		ubifs_err(c, "%s compression is not compiled in", compr->name);
+		return -EINVAL;
+	}
+
+	if (compr_type == UBIFS_COMPR_NONE) {
+		if (out_folio)
+			memcpy_to_folio(out_ptr, out_offset, in_buf, in_len);
+		else
+			memcpy(out_ptr, in_buf, in_len);
+		*out_len = in_len;
+		return 0;
+	}
+
+	do {
+		ACOMP_REQUEST_ON_STACK(req, compr->cc);
+		DECLARE_CRYPTO_WAIT(wait);
+
+		acomp_request_set_callback(req, CRYPTO_TFM_REQ_MAY_BACKLOG,
+					   crypto_req_done, &wait);
+		acomp_request_set_src_dma(req, in_buf, in_len);
+		if (out_folio)
+			acomp_request_set_dst_folio(req, out_ptr, out_offset,
+						    dlen);
+		else
+			acomp_request_set_dst_dma(req, out_ptr, dlen);
+		err = crypto_acomp_decompress(req);
+		dlen = req->dlen;
+		if (err != -EAGAIN)
+			break;
+
+		req = ACOMP_REQUEST_CLONE(req, GFP_NOFS | __GFP_NOWARN);
+		err = crypto_acomp_decompress(req);
+		err = crypto_wait_req(err, &wait);
+		dlen = req->dlen;
+		acomp_request_free(req);
+	} while (0);
+
+	*out_len = dlen;
+	if (err)
+		ubifs_err(c, "cannot decompress %d bytes, compressor %s, error %d",
+			  in_len, compr->name, err);
+
+	return err;
+}
+
+/**
+ * ubifs_decompress - decompress data.
+ * @c: UBIFS file-system description object
+ * @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 struct ubifs_info *c, const void *in_buf,
+		     int in_len, void *out_buf, int *out_len, int compr_type)
+{
+	return ubifs_decompress_common(c, in_buf, in_len, out_buf, 0, out_len,
+				       false, compr_type);
+}
+
+/**
+ * ubifs_decompress_folio - decompress folio.
+ * @c: UBIFS file-system description object
+ * @in_buf: data to decompress
+ * @in_len: length of the data to decompress
+ * @out_folio: output folio where decompressed data should
+ * @out_offset: offset into @out_folio
+ * @out_len: output length is returned here
+ * @compr_type: type of compression
+ *
+ * This function decompresses data from buffer @in_buf into folio
+ * @out_folio.  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_folio(const struct ubifs_info *c, const void *in_buf,
+			   int in_len, struct folio *out_folio,
+			   size_t out_offset, int *out_len, int compr_type)
+{
+	return ubifs_decompress_common(c, in_buf, in_len, out_folio,
+				       out_offset, out_len, true, compr_type);
+}
+
+/**
+ * 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)
+{
+	if (compr->capi_name) {
+		compr->cc = crypto_alloc_acomp(compr->capi_name, 0, 0);
+		if (IS_ERR(compr->cc)) {
+			pr_err("UBIFS error (pid %d): cannot initialize compressor %s, error %ld",
+			       current->pid, compr->name, PTR_ERR(compr->cc));
+			return PTR_ERR(compr->cc);
+		}
+	}
+
+	ubifs_compressors[compr->compr_type] = compr;
+	return 0;
+}
+
+/**
+ * compr_exit - de-initialize a compressor.
+ * @compr: compressor description object
+ */
+static void compr_exit(struct ubifs_compressor *compr)
+{
+	if (compr->capi_name)
+		crypto_free_acomp(compr->cc);
+}
+
+/**
+ * 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(&zstd_compr);
+	if (err)
+		goto out_lzo;
+
+	err = compr_init(&zlib_compr);
+	if (err)
+		goto out_zstd;
+
+	ubifs_compressors[UBIFS_COMPR_NONE] = &none_compr;
+	return 0;
+
+out_zstd:
+	compr_exit(&zstd_compr);
+out_lzo:
+	compr_exit(&lzo_compr);
+	return err;
+}
+
+/**
+ * ubifs_compressors_exit - de-initialize UBIFS compressors.
+ */
+void ubifs_compressors_exit(void)
+{
+	compr_exit(&lzo_compr);
+	compr_exit(&zlib_compr);
+	compr_exit(&zstd_compr);
+}

linux-7.2.3/fs/ubifs/crypto.c

@@ -0,0 +1,97 @@
+// SPDX-License-Identifier: GPL-2.0
+#include "ubifs.h"
+
+static int ubifs_crypt_get_context(struct inode *inode, void *ctx, size_t len)
+{
+	return ubifs_xattr_get(inode, UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT,
+			       ctx, len);
+}
+
+static int ubifs_crypt_set_context(struct inode *inode, const void *ctx,
+				   size_t len, void *fs_data)
+{
+	/*
+	 * Creating an encryption context is done unlocked since we
+	 * operate on a new inode which is not visible to other users
+	 * at this point. So, no need to check whether inode is locked.
+	 */
+	return ubifs_xattr_set(inode, UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT,
+			       ctx, len, 0, false);
+}
+
+static bool ubifs_crypt_empty_dir(struct inode *inode)
+{
+	return ubifs_check_dir_empty(inode) == 0;
+}
+
+/**
+ * ubifs_encrypt - Encrypt data.
+ * @inode: inode which refers to the data node
+ * @dn: data node to encrypt
+ * @in_len: length of data to be compressed
+ * @out_len: allocated memory size for the data area of @dn
+ * @block: logical block number of the block
+ *
+ * This function encrypt a possibly-compressed data in the data node.
+ * The encrypted data length will store in @out_len.
+ */
+int ubifs_encrypt(const struct inode *inode, struct ubifs_data_node *dn,
+		  unsigned int in_len, unsigned int *out_len, int block)
+{
+	struct ubifs_info *c = inode->i_sb->s_fs_info;
+	void *p = &dn->data;
+	unsigned int pad_len = round_up(in_len, UBIFS_CIPHER_BLOCK_SIZE);
+	int err;
+
+	ubifs_assert(c, pad_len <= *out_len);
+	dn->compr_size = cpu_to_le16(in_len);
+
+	/* pad to full block cipher length */
+	if (pad_len != in_len)
+		memset(p + in_len, 0, pad_len - in_len);
+
+	err = fscrypt_encrypt_block_inplace(inode, virt_to_page(p), pad_len,
+					    offset_in_page(p), block);
+	if (err) {
+		ubifs_err(c, "fscrypt_encrypt_block_inplace() failed: %d", err);
+		return err;
+	}
+	*out_len = pad_len;
+
+	return 0;
+}
+
+int ubifs_decrypt(const struct inode *inode, struct ubifs_data_node *dn,
+		  unsigned int *out_len, int block)
+{
+	struct ubifs_info *c = inode->i_sb->s_fs_info;
+	int err;
+	unsigned int clen = le16_to_cpu(dn->compr_size);
+	unsigned int dlen = *out_len;
+
+	if (clen <= 0 || clen > UBIFS_BLOCK_SIZE || clen > dlen) {
+		ubifs_err(c, "bad compr_size: %i", clen);
+		return -EINVAL;
+	}
+
+	ubifs_assert(c, dlen <= UBIFS_BLOCK_SIZE);
+	err = fscrypt_decrypt_block_inplace(inode, virt_to_page(&dn->data),
+					    dlen, offset_in_page(&dn->data),
+					    block);
+	if (err) {
+		ubifs_err(c, "fscrypt_decrypt_block_inplace() failed: %d", err);
+		return err;
+	}
+	*out_len = clen;
+
+	return 0;
+}
+
+const struct fscrypt_operations ubifs_crypt_operations = {
+	.inode_info_offs	= (int)offsetof(struct ubifs_inode, i_crypt_info) -
+				  (int)offsetof(struct ubifs_inode, vfs_inode),
+	.legacy_key_prefix	= "ubifs:",
+	.get_context		= ubifs_crypt_get_context,
+	.set_context		= ubifs_crypt_set_context,
+	.empty_dir		= ubifs_crypt_empty_dir,
+};

linux-7.2.3/fs/ubifs/debug.h

@@ -0,0 +1,305 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+#ifndef __UBIFS_DEBUG_H__
+#define __UBIFS_DEBUG_H__
+
+/* Checking helper functions */
+typedef int (*dbg_leaf_callback)(struct ubifs_info *c,
+				 struct ubifs_zbranch *zbr, void *priv);
+typedef int (*dbg_znode_callback)(struct ubifs_info *c,
+				  struct ubifs_znode *znode, void *priv);
+
+/*
+ * The UBIFS debugfs directory name pattern and maximum name length (3 for "ubi"
+ * + 1 for "_" and 2 for UBI device numbers and 3 for volume number and 1 for
+ * the trailing zero byte.
+ */
+#define UBIFS_DFS_DIR_NAME "ubi%d_%d"
+#define UBIFS_DFS_DIR_LEN  (3 + 1 + 2 + 3 + 1)
+
+/**
+ * ubifs_debug_info - per-FS debugging information.
+ * @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()'
+ *
+ * @pc_happened: non-zero if an emulated power cut happened
+ * @pc_delay: 0=>don't delay, 1=>delay a time, 2=>delay a number of calls
+ * @pc_timeout: time in jiffies when delay of failure mode expires
+ * @pc_cnt: current number of calls to failure mode I/O functions
+ * @pc_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_bi: saved budgeting information
+ * @saved_free: saved amount of free space
+ * @saved_idx_gc_cnt: saved value of @c->idx_gc_cnt
+ *
+ * @chk_gen: if general extra checks are enabled
+ * @chk_index: if index xtra checks are enabled
+ * @chk_orph: if orphans extra checks are enabled
+ * @chk_lprops: if lprops extra checks are enabled
+ * @chk_fs: if UBIFS contents extra checks are enabled
+ * @tst_rcvry: if UBIFS recovery testing mode enabled
+ *
+ * @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
+ * @dfs_chk_gen: debugfs knob to enable UBIFS general extra checks
+ * @dfs_chk_index: debugfs knob to enable UBIFS index extra checks
+ * @dfs_chk_orph: debugfs knob to enable UBIFS orphans extra checks
+ * @dfs_chk_lprops: debugfs knob to enable UBIFS LEP properties extra checks
+ * @dfs_chk_fs: debugfs knob to enable UBIFS contents extra checks
+ * @dfs_tst_rcvry: debugfs knob to enable UBIFS recovery testing
+ * @dfs_ro_error: debugfs knob to switch UBIFS to R/O mode (different to
+ *                re-mounting to R/O mode because it does not flush any buffers
+ *                and UBIFS just starts returning -EROFS on all write
+ *               operations)
+ */
+struct ubifs_debug_info {
+	struct ubifs_zbranch old_zroot;
+	int old_zroot_level;
+	unsigned long long old_zroot_sqnum;
+
+	int pc_happened;
+	int pc_delay;
+	unsigned long pc_timeout;
+	unsigned int pc_cnt;
+	unsigned int pc_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;
+	struct ubifs_budg_info saved_bi;
+	long long saved_free;
+	int saved_idx_gc_cnt;
+
+	unsigned int chk_gen:1;
+	unsigned int chk_index:1;
+	unsigned int chk_orph:1;
+	unsigned int chk_lprops:1;
+	unsigned int chk_fs:1;
+	unsigned int tst_rcvry:1;
+
+	char dfs_dir_name[UBIFS_DFS_DIR_LEN];
+	struct dentry *dfs_dir;
+	struct dentry *dfs_dump_lprops;
+	struct dentry *dfs_dump_budg;
+	struct dentry *dfs_dump_tnc;
+	struct dentry *dfs_chk_gen;
+	struct dentry *dfs_chk_index;
+	struct dentry *dfs_chk_orph;
+	struct dentry *dfs_chk_lprops;
+	struct dentry *dfs_chk_fs;
+	struct dentry *dfs_tst_rcvry;
+	struct dentry *dfs_ro_error;
+};
+
+/**
+ * ubifs_global_debug_info - global (not per-FS) UBIFS debugging information.
+ *
+ * @chk_gen: if general extra checks are enabled
+ * @chk_index: if index xtra checks are enabled
+ * @chk_orph: if orphans extra checks are enabled
+ * @chk_lprops: if lprops extra checks are enabled
+ * @chk_fs: if UBIFS contents extra checks are enabled
+ * @tst_rcvry: if UBIFS recovery testing mode enabled
+ */
+struct ubifs_global_debug_info {
+	unsigned int chk_gen:1;
+	unsigned int chk_index:1;
+	unsigned int chk_orph:1;
+	unsigned int chk_lprops:1;
+	unsigned int chk_fs:1;
+	unsigned int tst_rcvry:1;
+};
+
+void ubifs_assert_failed(struct ubifs_info *c, const char *expr,
+	const char *file, int line);
+
+#define ubifs_assert(c, expr) do {                                             \
+	if (unlikely(!(expr))) {                                               \
+		ubifs_assert_failed((struct ubifs_info *)c, #expr, __FILE__,   \
+		 __LINE__);                                                    \
+	}                                                                      \
+} while (0)
+
+#define ubifs_assert_cmt_locked(c) do {                                        \
+	if (unlikely(down_write_trylock(&(c)->commit_sem))) {                  \
+		up_write(&(c)->commit_sem);                                    \
+		ubifs_err(c, "commit lock is not locked!\n");                  \
+		ubifs_assert(c, 0);                                            \
+	}                                                                      \
+} while (0)
+
+#define ubifs_dbg_msg(type, fmt, ...) \
+	pr_debug("UBIFS DBG " type " (pid %d): " fmt "\n", current->pid,       \
+		 ##__VA_ARGS__)
+
+#define DBG_KEY_BUF_LEN 48
+#define ubifs_dbg_msg_key(type, key, fmt, ...) do {                            \
+	char __tmp_key_buf[DBG_KEY_BUF_LEN];                                   \
+	pr_debug("UBIFS DBG " type " (pid %d): " fmt "%s\n", current->pid,     \
+		 ##__VA_ARGS__,                                                \
+		 dbg_snprintf_key(c, key, __tmp_key_buf, DBG_KEY_BUF_LEN));    \
+} while (0)
+
+/* General messages */
+#define dbg_gen(fmt, ...)   ubifs_dbg_msg("gen", fmt, ##__VA_ARGS__)
+/* Additional journal messages */
+#define dbg_jnl(fmt, ...)   ubifs_dbg_msg("jnl", fmt, ##__VA_ARGS__)
+#define dbg_jnlk(key, fmt, ...) \
+	ubifs_dbg_msg_key("jnl", key, fmt, ##__VA_ARGS__)
+/* Additional TNC messages */
+#define dbg_tnc(fmt, ...)   ubifs_dbg_msg("tnc", fmt, ##__VA_ARGS__)
+#define dbg_tnck(key, fmt, ...) \
+	ubifs_dbg_msg_key("tnc", key, fmt, ##__VA_ARGS__)
+/* Additional lprops messages */
+#define dbg_lp(fmt, ...)    ubifs_dbg_msg("lp", fmt, ##__VA_ARGS__)
+/* Additional LEB find messages */
+#define dbg_find(fmt, ...)  ubifs_dbg_msg("find", fmt, ##__VA_ARGS__)
+/* Additional mount messages */
+#define dbg_mnt(fmt, ...)   ubifs_dbg_msg("mnt", fmt, ##__VA_ARGS__)
+#define dbg_mntk(key, fmt, ...) \
+	ubifs_dbg_msg_key("mnt", key, fmt, ##__VA_ARGS__)
+/* Additional I/O messages */
+#define dbg_io(fmt, ...)    ubifs_dbg_msg("io", fmt, ##__VA_ARGS__)
+/* Additional commit messages */
+#define dbg_cmt(fmt, ...)   ubifs_dbg_msg("cmt", fmt, ##__VA_ARGS__)
+/* Additional budgeting messages */
+#define dbg_budg(fmt, ...)  ubifs_dbg_msg("budg", fmt, ##__VA_ARGS__)
+/* Additional log messages */
+#define dbg_log(fmt, ...)   ubifs_dbg_msg("log", fmt, ##__VA_ARGS__)
+/* Additional gc messages */
+#define dbg_gc(fmt, ...)    ubifs_dbg_msg("gc", fmt, ##__VA_ARGS__)
+/* Additional scan messages */
+#define dbg_scan(fmt, ...)  ubifs_dbg_msg("scan", fmt, ##__VA_ARGS__)
+/* Additional recovery messages */
+#define dbg_rcvry(fmt, ...) ubifs_dbg_msg("rcvry", fmt, ##__VA_ARGS__)
+
+extern struct ubifs_global_debug_info ubifs_dbg;
+
+static inline int dbg_is_chk_gen(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.chk_gen || c->dbg->chk_gen);
+}
+static inline int dbg_is_chk_index(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.chk_index || c->dbg->chk_index);
+}
+static inline int dbg_is_chk_orph(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.chk_orph || c->dbg->chk_orph);
+}
+static inline int dbg_is_chk_lprops(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.chk_lprops || c->dbg->chk_lprops);
+}
+static inline int dbg_is_chk_fs(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.chk_fs || c->dbg->chk_fs);
+}
+static inline int dbg_is_tst_rcvry(const struct ubifs_info *c)
+{
+	return !!(ubifs_dbg.tst_rcvry || c->dbg->tst_rcvry);
+}
+static inline int dbg_is_power_cut(const struct ubifs_info *c)
+{
+	return !!c->dbg->pc_happened;
+}
+
+int ubifs_debugging_init(struct ubifs_info *c);
+void ubifs_debugging_exit(struct ubifs_info *c);
+
+/* Dump functions */
+const char *dbg_ntype(int type);
+const char *dbg_cstate(int cmt_state);
+const char *dbg_jhead(int jhead);
+const char *dbg_get_key_dump(const struct ubifs_info *c,
+			     const union ubifs_key *key);
+const char *dbg_snprintf_key(const struct ubifs_info *c,
+			     const union ubifs_key *key, char *buffer, int len);
+void ubifs_dump_inode(struct ubifs_info *c, const struct inode *inode);
+void ubifs_dump_node(const struct ubifs_info *c, const void *node,
+		     int node_len);
+void ubifs_dump_budget_req(const struct ubifs_budget_req *req);
+void ubifs_dump_lstats(const struct ubifs_lp_stats *lst);
+void ubifs_dump_budg(struct ubifs_info *c, const struct ubifs_budg_info *bi);
+void ubifs_dump_lprop(const struct ubifs_info *c,
+		      const struct ubifs_lprops *lp);
+void ubifs_dump_lprops(struct ubifs_info *c);
+void ubifs_dump_lpt_info(struct ubifs_info *c);
+void ubifs_dump_leb(const struct ubifs_info *c, int lnum);
+void ubifs_dump_znode(const struct ubifs_info *c,
+		      const struct ubifs_znode *znode);
+void ubifs_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
+		     int cat);
+void ubifs_dump_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
+		      struct ubifs_nnode *parent, int iip);
+void ubifs_dump_tnc(struct ubifs_info *c);
+void ubifs_dump_index(struct ubifs_info *c);
+void ubifs_dump_lpt_lebs(const struct ubifs_info *c);
+
+int dbg_walk_index(struct ubifs_info *c, dbg_leaf_callback leaf_cb,
+		   dbg_znode_callback znode_cb, void *priv);
+
+/* Checking functions */
+void dbg_save_space_info(struct ubifs_info *c);
+int dbg_check_space_info(struct ubifs_info *c);
+int dbg_check_lprops(struct ubifs_info *c);
+int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot);
+int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot);
+int dbg_check_cats(struct ubifs_info *c);
+int dbg_check_ltab(struct ubifs_info *c);
+int dbg_chk_lpt_free_spc(struct ubifs_info *c);
+int dbg_chk_lpt_sz(struct ubifs_info *c, int action, int len);
+int dbg_check_synced_i_size(const struct ubifs_info *c, struct inode *inode);
+int dbg_check_dir(struct ubifs_info *c, const struct inode *dir);
+int dbg_check_tnc(struct ubifs_info *c, int extra);
+int dbg_check_idx_size(struct ubifs_info *c, long long idx_size);
+int dbg_check_filesystem(struct ubifs_info *c);
+void dbg_check_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat,
+		    int add_pos);
+int dbg_check_lpt_nodes(struct ubifs_info *c, struct ubifs_cnode *cnode,
+			int row, int col);
+int dbg_check_inode_size(struct ubifs_info *c, const struct inode *inode,
+			 loff_t size);
+int dbg_check_data_nodes_order(struct ubifs_info *c, struct list_head *head);
+int dbg_check_nondata_nodes_order(struct ubifs_info *c, struct list_head *head);
+
+int dbg_leb_write(struct ubifs_info *c, int lnum, const void *buf, int offs,
+		  int len);
+int dbg_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len);
+int dbg_leb_unmap(struct ubifs_info *c, int lnum);
+int dbg_leb_map(struct ubifs_info *c, int lnum);
+
+/* Debugfs-related stuff */
+void dbg_debugfs_init(void);
+void dbg_debugfs_exit(void);
+void dbg_debugfs_init_fs(struct ubifs_info *c);
+void dbg_debugfs_exit_fs(struct ubifs_info *c);
+
+#endif /* !__UBIFS_DEBUG_H__ */

linux-7.2.3/fs/ubifs/find.c

@@ -0,0 +1,963 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+/*
+ * This file contains functions for finding LEBs for various purposes e.g.
+ * garbage collection. In general, lprops category heaps and lists are used
+ * for fast access, falling back on scanning the LPT as a last resort.
+ */
+
+#include <linux/sort.h>
+#include "ubifs.h"
+
+/**
+ * struct scan_data - data provided to scan callback functions
+ * @min_space: minimum number of bytes for which to scan
+ * @pick_free: whether it is OK to scan for empty LEBs
+ * @lnum: LEB number found is returned here
+ * @exclude_index: whether to exclude index LEBs
+ */
+struct scan_data {
+	int min_space;
+	int pick_free;
+	int lnum;
+	int exclude_index;
+};
+
+/**
+ * valuable - determine whether LEB properties are valuable.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties
+ *
+ * This function return %1 if the LEB properties should be added to the LEB
+ * properties tree in memory. Otherwise %0 is returned.
+ */
+static int valuable(struct ubifs_info *c, const struct ubifs_lprops *lprops)
+{
+	int n, cat = lprops->flags & LPROPS_CAT_MASK;
+	struct ubifs_lpt_heap *heap;
+
+	switch (cat) {
+	case LPROPS_DIRTY:
+	case LPROPS_DIRTY_IDX:
+	case LPROPS_FREE:
+		heap = &c->lpt_heap[cat - 1];
+		if (heap->cnt < heap->max_cnt)
+			return 1;
+		if (lprops->free + lprops->dirty >= c->dark_wm)
+			return 1;
+		return 0;
+	case LPROPS_EMPTY:
+		n = c->lst.empty_lebs + c->freeable_cnt -
+		    c->lst.taken_empty_lebs;
+		if (n < c->lsave_cnt)
+			return 1;
+		return 0;
+	case LPROPS_FREEABLE:
+		return 1;
+	case LPROPS_FRDI_IDX:
+		return 1;
+	}
+	return 0;
+}
+
+/**
+ * scan_for_dirty_cb - dirty space scan callback.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to scan
+ * @in_tree: whether the LEB properties are in main memory
+ * @arg: information passed to and from the caller of the scan
+ *
+ * This function returns a code that indicates whether the scan should continue
+ * (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
+ * in main memory (%LPT_SCAN_ADD), or whether the scan should stop
+ * (%LPT_SCAN_STOP).
+ */
+static int scan_for_dirty_cb(struct ubifs_info *c,
+			     const struct ubifs_lprops *lprops, int in_tree,
+			     void *arg)
+{
+	struct scan_data *data = arg;
+	int ret = LPT_SCAN_CONTINUE;
+
+	/* Exclude LEBs that are currently in use */
+	if (lprops->flags & LPROPS_TAKEN)
+		return LPT_SCAN_CONTINUE;
+	/* Determine whether to add these LEB properties to the tree */
+	if (!in_tree && valuable(c, lprops))
+		ret |= LPT_SCAN_ADD;
+	/* Exclude LEBs with too little space */
+	if (lprops->free + lprops->dirty < data->min_space)
+		return ret;
+	/* If specified, exclude index LEBs */
+	if (data->exclude_index && lprops->flags & LPROPS_INDEX)
+		return ret;
+	/* If specified, exclude empty or freeable LEBs */
+	if (lprops->free + lprops->dirty == c->leb_size) {
+		if (!data->pick_free)
+			return ret;
+	/* Exclude LEBs with too little dirty space (unless it is empty) */
+	} else if (lprops->dirty < c->dead_wm)
+		return ret;
+	/* Finally we found space */
+	data->lnum = lprops->lnum;
+	return LPT_SCAN_ADD | LPT_SCAN_STOP;
+}
+
+/**
+ * scan_for_dirty - find a data LEB with free space.
+ * @c: the UBIFS file-system description object
+ * @min_space: minimum amount free plus dirty space the returned LEB has to
+ *             have
+ * @pick_free: if it is OK to return a free or freeable LEB
+ * @exclude_index: whether to exclude index LEBs
+ *
+ * This function returns a pointer to the LEB properties found or a negative
+ * error code.
+ */
+static const struct ubifs_lprops *scan_for_dirty(struct ubifs_info *c,
+						 int min_space, int pick_free,
+						 int exclude_index)
+{
+	const struct ubifs_lprops *lprops;
+	struct ubifs_lpt_heap *heap;
+	struct scan_data data;
+	int err, i;
+
+	/* There may be an LEB with enough dirty space on the free heap */
+	heap = &c->lpt_heap[LPROPS_FREE - 1];
+	for (i = 0; i < heap->cnt; i++) {
+		lprops = heap->arr[i];
+		if (lprops->free + lprops->dirty < min_space)
+			continue;
+		if (lprops->dirty < c->dead_wm)
+			continue;
+		return lprops;
+	}
+	/*
+	 * A LEB may have fallen off of the bottom of the dirty heap, and ended
+	 * up as uncategorized even though it has enough dirty space for us now,
+	 * so check the uncategorized list. N.B. neither empty nor freeable LEBs
+	 * can end up as uncategorized because they are kept on lists not
+	 * finite-sized heaps.
+	 */
+	list_for_each_entry(lprops, &c->uncat_list, list) {
+		if (lprops->flags & LPROPS_TAKEN)
+			continue;
+		if (lprops->free + lprops->dirty < min_space)
+			continue;
+		if (exclude_index && (lprops->flags & LPROPS_INDEX))
+			continue;
+		if (lprops->dirty < c->dead_wm)
+			continue;
+		return lprops;
+	}
+	/* We have looked everywhere in main memory, now scan the flash */
+	if (c->pnodes_have >= c->pnode_cnt)
+		/* All pnodes are in memory, so skip scan */
+		return ERR_PTR(-ENOSPC);
+	data.min_space = min_space;
+	data.pick_free = pick_free;
+	data.lnum = -1;
+	data.exclude_index = exclude_index;
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_dirty_cb,
+				    &data);
+	if (err)
+		return ERR_PTR(err);
+	ubifs_assert(c, data.lnum >= c->main_first && data.lnum < c->leb_cnt);
+	c->lscan_lnum = data.lnum;
+	lprops = ubifs_lpt_lookup_dirty(c, data.lnum);
+	if (IS_ERR(lprops))
+		return lprops;
+	ubifs_assert(c, lprops->lnum == data.lnum);
+	ubifs_assert(c, lprops->free + lprops->dirty >= min_space);
+	ubifs_assert(c, lprops->dirty >= c->dead_wm ||
+		     (pick_free &&
+		      lprops->free + lprops->dirty == c->leb_size));
+	ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+	ubifs_assert(c, !exclude_index || !(lprops->flags & LPROPS_INDEX));
+	return lprops;
+}
+
+/**
+ * ubifs_find_dirty_leb - find a dirty LEB for the Garbage Collector.
+ * @c: the UBIFS file-system description object
+ * @ret_lp: LEB properties are returned here on exit
+ * @min_space: minimum amount free plus dirty space the returned LEB has to
+ *             have
+ * @pick_free: controls whether it is OK to pick empty or index LEBs
+ *
+ * This function tries to find a dirty logical eraseblock which has at least
+ * @min_space free and dirty space. It prefers to take an LEB from the dirty or
+ * dirty index heap, and it falls-back to LPT scanning if the heaps are empty
+ * or do not have an LEB which satisfies the @min_space criteria.
+ *
+ * Note, LEBs which have less than dead watermark of free + dirty space are
+ * never picked by this function.
+ *
+ * The additional @pick_free argument controls if this function has to return a
+ * free or freeable LEB if one is present. For example, GC must to set it to %1,
+ * when called from the journal space reservation function, because the
+ * appearance of free space may coincide with the loss of enough dirty space
+ * for GC to succeed anyway.
+ *
+ * In contrast, if the Garbage Collector is called from budgeting, it should
+ * just make free space, not return LEBs which are already free or freeable.
+ *
+ * In addition @pick_free is set to %2 by the recovery process in order to
+ * recover gc_lnum in which case an index LEB must not be returned.
+ *
+ * This function returns zero and the LEB properties of found dirty LEB in case
+ * of success, %-ENOSPC if no dirty LEB was found and a negative error code in
+ * case of other failures. The returned LEB is marked as "taken".
+ */
+int ubifs_find_dirty_leb(struct ubifs_info *c, struct ubifs_lprops *ret_lp,
+			 int min_space, int pick_free)
+{
+	int err = 0, sum, exclude_index = pick_free == 2 ? 1 : 0;
+	const struct ubifs_lprops *lp = NULL, *idx_lp = NULL;
+	struct ubifs_lpt_heap *heap, *idx_heap;
+
+	ubifs_get_lprops(c);
+
+	if (pick_free) {
+		int lebs, rsvd_idx_lebs = 0;
+
+		spin_lock(&c->space_lock);
+		lebs = c->lst.empty_lebs + c->idx_gc_cnt;
+		lebs += c->freeable_cnt - c->lst.taken_empty_lebs;
+
+		/*
+		 * Note, the index may consume more LEBs than have been reserved
+		 * for it. It is OK because it might be consolidated by GC.
+		 * But if the index takes fewer LEBs than it is reserved for it,
+		 * this function must avoid picking those reserved LEBs.
+		 */
+		if (c->bi.min_idx_lebs >= c->lst.idx_lebs) {
+			rsvd_idx_lebs = c->bi.min_idx_lebs -  c->lst.idx_lebs;
+			exclude_index = 1;
+		}
+		spin_unlock(&c->space_lock);
+
+		/* Check if there are enough free LEBs for the index */
+		if (rsvd_idx_lebs < lebs) {
+			/* OK, try to find an empty LEB */
+			lp = ubifs_fast_find_empty(c);
+			if (lp)
+				goto found;
+
+			/* Or a freeable LEB */
+			lp = ubifs_fast_find_freeable(c);
+			if (lp)
+				goto found;
+		} else
+			/*
+			 * We cannot pick free/freeable LEBs in the below code.
+			 */
+			pick_free = 0;
+	} else {
+		spin_lock(&c->space_lock);
+		exclude_index = (c->bi.min_idx_lebs >= c->lst.idx_lebs);
+		spin_unlock(&c->space_lock);
+	}
+
+	/* Look on the dirty and dirty index heaps */
+	heap = &c->lpt_heap[LPROPS_DIRTY - 1];
+	idx_heap = &c->lpt_heap[LPROPS_DIRTY_IDX - 1];
+
+	if (idx_heap->cnt && !exclude_index) {
+		idx_lp = idx_heap->arr[0];
+		sum = idx_lp->free + idx_lp->dirty;
+		/*
+		 * Since we reserve thrice as much space for the index than it
+		 * actually takes, it does not make sense to pick indexing LEBs
+		 * with less than, say, half LEB of dirty space. May be half is
+		 * not the optimal boundary - this should be tested and
+		 * checked. This boundary should determine how much we use
+		 * in-the-gaps to consolidate the index comparing to how much
+		 * we use garbage collector to consolidate it. The "half"
+		 * criteria just feels to be fine.
+		 */
+		if (sum < min_space || sum < c->half_leb_size)
+			idx_lp = NULL;
+	}
+
+	if (heap->cnt) {
+		lp = heap->arr[0];
+		if (lp->dirty + lp->free < min_space)
+			lp = NULL;
+	}
+
+	/* Pick the LEB with most space */
+	if (idx_lp && lp) {
+		if (idx_lp->free + idx_lp->dirty >= lp->free + lp->dirty)
+			lp = idx_lp;
+	} else if (idx_lp && !lp)
+		lp = idx_lp;
+
+	if (lp) {
+		ubifs_assert(c, lp->free + lp->dirty >= c->dead_wm);
+		goto found;
+	}
+
+	/* Did not find a dirty LEB on the dirty heaps, have to scan */
+	dbg_find("scanning LPT for a dirty LEB");
+	lp = scan_for_dirty(c, min_space, pick_free, exclude_index);
+	if (IS_ERR(lp)) {
+		err = PTR_ERR(lp);
+		goto out;
+	}
+	ubifs_assert(c, lp->dirty >= c->dead_wm ||
+		     (pick_free && lp->free + lp->dirty == c->leb_size));
+
+found:
+	dbg_find("found LEB %d, free %d, dirty %d, flags %#x",
+		 lp->lnum, lp->free, lp->dirty, lp->flags);
+
+	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;
+	}
+
+	memcpy(ret_lp, lp, sizeof(struct ubifs_lprops));
+
+out:
+	ubifs_release_lprops(c);
+	return err;
+}
+
+/**
+ * scan_for_free_cb - free space scan callback.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to scan
+ * @in_tree: whether the LEB properties are in main memory
+ * @arg: information passed to and from the caller of the scan
+ *
+ * This function returns a code that indicates whether the scan should continue
+ * (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
+ * in main memory (%LPT_SCAN_ADD), or whether the scan should stop
+ * (%LPT_SCAN_STOP).
+ */
+static int scan_for_free_cb(struct ubifs_info *c,
+			    const struct ubifs_lprops *lprops, int in_tree,
+			    void *arg)
+{
+	struct scan_data *data = arg;
+	int ret = LPT_SCAN_CONTINUE;
+
+	/* Exclude LEBs that are currently in use */
+	if (lprops->flags & LPROPS_TAKEN)
+		return LPT_SCAN_CONTINUE;
+	/* Determine whether to add these LEB properties to the tree */
+	if (!in_tree && valuable(c, lprops))
+		ret |= LPT_SCAN_ADD;
+	/* Exclude index LEBs */
+	if (lprops->flags & LPROPS_INDEX)
+		return ret;
+	/* Exclude LEBs with too little space */
+	if (lprops->free < data->min_space)
+		return ret;
+	/* If specified, exclude empty LEBs */
+	if (!data->pick_free && lprops->free == c->leb_size)
+		return ret;
+	/*
+	 * LEBs that have only free and dirty space must not be allocated
+	 * because they may have been unmapped already or they may have data
+	 * that is obsolete only because of nodes that are still sitting in a
+	 * wbuf.
+	 */
+	if (lprops->free + lprops->dirty == c->leb_size && lprops->dirty > 0)
+		return ret;
+	/* Finally we found space */
+	data->lnum = lprops->lnum;
+	return LPT_SCAN_ADD | LPT_SCAN_STOP;
+}
+
+/**
+ * do_find_free_space - find a data LEB with free space.
+ * @c: the UBIFS file-system description object
+ * @min_space: minimum amount of free space required
+ * @pick_free: whether it is OK to scan for empty LEBs
+ * @squeeze: whether to try to find space in a non-empty LEB first
+ *
+ * This function returns a pointer to the LEB properties found or a negative
+ * error code.
+ */
+static
+const struct ubifs_lprops *do_find_free_space(struct ubifs_info *c,
+					      int min_space, int pick_free,
+					      int squeeze)
+{
+	const struct ubifs_lprops *lprops;
+	struct ubifs_lpt_heap *heap;
+	struct scan_data data;
+	int err, i;
+
+	if (squeeze) {
+		lprops = ubifs_fast_find_free(c);
+		if (lprops && lprops->free >= min_space)
+			return lprops;
+	}
+	if (pick_free) {
+		lprops = ubifs_fast_find_empty(c);
+		if (lprops)
+			return lprops;
+	}
+	if (!squeeze) {
+		lprops = ubifs_fast_find_free(c);
+		if (lprops && lprops->free >= min_space)
+			return lprops;
+	}
+	/* There may be an LEB with enough free space on the dirty heap */
+	heap = &c->lpt_heap[LPROPS_DIRTY - 1];
+	for (i = 0; i < heap->cnt; i++) {
+		lprops = heap->arr[i];
+		if (lprops->free >= min_space)
+			return lprops;
+	}
+	/*
+	 * A LEB may have fallen off of the bottom of the free heap, and ended
+	 * up as uncategorized even though it has enough free space for us now,
+	 * so check the uncategorized list. N.B. neither empty nor freeable LEBs
+	 * can end up as uncategorized because they are kept on lists not
+	 * finite-sized heaps.
+	 */
+	list_for_each_entry(lprops, &c->uncat_list, list) {
+		if (lprops->flags & LPROPS_TAKEN)
+			continue;
+		if (lprops->flags & LPROPS_INDEX)
+			continue;
+		if (lprops->free >= min_space)
+			return lprops;
+	}
+	/* We have looked everywhere in main memory, now scan the flash */
+	if (c->pnodes_have >= c->pnode_cnt)
+		/* All pnodes are in memory, so skip scan */
+		return ERR_PTR(-ENOSPC);
+	data.min_space = min_space;
+	data.pick_free = pick_free;
+	data.lnum = -1;
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum,
+				    scan_for_free_cb,
+				    &data);
+	if (err)
+		return ERR_PTR(err);
+	ubifs_assert(c, data.lnum >= c->main_first && data.lnum < c->leb_cnt);
+	c->lscan_lnum = data.lnum;
+	lprops = ubifs_lpt_lookup_dirty(c, data.lnum);
+	if (IS_ERR(lprops))
+		return lprops;
+	ubifs_assert(c, lprops->lnum == data.lnum);
+	ubifs_assert(c, lprops->free >= min_space);
+	ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+	ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+	return lprops;
+}
+
+/**
+ * ubifs_find_free_space - find a data LEB with free space.
+ * @c: the UBIFS file-system description object
+ * @min_space: minimum amount of required free space
+ * @offs: contains offset of where free space starts on exit
+ * @squeeze: whether to try to find space in a non-empty LEB first
+ *
+ * This function looks for an LEB with at least @min_space bytes of free space.
+ * It tries to find an empty LEB if possible. If no empty LEBs are available,
+ * this function searches for a non-empty data LEB. The returned LEB is marked
+ * as "taken".
+ *
+ * This function returns found LEB number in case of success, %-ENOSPC if it
+ * failed to find a LEB with @min_space bytes of free space and other a negative
+ * error codes in case of failure.
+ */
+int ubifs_find_free_space(struct ubifs_info *c, int min_space, int *offs,
+			  int squeeze)
+{
+	const struct ubifs_lprops *lprops;
+	int lebs, rsvd_idx_lebs, pick_free = 0, err, lnum, flags;
+
+	dbg_find("min_space %d", min_space);
+	ubifs_get_lprops(c);
+
+	/* Check if there are enough empty LEBs for commit */
+	spin_lock(&c->space_lock);
+	if (c->bi.min_idx_lebs > c->lst.idx_lebs)
+		rsvd_idx_lebs = c->bi.min_idx_lebs -  c->lst.idx_lebs;
+	else
+		rsvd_idx_lebs = 0;
+	lebs = c->lst.empty_lebs + c->freeable_cnt + c->idx_gc_cnt -
+	       c->lst.taken_empty_lebs;
+	if (rsvd_idx_lebs < lebs)
+		/*
+		 * OK to allocate an empty LEB, but we still don't want to go
+		 * looking for one if there aren't any.
+		 */
+		if (c->lst.empty_lebs - c->lst.taken_empty_lebs > 0) {
+			pick_free = 1;
+			/*
+			 * Because we release the space lock, we must account
+			 * for this allocation here. After the LEB properties
+			 * flags have been updated, we subtract one. Note, the
+			 * result of this is that lprops also decreases
+			 * @taken_empty_lebs in 'ubifs_change_lp()', so it is
+			 * off by one for a short period of time which may
+			 * introduce a small disturbance to budgeting
+			 * calculations, but this is harmless because at the
+			 * worst case this would make the budgeting subsystem
+			 * be more pessimistic than needed.
+			 *
+			 * Fundamentally, this is about serialization of the
+			 * budgeting and lprops subsystems. We could make the
+			 * @space_lock a mutex and avoid dropping it before
+			 * calling 'ubifs_change_lp()', but mutex is more
+			 * heavy-weight, and we want budgeting to be as fast as
+			 * possible.
+			 */
+			c->lst.taken_empty_lebs += 1;
+		}
+	spin_unlock(&c->space_lock);
+
+	lprops = do_find_free_space(c, min_space, pick_free, squeeze);
+	if (IS_ERR(lprops)) {
+		err = PTR_ERR(lprops);
+		goto out;
+	}
+
+	lnum = lprops->lnum;
+	flags = lprops->flags | LPROPS_TAKEN;
+
+	lprops = ubifs_change_lp(c, lprops, LPROPS_NC, LPROPS_NC, flags, 0);
+	if (IS_ERR(lprops)) {
+		err = PTR_ERR(lprops);
+		goto out;
+	}
+
+	if (pick_free) {
+		spin_lock(&c->space_lock);
+		c->lst.taken_empty_lebs -= 1;
+		spin_unlock(&c->space_lock);
+	}
+
+	*offs = c->leb_size - lprops->free;
+	ubifs_release_lprops(c);
+
+	if (*offs == 0) {
+		/*
+		 * Ensure that empty LEBs have been unmapped. They may not have
+		 * been, for example, because of an unclean unmount.  Also
+		 * LEBs that were freeable LEBs (free + dirty == leb_size) will
+		 * not have been unmapped.
+		 */
+		err = ubifs_leb_unmap(c, lnum);
+		if (err)
+			return err;
+	}
+
+	dbg_find("found LEB %d, free %d", lnum, c->leb_size - *offs);
+	ubifs_assert(c, *offs <= c->leb_size - min_space);
+	return lnum;
+
+out:
+	if (pick_free) {
+		spin_lock(&c->space_lock);
+		c->lst.taken_empty_lebs -= 1;
+		spin_unlock(&c->space_lock);
+	}
+	ubifs_release_lprops(c);
+	return err;
+}
+
+/**
+ * scan_for_idx_cb - callback used by the scan for a free LEB for the index.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to scan
+ * @in_tree: whether the LEB properties are in main memory
+ * @arg: information passed to and from the caller of the scan
+ *
+ * This function returns a code that indicates whether the scan should continue
+ * (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
+ * in main memory (%LPT_SCAN_ADD), or whether the scan should stop
+ * (%LPT_SCAN_STOP).
+ */
+static int scan_for_idx_cb(struct ubifs_info *c,
+			   const struct ubifs_lprops *lprops, int in_tree,
+			   void *arg)
+{
+	struct scan_data *data = arg;
+	int ret = LPT_SCAN_CONTINUE;
+
+	/* Exclude LEBs that are currently in use */
+	if (lprops->flags & LPROPS_TAKEN)
+		return LPT_SCAN_CONTINUE;
+	/* Determine whether to add these LEB properties to the tree */
+	if (!in_tree && valuable(c, lprops))
+		ret |= LPT_SCAN_ADD;
+	/* Exclude index LEBS */
+	if (lprops->flags & LPROPS_INDEX)
+		return ret;
+	/* Exclude LEBs that cannot be made empty */
+	if (lprops->free + lprops->dirty != c->leb_size)
+		return ret;
+	/*
+	 * We are allocating for the index so it is safe to allocate LEBs with
+	 * only free and dirty space, because write buffers are sync'd at commit
+	 * start.
+	 */
+	data->lnum = lprops->lnum;
+	return LPT_SCAN_ADD | LPT_SCAN_STOP;
+}
+
+/**
+ * scan_for_leb_for_idx - scan for a free LEB for the index.
+ * @c: the UBIFS file-system description object
+ */
+static const struct ubifs_lprops *scan_for_leb_for_idx(struct ubifs_info *c)
+{
+	const struct ubifs_lprops *lprops;
+	struct scan_data data;
+	int err;
+
+	data.lnum = -1;
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_idx_cb,
+				    &data);
+	if (err)
+		return ERR_PTR(err);
+	ubifs_assert(c, data.lnum >= c->main_first && data.lnum < c->leb_cnt);
+	c->lscan_lnum = data.lnum;
+	lprops = ubifs_lpt_lookup_dirty(c, data.lnum);
+	if (IS_ERR(lprops))
+		return lprops;
+	ubifs_assert(c, lprops->lnum == data.lnum);
+	ubifs_assert(c, lprops->free + lprops->dirty == c->leb_size);
+	ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+	ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+	return lprops;
+}
+
+/**
+ * ubifs_find_free_leb_for_idx - find a free LEB for the index.
+ * @c: the UBIFS file-system description object
+ *
+ * This function looks for a free LEB and returns that LEB number. The returned
+ * LEB is marked as "taken", "index".
+ *
+ * Only empty LEBs are allocated. This is for two reasons. First, the commit
+ * calculates the number of LEBs to allocate based on the assumption that they
+ * will be empty. Secondly, free space at the end of an index LEB is not
+ * guaranteed to be empty because it may have been used by the in-the-gaps
+ * method prior to an unclean unmount.
+ *
+ * If no LEB is found %-ENOSPC is returned. For other failures another negative
+ * error code is returned.
+ */
+int ubifs_find_free_leb_for_idx(struct ubifs_info *c)
+{
+	const struct ubifs_lprops *lprops;
+	int lnum = -1, err, flags;
+
+	ubifs_get_lprops(c);
+
+	lprops = ubifs_fast_find_empty(c);
+	if (!lprops) {
+		lprops = ubifs_fast_find_freeable(c);
+		if (!lprops) {
+			/*
+			 * The first condition means the following: go scan the
+			 * LPT if there are uncategorized lprops, which means
+			 * there may be freeable LEBs there (UBIFS does not
+			 * store the information about freeable LEBs in the
+			 * master node).
+			 */
+			if (c->in_a_category_cnt != c->main_lebs ||
+			    c->lst.empty_lebs - c->lst.taken_empty_lebs > 0) {
+				ubifs_assert(c, c->freeable_cnt == 0);
+				lprops = scan_for_leb_for_idx(c);
+				if (IS_ERR(lprops)) {
+					err = PTR_ERR(lprops);
+					goto out;
+				}
+			}
+		}
+	}
+
+	if (!lprops) {
+		err = -ENOSPC;
+		goto out;
+	}
+
+	lnum = lprops->lnum;
+
+	dbg_find("found LEB %d, free %d, dirty %d, flags %#x",
+		 lnum, lprops->free, lprops->dirty, lprops->flags);
+
+	flags = lprops->flags | LPROPS_TAKEN | LPROPS_INDEX;
+	lprops = ubifs_change_lp(c, lprops, c->leb_size, 0, flags, 0);
+	if (IS_ERR(lprops)) {
+		err = PTR_ERR(lprops);
+		goto out;
+	}
+
+	ubifs_release_lprops(c);
+
+	/*
+	 * Ensure that empty LEBs have been unmapped. They may not have been,
+	 * for example, because of an unclean unmount. Also LEBs that were
+	 * freeable LEBs (free + dirty == leb_size) will not have been unmapped.
+	 */
+	err = ubifs_leb_unmap(c, lnum);
+	if (err) {
+		ubifs_change_one_lp(c, lnum, LPROPS_NC, LPROPS_NC, 0,
+				    LPROPS_TAKEN | LPROPS_INDEX, 0);
+		return err;
+	}
+
+	return lnum;
+
+out:
+	ubifs_release_lprops(c);
+	return err;
+}
+
+static int cmp_dirty_idx(const void *a, const void *b)
+{
+	const struct ubifs_lprops *lpa = *(const struct ubifs_lprops **)a;
+	const struct ubifs_lprops *lpb = *(const struct ubifs_lprops **)b;
+
+	return lpa->dirty + lpa->free - lpb->dirty - lpb->free;
+}
+
+/**
+ * ubifs_save_dirty_idx_lnums - save an array of the most dirty index LEB nos.
+ * @c: the UBIFS file-system description object
+ *
+ * This function is called each commit to create an array of LEB numbers of
+ * dirty index LEBs sorted in order of dirty and free space.  This is used by
+ * the in-the-gaps method of TNC commit.
+ */
+int ubifs_save_dirty_idx_lnums(struct ubifs_info *c)
+{
+	int i;
+
+	ubifs_get_lprops(c);
+	/* Copy the LPROPS_DIRTY_IDX heap */
+	c->dirty_idx.cnt = c->lpt_heap[LPROPS_DIRTY_IDX - 1].cnt;
+	memcpy(c->dirty_idx.arr, c->lpt_heap[LPROPS_DIRTY_IDX - 1].arr,
+	       sizeof(void *) * c->dirty_idx.cnt);
+	/* Sort it so that the dirtiest is now at the end */
+	sort(c->dirty_idx.arr, c->dirty_idx.cnt, sizeof(void *),
+	     cmp_dirty_idx, NULL);
+	dbg_find("found %d dirty index LEBs", c->dirty_idx.cnt);
+	if (c->dirty_idx.cnt)
+		dbg_find("dirtiest index LEB is %d with dirty %d and free %d",
+			 c->dirty_idx.arr[c->dirty_idx.cnt - 1]->lnum,
+			 c->dirty_idx.arr[c->dirty_idx.cnt - 1]->dirty,
+			 c->dirty_idx.arr[c->dirty_idx.cnt - 1]->free);
+	/* Replace the lprops pointers with LEB numbers */
+	for (i = 0; i < c->dirty_idx.cnt; i++)
+		c->dirty_idx.arr[i] = (void *)(size_t)c->dirty_idx.arr[i]->lnum;
+	ubifs_release_lprops(c);
+	return 0;
+}
+
+/**
+ * scan_dirty_idx_cb - callback used by the scan for a dirty index LEB.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to scan
+ * @in_tree: whether the LEB properties are in main memory
+ * @arg: information passed to and from the caller of the scan
+ *
+ * This function returns a code that indicates whether the scan should continue
+ * (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
+ * in main memory (%LPT_SCAN_ADD), or whether the scan should stop
+ * (%LPT_SCAN_STOP).
+ */
+static int scan_dirty_idx_cb(struct ubifs_info *c,
+			   const struct ubifs_lprops *lprops, int in_tree,
+			   void *arg)
+{
+	struct scan_data *data = arg;
+	int ret = LPT_SCAN_CONTINUE;
+
+	/* Exclude LEBs that are currently in use */
+	if (lprops->flags & LPROPS_TAKEN)
+		return LPT_SCAN_CONTINUE;
+	/* Determine whether to add these LEB properties to the tree */
+	if (!in_tree && valuable(c, lprops))
+		ret |= LPT_SCAN_ADD;
+	/* Exclude non-index LEBs */
+	if (!(lprops->flags & LPROPS_INDEX))
+		return ret;
+	/* Exclude LEBs with too little space */
+	if (lprops->free + lprops->dirty < c->min_idx_node_sz)
+		return ret;
+	/* Finally we found space */
+	data->lnum = lprops->lnum;
+	return LPT_SCAN_ADD | LPT_SCAN_STOP;
+}
+
+/**
+ * find_dirty_idx_leb - find a dirty index LEB.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB number upon success and a negative error code upon
+ * failure.  In particular, -ENOSPC is returned if a dirty index LEB is not
+ * found.
+ *
+ * Note that this function scans the entire LPT but it is called very rarely.
+ */
+static int find_dirty_idx_leb(struct ubifs_info *c)
+{
+	const struct ubifs_lprops *lprops;
+	struct ubifs_lpt_heap *heap;
+	struct scan_data data;
+	int err, i, ret;
+
+	/* Check all structures in memory first */
+	data.lnum = -1;
+	heap = &c->lpt_heap[LPROPS_DIRTY_IDX - 1];
+	for (i = 0; i < heap->cnt; i++) {
+		lprops = heap->arr[i];
+		ret = scan_dirty_idx_cb(c, lprops, 1, &data);
+		if (ret & LPT_SCAN_STOP)
+			goto found;
+	}
+	list_for_each_entry(lprops, &c->frdi_idx_list, list) {
+		ret = scan_dirty_idx_cb(c, lprops, 1, &data);
+		if (ret & LPT_SCAN_STOP)
+			goto found;
+	}
+	list_for_each_entry(lprops, &c->uncat_list, list) {
+		ret = scan_dirty_idx_cb(c, lprops, 1, &data);
+		if (ret & LPT_SCAN_STOP)
+			goto found;
+	}
+	if (c->pnodes_have >= c->pnode_cnt)
+		/* All pnodes are in memory, so skip scan */
+		return -ENOSPC;
+	err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_dirty_idx_cb,
+				    &data);
+	if (err)
+		return err;
+found:
+	ubifs_assert(c, data.lnum >= c->main_first && data.lnum < c->leb_cnt);
+	c->lscan_lnum = data.lnum;
+	lprops = ubifs_lpt_lookup_dirty(c, data.lnum);
+	if (IS_ERR(lprops))
+		return PTR_ERR(lprops);
+	ubifs_assert(c, lprops->lnum == data.lnum);
+	ubifs_assert(c, lprops->free + lprops->dirty >= c->min_idx_node_sz);
+	ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+	ubifs_assert(c, (lprops->flags & LPROPS_INDEX));
+
+	dbg_find("found dirty LEB %d, free %d, dirty %d, flags %#x",
+		 lprops->lnum, lprops->free, lprops->dirty, lprops->flags);
+
+	lprops = ubifs_change_lp(c, lprops, LPROPS_NC, LPROPS_NC,
+				 lprops->flags | LPROPS_TAKEN, 0);
+	if (IS_ERR(lprops))
+		return PTR_ERR(lprops);
+
+	return lprops->lnum;
+}
+
+/**
+ * get_idx_gc_leb - try to get a LEB number from trivial GC.
+ * @c: the UBIFS file-system description object
+ */
+static int get_idx_gc_leb(struct ubifs_info *c)
+{
+	const struct ubifs_lprops *lp;
+	int err, lnum;
+
+	err = ubifs_get_idx_gc_leb(c);
+	if (err < 0)
+		return err;
+	lnum = err;
+	/*
+	 * The LEB was due to be unmapped after the commit but
+	 * it is needed now for this commit.
+	 */
+	lp = ubifs_lpt_lookup_dirty(c, lnum);
+	if (IS_ERR(lp))
+		return PTR_ERR(lp);
+	lp = ubifs_change_lp(c, lp, LPROPS_NC, LPROPS_NC,
+			     lp->flags | LPROPS_INDEX, -1);
+	if (IS_ERR(lp))
+		return PTR_ERR(lp);
+	dbg_find("LEB %d, dirty %d and free %d flags %#x",
+		 lp->lnum, lp->dirty, lp->free, lp->flags);
+	return lnum;
+}
+
+/**
+ * find_dirtiest_idx_leb - find dirtiest index LEB from dirtiest array.
+ * @c: the UBIFS file-system description object
+ */
+static int find_dirtiest_idx_leb(struct ubifs_info *c)
+{
+	const struct ubifs_lprops *lp;
+	int lnum;
+
+	while (1) {
+		if (!c->dirty_idx.cnt)
+			return -ENOSPC;
+		/* The lprops pointers were replaced by LEB numbers */
+		lnum = (size_t)c->dirty_idx.arr[--c->dirty_idx.cnt];
+		lp = ubifs_lpt_lookup(c, lnum);
+		if (IS_ERR(lp))
+			return PTR_ERR(lp);
+		if ((lp->flags & LPROPS_TAKEN) || !(lp->flags & LPROPS_INDEX))
+			continue;
+		lp = ubifs_change_lp(c, lp, LPROPS_NC, LPROPS_NC,
+				     lp->flags | LPROPS_TAKEN, 0);
+		if (IS_ERR(lp))
+			return PTR_ERR(lp);
+		break;
+	}
+	dbg_find("LEB %d, dirty %d and free %d flags %#x", lp->lnum, lp->dirty,
+		 lp->free, lp->flags);
+	ubifs_assert(c, lp->flags & LPROPS_TAKEN);
+	ubifs_assert(c, lp->flags & LPROPS_INDEX);
+	return lnum;
+}
+
+/**
+ * ubifs_find_dirty_idx_leb - try to find dirtiest index LEB as at last commit.
+ * @c: the UBIFS file-system description object
+ *
+ * This function attempts to find an untaken index LEB with the most free and
+ * dirty space that can be used without overwriting index nodes that were in the
+ * last index committed.
+ */
+int ubifs_find_dirty_idx_leb(struct ubifs_info *c)
+{
+	int err;
+
+	ubifs_get_lprops(c);
+
+	/*
+	 * We made an array of the dirtiest index LEB numbers as at the start of
+	 * last commit.  Try that array first.
+	 */
+	err = find_dirtiest_idx_leb(c);
+
+	/* Next try scanning the entire LPT */
+	if (err == -ENOSPC)
+		err = find_dirty_idx_leb(c);
+
+	/* Finally take any index LEBs awaiting trivial GC */
+	if (err == -ENOSPC)
+		err = get_idx_gc_leb(c);
+
+	ubifs_release_lprops(c);
+	return err;
+}

linux-7.2.3/fs/ubifs/ioctl.c

@@ -0,0 +1,230 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ * Copyright (C) 2006, 2007 University of Szeged, Hungary
+ *
+ * Authors: Zoltan Sogor
+ *          Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+/* This file implements EXT2-compatible extended attribute ioctl() calls */
+
+#include <linux/compat.h>
+#include <linux/mount.h>
+#include <linux/fileattr.h>
+#include "ubifs.h"
+
+/* Need to be kept consistent with checked flags in ioctl2ubifs() */
+#define UBIFS_SETTABLE_IOCTL_FLAGS \
+	(FS_COMPR_FL | FS_SYNC_FL | FS_APPEND_FL | \
+	 FS_IMMUTABLE_FL | FS_DIRSYNC_FL)
+
+/* Need to be kept consistent with checked flags in ubifs2ioctl() */
+#define UBIFS_GETTABLE_IOCTL_FLAGS \
+	(UBIFS_SETTABLE_IOCTL_FLAGS | FS_ENCRYPT_FL)
+
+/**
+ * ubifs_set_inode_flags - set VFS inode flags.
+ * @inode: VFS inode to set flags for
+ *
+ * This function propagates flags from UBIFS inode object to VFS inode object.
+ */
+void ubifs_set_inode_flags(struct inode *inode)
+{
+	unsigned int flags = ubifs_inode(inode)->flags;
+
+	inode->i_flags &= ~(S_SYNC | S_APPEND | S_IMMUTABLE | S_DIRSYNC |
+			    S_ENCRYPTED);
+	if (flags & UBIFS_SYNC_FL)
+		inode->i_flags |= S_SYNC;
+	if (flags & UBIFS_APPEND_FL)
+		inode->i_flags |= S_APPEND;
+	if (flags & UBIFS_IMMUTABLE_FL)
+		inode->i_flags |= S_IMMUTABLE;
+	if (flags & UBIFS_DIRSYNC_FL)
+		inode->i_flags |= S_DIRSYNC;
+	if (flags & UBIFS_CRYPT_FL)
+		inode->i_flags |= S_ENCRYPTED;
+}
+
+/*
+ * ioctl2ubifs - convert ioctl inode flags to UBIFS inode flags.
+ * @ioctl_flags: flags to convert
+ *
+ * This function converts ioctl flags (@FS_COMPR_FL, etc) to UBIFS inode flags
+ * (@UBIFS_COMPR_FL, etc).
+ */
+static int ioctl2ubifs(int ioctl_flags)
+{
+	int ubifs_flags = 0;
+
+	if (ioctl_flags & FS_COMPR_FL)
+		ubifs_flags |= UBIFS_COMPR_FL;
+	if (ioctl_flags & FS_SYNC_FL)
+		ubifs_flags |= UBIFS_SYNC_FL;
+	if (ioctl_flags & FS_APPEND_FL)
+		ubifs_flags |= UBIFS_APPEND_FL;
+	if (ioctl_flags & FS_IMMUTABLE_FL)
+		ubifs_flags |= UBIFS_IMMUTABLE_FL;
+	if (ioctl_flags & FS_DIRSYNC_FL)
+		ubifs_flags |= UBIFS_DIRSYNC_FL;
+
+	return ubifs_flags;
+}
+
+/*
+ * ubifs2ioctl - convert UBIFS inode flags to ioctl inode flags.
+ * @ubifs_flags: flags to convert
+ *
+ * This function converts UBIFS inode flags (@UBIFS_COMPR_FL, etc) to ioctl
+ * flags (@FS_COMPR_FL, etc).
+ */
+static int ubifs2ioctl(int ubifs_flags)
+{
+	int ioctl_flags = 0;
+
+	if (ubifs_flags & UBIFS_COMPR_FL)
+		ioctl_flags |= FS_COMPR_FL;
+	if (ubifs_flags & UBIFS_SYNC_FL)
+		ioctl_flags |= FS_SYNC_FL;
+	if (ubifs_flags & UBIFS_APPEND_FL)
+		ioctl_flags |= FS_APPEND_FL;
+	if (ubifs_flags & UBIFS_IMMUTABLE_FL)
+		ioctl_flags |= FS_IMMUTABLE_FL;
+	if (ubifs_flags & UBIFS_DIRSYNC_FL)
+		ioctl_flags |= FS_DIRSYNC_FL;
+	if (ubifs_flags & UBIFS_CRYPT_FL)
+		ioctl_flags |= FS_ENCRYPT_FL;
+
+	return ioctl_flags;
+}
+
+static int setflags(struct inode *inode, int flags)
+{
+	int err, release;
+	struct ubifs_inode *ui = ubifs_inode(inode);
+	struct ubifs_info *c = inode->i_sb->s_fs_info;
+	struct ubifs_budget_req req = { .dirtied_ino = 1,
+			.dirtied_ino_d = ALIGN(ui->data_len, 8) };
+
+	err = ubifs_budget_space(c, &req);
+	if (err)
+		return err;
+
+	mutex_lock(&ui->ui_mutex);
+	ui->flags &= ~ioctl2ubifs(UBIFS_SETTABLE_IOCTL_FLAGS);
+	ui->flags |= ioctl2ubifs(flags);
+	ubifs_set_inode_flags(inode);
+	inode_set_ctime_current(inode);
+	release = ui->dirty;
+	mark_inode_dirty_sync(inode);
+	mutex_unlock(&ui->ui_mutex);
+
+	if (release)
+		ubifs_release_budget(c, &req);
+	if (IS_SYNC(inode))
+		err = write_inode_now(inode, 1);
+	return err;
+}
+
+int ubifs_fileattr_get(struct dentry *dentry, struct file_kattr *fa)
+{
+	struct inode *inode = d_inode(dentry);
+	int flags = ubifs2ioctl(ubifs_inode(inode)->flags);
+
+	if (d_is_special(dentry))
+		return -ENOTTY;
+
+	dbg_gen("get flags: %#x, i_flags %#x", flags, inode->i_flags);
+	fileattr_fill_flags(fa, flags);
+
+	return 0;
+}
+
+int ubifs_fileattr_set(struct mnt_idmap *idmap,
+		       struct dentry *dentry, struct file_kattr *fa)
+{
+	struct inode *inode = d_inode(dentry);
+	int flags = fa->flags;
+
+	if (d_is_special(dentry))
+		return -ENOTTY;
+
+	if (fileattr_has_fsx(fa))
+		return -EOPNOTSUPP;
+
+	if (flags & ~UBIFS_GETTABLE_IOCTL_FLAGS)
+		return -EOPNOTSUPP;
+
+	flags &= UBIFS_SETTABLE_IOCTL_FLAGS;
+
+	if (!S_ISDIR(inode->i_mode))
+		flags &= ~FS_DIRSYNC_FL;
+
+	dbg_gen("set flags: %#x, i_flags %#x", flags, inode->i_flags);
+	return setflags(inode, flags);
+}
+
+long ubifs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+	int err;
+	struct inode *inode = file_inode(file);
+
+	switch (cmd) {
+	case FS_IOC_SET_ENCRYPTION_POLICY: {
+		struct ubifs_info *c = inode->i_sb->s_fs_info;
+
+		err = ubifs_enable_encryption(c);
+		if (err)
+			return err;
+
+		return fscrypt_ioctl_set_policy(file, (const void __user *)arg);
+	}
+	case FS_IOC_GET_ENCRYPTION_POLICY:
+		return fscrypt_ioctl_get_policy(file, (void __user *)arg);
+
+	case FS_IOC_GET_ENCRYPTION_POLICY_EX:
+		return fscrypt_ioctl_get_policy_ex(file, (void __user *)arg);
+
+	case FS_IOC_ADD_ENCRYPTION_KEY:
+		return fscrypt_ioctl_add_key(file, (void __user *)arg);
+
+	case FS_IOC_REMOVE_ENCRYPTION_KEY:
+		return fscrypt_ioctl_remove_key(file, (void __user *)arg);
+
+	case FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS:
+		return fscrypt_ioctl_remove_key_all_users(file,
+							  (void __user *)arg);
+	case FS_IOC_GET_ENCRYPTION_KEY_STATUS:
+		return fscrypt_ioctl_get_key_status(file, (void __user *)arg);
+
+	case FS_IOC_GET_ENCRYPTION_NONCE:
+		return fscrypt_ioctl_get_nonce(file, (void __user *)arg);
+
+	default:
+		return -ENOTTY;
+	}
+}
+
+#ifdef CONFIG_COMPAT
+long ubifs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
+{
+	switch (cmd) {
+	case FS_IOC_SET_ENCRYPTION_POLICY:
+	case FS_IOC_GET_ENCRYPTION_POLICY:
+	case FS_IOC_GET_ENCRYPTION_POLICY_EX:
+	case FS_IOC_ADD_ENCRYPTION_KEY:
+	case FS_IOC_REMOVE_ENCRYPTION_KEY:
+	case FS_IOC_REMOVE_ENCRYPTION_KEY_ALL_USERS:
+	case FS_IOC_GET_ENCRYPTION_KEY_STATUS:
+	case FS_IOC_GET_ENCRYPTION_NONCE:
+		break;
+	default:
+		return -ENOIOCTLCMD;
+	}
+	return ubifs_ioctl(file, cmd, (unsigned long)compat_ptr(arg));
+}
+#endif

linux-7.2.3/fs/ubifs/key.h

@@ -0,0 +1,543 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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.
+ */
+
+/*
+ * Lot's of the key helpers require a struct ubifs_info *c as the first parameter.
+ * But we are not using it at all currently. That's designed for future extensions of
+ * different c->key_format. But right now, there is only one key type, UBIFS_SIMPLE_KEY_FMT.
+ */
+
+#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 (len--) {
+		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. Not a string when encrypted!
+ */
+static inline void dent_key_init(const struct ubifs_info *c,
+				 union ubifs_key *key, ino_t inum,
+				 const struct fscrypt_name *nm)
+{
+	uint32_t hash = c->key_hash(fname_name(nm), fname_len(nm));
+
+	ubifs_assert(c, !(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(c, !(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 fscrypt_name *nm)
+{
+	union ubifs_key *key = k;
+	uint32_t hash = c->key_hash(fname_name(nm), fname_len(nm));
+
+	ubifs_assert(c, !(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 fscrypt_name *nm)
+{
+	uint32_t hash = c->key_hash(fname_name(nm), fname_len(nm));
+
+	ubifs_assert(c, !(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 fscrypt_name *nm)
+{
+	union ubifs_key *key = k;
+	uint32_t hash = c->key_hash(fname_name(nm), fname_len(nm));
+
+	ubifs_assert(c, !(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(c, !(block & ~UBIFS_S_KEY_BLOCK_MASK));
+	key->u32[0] = inum;
+	key->u32[1] = block | (UBIFS_DATA_KEY << UBIFS_S_KEY_BLOCK_BITS);
+}
+
+/**
+ * highest_data_key - get the highest possible data key for an inode.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ * @inum: inode number
+ */
+static inline void highest_data_key(const struct ubifs_info *c,
+				   union ubifs_key *key, ino_t inum)
+{
+	data_key_init(c, key, inum, UBIFS_S_KEY_BLOCK_MASK);
+}
+
+/**
+ * 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;
+}
+
+/**
+ * invalid_key_init - initialize invalid node key.
+ * @c: UBIFS file-system description object
+ * @key: key to initialize
+ *
+ * This is a helper function which marks a @key object as invalid.
+ */
+static inline void invalid_key_init(const struct ubifs_info *c,
+				    union ubifs_key *key)
+{
+	key->u32[0] = 0xDEADBEAF;
+	key->u32[1] = UBIFS_INVALID_KEY;
+}
+
+/**
+ * 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 uint32_t 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 uint32_t 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__ */

linux-7.2.3/fs/ubifs/log.c

@@ -0,0 +1,762 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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"
+
+static int dbg_check_bud_bytes(struct ubifs_info *c);
+
+/**
+ * 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_get_wbuf - get the wbuf associated with a LEB, if there is one.
+ * @c: UBIFS file-system description object
+ * @lnum: logical eraseblock number to search
+ *
+ * This functions returns the wbuf for @lnum or %NULL if there is not one.
+ */
+struct ubifs_wbuf *ubifs_get_wbuf(struct ubifs_info *c, int lnum)
+{
+	struct rb_node *p;
+	struct ubifs_bud *bud;
+	int jhead;
+
+	if (!c->jheads)
+		return NULL;
+
+	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 {
+			jhead = bud->jhead;
+			spin_unlock(&c->buds_lock);
+			return &c->jheads[jhead].wbuf;
+		}
+	}
+	spin_unlock(&c->buds_lock);
+	return NULL;
+}
+
+/**
+ * empty_log_bytes - calculate amount of empty space in the log.
+ * @c: UBIFS file-system description object
+ */
+static inline long long empty_log_bytes(const struct ubifs_info *c)
+{
+	long long h, t;
+
+	h = (long long)c->lhead_lnum * c->leb_size + c->lhead_offs;
+	t = (long long)c->ltail_lnum * c->leb_size;
+
+	if (h > t)
+		return c->log_bytes - h + t;
+	else if (h != t)
+		return t - h;
+	else if (c->lhead_lnum != c->ltail_lnum)
+		return 0;
+	else
+		return c->log_bytes;
+}
+
+/**
+ * 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(c, 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, c->replaying && c->ro_mount);
+
+	/*
+	 * 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 %s, bud_bytes %lld", bud->lnum,
+		bud->start, dbg_jhead(bud->jhead), c->bud_bytes);
+	spin_unlock(&c->buds_lock);
+}
+
+/**
+ * ubifs_add_bud_to_log - add a new bud to the log.
+ * @c: UBIFS file-system description object
+ * @jhead: journal head the bud belongs to
+ * @lnum: LEB number of the bud
+ * @offs: starting offset of the bud
+ *
+ * This function writes a reference node for the new bud LEB @lnum to the log,
+ * and adds it to the buds trees. It also makes sure that log size does not
+ * exceed the 'c->max_bud_bytes' limit. Returns zero in case of success,
+ * %-EAGAIN if commit is required, and a negative error code in case of
+ * failure.
+ */
+int ubifs_add_bud_to_log(struct ubifs_info *c, int jhead, int lnum, int offs)
+{
+	int err;
+	struct ubifs_bud *bud;
+	struct ubifs_ref_node *ref;
+
+	bud = kmalloc_obj(struct ubifs_bud, GFP_NOFS);
+	if (!bud)
+		return -ENOMEM;
+	ref = kzalloc(c->ref_node_alsz, GFP_NOFS);
+	if (!ref) {
+		kfree(bud);
+		return -ENOMEM;
+	}
+
+	mutex_lock(&c->log_mutex);
+	ubifs_assert(c, !c->ro_media && !c->ro_mount);
+	if (c->ro_error) {
+		err = -EROFS;
+		goto out_unlock;
+	}
+
+	/* Make sure we have enough space in the log */
+	if (empty_log_bytes(c) - c->ref_node_alsz < c->min_log_bytes) {
+		dbg_log("not enough log space - %lld, required %d",
+			empty_log_bytes(c), c->min_log_bytes);
+		ubifs_commit_required(c);
+		err = -EAGAIN;
+		goto out_unlock;
+	}
+
+	/*
+	 * Make sure the amount of space in buds will not exceed the
+	 * 'c->max_bud_bytes' limit, because we want to guarantee mount time
+	 * limits.
+	 *
+	 * It is not necessary to hold @c->buds_lock when reading @c->bud_bytes
+	 * because we are holding @c->log_mutex. All @c->bud_bytes take place
+	 * when both @c->log_mutex and @c->bud_bytes are locked.
+	 */
+	if (c->bud_bytes + c->leb_size - offs > c->max_bud_bytes) {
+		dbg_log("bud bytes %lld (%lld max), require commit",
+			c->bud_bytes, c->max_bud_bytes);
+		ubifs_commit_required(c);
+		err = -EAGAIN;
+		goto out_unlock;
+	}
+
+	/*
+	 * If the journal is full enough - start background commit. Note, it is
+	 * OK to read 'c->cmt_state' without spinlock because integer reads
+	 * are atomic in the kernel.
+	 */
+	if (c->bud_bytes >= c->bg_bud_bytes &&
+	    c->cmt_state == COMMIT_RESTING) {
+		dbg_log("bud bytes %lld (%lld max), initiate BG commit",
+			c->bud_bytes, c->max_bud_bytes);
+		ubifs_request_bg_commit(c);
+	}
+
+	bud->lnum = lnum;
+	bud->start = offs;
+	bud->jhead = jhead;
+	bud->log_hash = NULL;
+
+	ref->ch.node_type = UBIFS_REF_NODE;
+	ref->lnum = cpu_to_le32(bud->lnum);
+	ref->offs = cpu_to_le32(bud->start);
+	ref->jhead = cpu_to_le32(jhead);
+
+	if (c->lhead_offs > c->leb_size - c->ref_node_alsz) {
+		c->lhead_lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
+		ubifs_assert(c, c->lhead_lnum != c->ltail_lnum);
+		c->lhead_offs = 0;
+	}
+
+	if (c->lhead_offs == 0) {
+		/* Must ensure next log LEB has been unmapped */
+		err = ubifs_leb_unmap(c, c->lhead_lnum);
+		if (err)
+			goto out_unlock;
+	}
+
+	if (bud->start == 0) {
+		/*
+		 * Before writing the LEB reference which refers an empty LEB
+		 * to the log, we have to make sure it is mapped, because
+		 * otherwise we'd risk to refer an LEB with garbage in case of
+		 * an unclean reboot, because the target LEB might have been
+		 * unmapped, but not yet physically erased.
+		 */
+		err = ubifs_leb_map(c, bud->lnum);
+		if (err)
+			goto out_unlock;
+	}
+
+	dbg_log("write ref LEB %d:%d",
+		c->lhead_lnum, c->lhead_offs);
+	err = ubifs_write_node(c, ref, UBIFS_REF_NODE_SZ, c->lhead_lnum,
+			       c->lhead_offs);
+	if (err)
+		goto out_unlock;
+
+	err = ubifs_shash_update(c, c->log_hash, ref, UBIFS_REF_NODE_SZ);
+	if (err)
+		goto out_unlock;
+
+	err = ubifs_shash_copy_state(c, c->log_hash, c->jheads[jhead].log_hash);
+	if (err)
+		goto out_unlock;
+
+	c->lhead_offs += c->ref_node_alsz;
+
+	ubifs_add_bud(c, bud);
+
+	mutex_unlock(&c->log_mutex);
+	kfree(ref);
+	return 0;
+
+out_unlock:
+	mutex_unlock(&c->log_mutex);
+	kfree(ref);
+	kfree(bud);
+	return err;
+}
+
+/**
+ * remove_buds - remove used buds.
+ * @c: UBIFS file-system description object
+ *
+ * This function removes use buds from the buds tree. It does not remove the
+ * buds which are pointed to by journal heads.
+ */
+static void remove_buds(struct ubifs_info *c)
+{
+	struct rb_node *p;
+
+	ubifs_assert(c, list_empty(&c->old_buds));
+	c->cmt_bud_bytes = 0;
+	spin_lock(&c->buds_lock);
+	p = rb_first(&c->buds);
+	while (p) {
+		struct rb_node *p1 = p;
+		struct ubifs_bud *bud;
+		struct ubifs_wbuf *wbuf;
+
+		p = rb_next(p);
+		bud = rb_entry(p1, struct ubifs_bud, rb);
+		wbuf = &c->jheads[bud->jhead].wbuf;
+
+		if (wbuf->lnum == bud->lnum) {
+			/*
+			 * Do not remove buds which are pointed to by journal
+			 * heads (non-closed buds).
+			 */
+			c->cmt_bud_bytes += wbuf->offs - bud->start;
+			dbg_log("preserve %d:%d, jhead %s, bud bytes %d, cmt_bud_bytes %lld",
+				bud->lnum, bud->start, dbg_jhead(bud->jhead),
+				wbuf->offs - bud->start, c->cmt_bud_bytes);
+			bud->start = wbuf->offs;
+		} else {
+			c->cmt_bud_bytes += c->leb_size - bud->start;
+			dbg_log("remove %d:%d, jhead %s, bud bytes %d, cmt_bud_bytes %lld",
+				bud->lnum, bud->start, dbg_jhead(bud->jhead),
+				c->leb_size - bud->start, c->cmt_bud_bytes);
+			rb_erase(p1, &c->buds);
+			/*
+			 * If the commit does not finish, the recovery will need
+			 * to replay the journal, in which case the old buds
+			 * must be unchanged. Do not release them until post
+			 * commit i.e. do not allow them to be garbage
+			 * collected.
+			 */
+			list_move(&bud->list, &c->old_buds);
+		}
+	}
+	spin_unlock(&c->buds_lock);
+}
+
+/**
+ * ubifs_log_start_commit - start commit.
+ * @c: UBIFS file-system description object
+ * @ltail_lnum: return new log tail LEB number
+ *
+ * The commit operation starts with writing "commit start" node to the log and
+ * reference nodes for all journal heads which will define new journal after
+ * the commit has been finished. The commit start and reference nodes are
+ * written in one go to the nearest empty log LEB (hence, when commit is
+ * finished UBIFS may safely unmap all the previous log LEBs). This function
+ * returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_log_start_commit(struct ubifs_info *c, int *ltail_lnum)
+{
+	void *buf;
+	struct ubifs_cs_node *cs;
+	struct ubifs_ref_node *ref;
+	int err, i, max_len, len;
+
+	err = dbg_check_bud_bytes(c);
+	if (err)
+		return err;
+
+	max_len = UBIFS_CS_NODE_SZ + c->jhead_cnt * UBIFS_REF_NODE_SZ;
+	max_len = ALIGN(max_len, c->min_io_size);
+	buf = cs = kmalloc(max_len, GFP_NOFS);
+	if (!buf)
+		return -ENOMEM;
+
+	cs->ch.node_type = UBIFS_CS_NODE;
+	cs->cmt_no = cpu_to_le64(c->cmt_no);
+	ubifs_prepare_node(c, cs, UBIFS_CS_NODE_SZ, 0);
+
+	err = ubifs_shash_init(c, c->log_hash);
+	if (err)
+		goto out;
+
+	err = ubifs_shash_update(c, c->log_hash, cs, UBIFS_CS_NODE_SZ);
+	if (err < 0)
+		goto out;
+
+	/*
+	 * Note, we do not lock 'c->log_mutex' because this is the commit start
+	 * phase and we are exclusively using the log. And we do not lock
+	 * write-buffer because nobody can write to the file-system at this
+	 * phase.
+	 */
+
+	len = UBIFS_CS_NODE_SZ;
+	for (i = 0; i < c->jhead_cnt; i++) {
+		int lnum = c->jheads[i].wbuf.lnum;
+		int offs = c->jheads[i].wbuf.offs;
+
+		if (lnum == -1 || offs == c->leb_size)
+			continue;
+
+		dbg_log("add ref to LEB %d:%d for jhead %s",
+			lnum, offs, dbg_jhead(i));
+		ref = buf + len;
+		ref->ch.node_type = UBIFS_REF_NODE;
+		ref->lnum = cpu_to_le32(lnum);
+		ref->offs = cpu_to_le32(offs);
+		ref->jhead = cpu_to_le32(i);
+
+		ubifs_prepare_node(c, ref, UBIFS_REF_NODE_SZ, 0);
+		len += UBIFS_REF_NODE_SZ;
+
+		err = ubifs_shash_update(c, c->log_hash, ref,
+					 UBIFS_REF_NODE_SZ);
+		if (err)
+			goto out;
+		ubifs_shash_copy_state(c, c->log_hash, c->jheads[i].log_hash);
+	}
+
+	ubifs_pad(c, buf + len, ALIGN(len, c->min_io_size) - len);
+
+	/* Switch to the next log LEB */
+	if (c->lhead_offs) {
+		c->lhead_lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
+		ubifs_assert(c, c->lhead_lnum != c->ltail_lnum);
+		c->lhead_offs = 0;
+	}
+
+	/* Must ensure next LEB has been unmapped */
+	err = ubifs_leb_unmap(c, c->lhead_lnum);
+	if (err)
+		goto out;
+
+	len = ALIGN(len, c->min_io_size);
+	dbg_log("writing commit start at LEB %d:0, len %d", c->lhead_lnum, len);
+	err = ubifs_leb_write(c, c->lhead_lnum, cs, 0, len);
+	if (err)
+		goto out;
+
+	*ltail_lnum = c->lhead_lnum;
+
+	c->lhead_offs += len;
+	ubifs_assert(c, c->lhead_offs < c->leb_size);
+
+	remove_buds(c);
+
+	/*
+	 * We have started the commit and now users may use the rest of the log
+	 * for new writes.
+	 */
+	c->min_log_bytes = 0;
+
+out:
+	kfree(buf);
+	return err;
+}
+
+/**
+ * ubifs_log_end_commit - end commit.
+ * @c: UBIFS file-system description object
+ * @ltail_lnum: new log tail LEB number
+ *
+ * This function is called on when the commit operation was finished. It
+ * moves log tail to new position and updates the master node so that it stores
+ * the new log tail LEB number. Returns zero in case of success and a negative
+ * error code in case of failure.
+ */
+int ubifs_log_end_commit(struct ubifs_info *c, int ltail_lnum)
+{
+	int err;
+
+	/*
+	 * At this phase we have to lock 'c->log_mutex' because UBIFS allows FS
+	 * writes during commit. Its only short "commit" start phase when
+	 * writers are blocked.
+	 */
+	mutex_lock(&c->log_mutex);
+
+	dbg_log("old tail was LEB %d:0, new tail is LEB %d:0",
+		c->ltail_lnum, ltail_lnum);
+
+	c->ltail_lnum = ltail_lnum;
+	/*
+	 * The commit is finished and from now on it must be guaranteed that
+	 * there is always enough space for the next commit.
+	 */
+	c->min_log_bytes = c->leb_size;
+
+	spin_lock(&c->buds_lock);
+	c->bud_bytes -= c->cmt_bud_bytes;
+	spin_unlock(&c->buds_lock);
+
+	err = dbg_check_bud_bytes(c);
+	if (err)
+		goto out;
+
+	err = ubifs_write_master(c);
+
+out:
+	mutex_unlock(&c->log_mutex);
+	return err;
+}
+
+/**
+ * ubifs_log_post_commit - things to do after commit is completed.
+ * @c: UBIFS file-system description object
+ * @old_ltail_lnum: old log tail LEB number
+ *
+ * Release buds only after commit is completed, because they must be unchanged
+ * if recovery is needed.
+ *
+ * Unmap log LEBs only after commit is completed, because they may be needed for
+ * recovery.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_log_post_commit(struct ubifs_info *c, int old_ltail_lnum)
+{
+	int lnum, err = 0;
+
+	while (!list_empty(&c->old_buds)) {
+		struct ubifs_bud *bud;
+
+		bud = list_entry(c->old_buds.next, struct ubifs_bud, list);
+		err = ubifs_return_leb(c, bud->lnum);
+		if (err)
+			return err;
+		list_del(&bud->list);
+		kfree(bud->log_hash);
+		kfree(bud);
+	}
+	mutex_lock(&c->log_mutex);
+	for (lnum = old_ltail_lnum; lnum != c->ltail_lnum;
+	     lnum = ubifs_next_log_lnum(c, lnum)) {
+		dbg_log("unmap log LEB %d", lnum);
+		err = ubifs_leb_unmap(c, lnum);
+		if (err)
+			goto out;
+	}
+out:
+	mutex_unlock(&c->log_mutex);
+	return err;
+}
+
+/**
+ * struct done_ref - references that have been done.
+ * @rb: rb-tree node
+ * @lnum: LEB number
+ */
+struct done_ref {
+	struct rb_node rb;
+	int lnum;
+};
+
+/**
+ * done_already - determine if a reference has been done already.
+ * @done_tree: rb-tree to store references that have been done
+ * @lnum: LEB number of reference
+ *
+ * This function returns %1 if the reference has been done, %0 if not, otherwise
+ * a negative error code is returned.
+ */
+static int done_already(struct rb_root *done_tree, int lnum)
+{
+	struct rb_node **p = &done_tree->rb_node, *parent = NULL;
+	struct done_ref *dr;
+
+	while (*p) {
+		parent = *p;
+		dr = rb_entry(parent, struct done_ref, rb);
+		if (lnum < dr->lnum)
+			p = &(*p)->rb_left;
+		else if (lnum > dr->lnum)
+			p = &(*p)->rb_right;
+		else
+			return 1;
+	}
+
+	dr = kzalloc_obj(struct done_ref, GFP_NOFS);
+	if (!dr)
+		return -ENOMEM;
+
+	dr->lnum = lnum;
+
+	rb_link_node(&dr->rb, parent, p);
+	rb_insert_color(&dr->rb, done_tree);
+
+	return 0;
+}
+
+/**
+ * destroy_done_tree - destroy the done tree.
+ * @done_tree: done tree to destroy
+ */
+static void destroy_done_tree(struct rb_root *done_tree)
+{
+	struct done_ref *dr, *n;
+
+	rbtree_postorder_for_each_entry_safe(dr, n, done_tree, rb)
+		kfree(dr);
+}
+
+/**
+ * add_node - add a node to the consolidated log.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to which to add
+ * @lnum: LEB number to which to write is passed and returned here
+ * @offs: offset to where to write is passed and returned here
+ * @node: node to add
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int add_node(struct ubifs_info *c, void *buf, int *lnum, int *offs,
+		    void *node)
+{
+	struct ubifs_ch *ch = node;
+	int len = le32_to_cpu(ch->len), remains = c->leb_size - *offs;
+
+	if (len > remains) {
+		int sz = ALIGN(*offs, c->min_io_size), err;
+
+		ubifs_pad(c, buf + *offs, sz - *offs);
+		err = ubifs_leb_change(c, *lnum, buf, sz);
+		if (err)
+			return err;
+		*lnum = ubifs_next_log_lnum(c, *lnum);
+		*offs = 0;
+	}
+	memcpy(buf + *offs, node, len);
+	*offs += ALIGN(len, 8);
+	return 0;
+}
+
+/**
+ * ubifs_consolidate_log - consolidate the log.
+ * @c: UBIFS file-system description object
+ *
+ * Repeated failed commits could cause the log to be full, but at least 1 LEB is
+ * needed for commit. This function rewrites the reference nodes in the log
+ * omitting duplicates, and failed CS nodes, and leaving no gaps.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_consolidate_log(struct ubifs_info *c)
+{
+	struct ubifs_scan_leb *sleb;
+	struct ubifs_scan_node *snod;
+	struct rb_root done_tree = RB_ROOT;
+	int lnum, err, first = 1, write_lnum, offs = 0;
+	void *buf;
+
+	dbg_rcvry("log tail LEB %d, log head LEB %d", c->ltail_lnum,
+		  c->lhead_lnum);
+	buf = vmalloc(c->leb_size);
+	if (!buf)
+		return -ENOMEM;
+	lnum = c->ltail_lnum;
+	write_lnum = lnum;
+	while (1) {
+		sleb = ubifs_scan(c, lnum, 0, c->sbuf, 0);
+		if (IS_ERR(sleb)) {
+			err = PTR_ERR(sleb);
+			goto out_free;
+		}
+		list_for_each_entry(snod, &sleb->nodes, list) {
+			switch (snod->type) {
+			case UBIFS_REF_NODE: {
+				struct ubifs_ref_node *ref = snod->node;
+				int ref_lnum = le32_to_cpu(ref->lnum);
+
+				err = done_already(&done_tree, ref_lnum);
+				if (err < 0)
+					goto out_scan;
+				if (err != 1) {
+					err = add_node(c, buf, &write_lnum,
+						       &offs, snod->node);
+					if (err)
+						goto out_scan;
+				}
+				break;
+			}
+			case UBIFS_CS_NODE:
+				if (!first)
+					break;
+				err = add_node(c, buf, &write_lnum, &offs,
+					       snod->node);
+				if (err)
+					goto out_scan;
+				first = 0;
+				break;
+			}
+		}
+		ubifs_scan_destroy(sleb);
+		if (lnum == c->lhead_lnum)
+			break;
+		lnum = ubifs_next_log_lnum(c, lnum);
+	}
+	if (offs) {
+		int sz = ALIGN(offs, c->min_io_size);
+
+		ubifs_pad(c, buf + offs, sz - offs);
+		err = ubifs_leb_change(c, write_lnum, buf, sz);
+		if (err)
+			goto out_free;
+		offs = ALIGN(offs, c->min_io_size);
+	}
+	destroy_done_tree(&done_tree);
+	vfree(buf);
+	if (write_lnum == c->lhead_lnum) {
+		ubifs_err(c, "log is too full");
+		return -EINVAL;
+	}
+	/* Unmap remaining LEBs */
+	lnum = write_lnum;
+	do {
+		lnum = ubifs_next_log_lnum(c, lnum);
+		err = ubifs_leb_unmap(c, lnum);
+		if (err)
+			return err;
+	} while (lnum != c->lhead_lnum);
+	c->lhead_lnum = write_lnum;
+	c->lhead_offs = offs;
+	dbg_rcvry("new log head at %d:%d", c->lhead_lnum, c->lhead_offs);
+	return 0;
+
+out_scan:
+	ubifs_scan_destroy(sleb);
+out_free:
+	destroy_done_tree(&done_tree);
+	vfree(buf);
+	return err;
+}
+
+/**
+ * dbg_check_bud_bytes - make sure bud bytes calculation are all right.
+ * @c: UBIFS file-system description object
+ *
+ * This function makes sure the amount of flash space used by closed buds
+ * ('c->bud_bytes' is correct). Returns zero in case of success and %-EINVAL in
+ * case of failure.
+ */
+static int dbg_check_bud_bytes(struct ubifs_info *c)
+{
+	int i, err = 0;
+	struct ubifs_bud *bud;
+	long long bud_bytes = 0;
+
+	if (!dbg_is_chk_gen(c))
+		return 0;
+
+	spin_lock(&c->buds_lock);
+	for (i = 0; i < c->jhead_cnt; i++)
+		list_for_each_entry(bud, &c->jheads[i].buds_list, list)
+			bud_bytes += c->leb_size - bud->start;
+
+	if (c->bud_bytes != bud_bytes) {
+		ubifs_err(c, "bad bud_bytes %lld, calculated %lld",
+			  c->bud_bytes, bud_bytes);
+		err = -EINVAL;
+	}
+	spin_unlock(&c->buds_lock);
+
+	return err;
+}

linux-7.2.3/fs/ubifs/master.c

@@ -0,0 +1,476 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+/* This file implements reading and writing the master node */
+
+#include "ubifs.h"
+
+/**
+ * ubifs_compare_master_node - compare two UBIFS master nodes
+ * @c: UBIFS file-system description object
+ * @m1: the first node
+ * @m2: the second node
+ *
+ * This function compares two UBIFS master nodes. Returns 0 if they are equal
+ * and nonzero if not.
+ */
+int ubifs_compare_master_node(struct ubifs_info *c, void *m1, void *m2)
+{
+	int ret;
+	int behind;
+	int hmac_offs = offsetof(struct ubifs_mst_node, hmac);
+
+	/*
+	 * Do not compare the common node header since the sequence number and
+	 * hence the CRC are different.
+	 */
+	ret = memcmp(m1 + UBIFS_CH_SZ, m2 + UBIFS_CH_SZ,
+		     hmac_offs - UBIFS_CH_SZ);
+	if (ret)
+		return ret;
+
+	/*
+	 * Do not compare the embedded HMAC as well which also must be different
+	 * due to the different common node header.
+	 */
+	behind = hmac_offs + UBIFS_MAX_HMAC_LEN;
+
+	if (UBIFS_MST_NODE_SZ > behind)
+		return memcmp(m1 + behind, m2 + behind, UBIFS_MST_NODE_SZ - behind);
+
+	return 0;
+}
+
+/* mst_node_check_hash - Check hash of a master node
+ * @c: UBIFS file-system description object
+ * @mst: The master node
+ * @expected: The expected hash of the master node
+ *
+ * This checks the hash of a master node against a given expected hash.
+ * Note that we have two master nodes on a UBIFS image which have different
+ * sequence numbers and consequently different CRCs. To be able to match
+ * both master nodes we exclude the common node header containing the sequence
+ * number and CRC from the hash.
+ *
+ * Returns 0 if the hashes are equal, a negative error code otherwise.
+ */
+static int mst_node_check_hash(const struct ubifs_info *c,
+			       const struct ubifs_mst_node *mst,
+			       const u8 *expected)
+{
+	u8 calc[UBIFS_MAX_HASH_LEN];
+	const void *node = mst;
+	int ret;
+
+	ret = crypto_shash_tfm_digest(c->hash_tfm, node + sizeof(struct ubifs_ch),
+				UBIFS_MST_NODE_SZ - sizeof(struct ubifs_ch),
+				calc);
+	if (ret)
+		return ret;
+
+	if (ubifs_check_hash(c, expected, calc))
+		return -EPERM;
+
+	return 0;
+}
+
+/**
+ * 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, %-EUCLEAN if there master area is
+ * corrupted and requires recovery, 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, err;
+
+	lnum = UBIFS_MST_LNUM;
+
+	sleb = ubifs_scan(c, lnum, 0, c->sbuf, 1);
+	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_dump;
+		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, 1);
+	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_dump;
+	if (snod->offs != offs)
+		goto out;
+	if (ubifs_compare_master_node(c, c->mst_node, snod->node))
+		goto out;
+
+	c->mst_offs = offs;
+	ubifs_scan_destroy(sleb);
+
+	if (!ubifs_authenticated(c))
+		return 0;
+
+	if (ubifs_hmac_zero(c, c->mst_node->hmac)) {
+		err = mst_node_check_hash(c, c->mst_node,
+					  c->sup_node->hash_mst);
+		if (err)
+			ubifs_err(c, "Failed to verify master node hash");
+	} else {
+		err = ubifs_node_verify_hmac(c, c->mst_node,
+					sizeof(struct ubifs_mst_node),
+					offsetof(struct ubifs_mst_node, hmac));
+		if (err)
+			ubifs_err(c, "Failed to verify master node HMAC");
+	}
+
+	if (err)
+		return -EPERM;
+
+	return 0;
+
+out:
+	ubifs_scan_destroy(sleb);
+	return -EUCLEAN;
+
+out_dump:
+	ubifs_err(c, "unexpected node type %d master LEB %d:%d",
+		  snod->type, lnum, snod->offs);
+	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->bi.old_idx_sz & 7 || c->bi.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->bi.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(c, "bad master node at offset %d error %d", c->mst_offs, err);
+	ubifs_dump_node(c, c->mst_node, c->mst_node_alsz);
+	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) {
+		if (err == -EUCLEAN)
+			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->bi.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);
+
+	ubifs_copy_hash(c, c->mst_node->hash_root_idx, c->zroot.hash);
+
+	c->calc_idx_sz = c->bi.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(c, "bad leb_cnt on master node");
+			ubifs_dump_node(c, c->mst_node, c->mst_node_alsz);
+			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;
+}
+
+/**
+ * ubifs_write_master - write master node.
+ * @c: UBIFS file-system description object
+ *
+ * This function writes the master node. Returns zero in case of success and a
+ * negative error code in case of failure. The master node is written twice to
+ * enable recovery.
+ */
+int ubifs_write_master(struct ubifs_info *c)
+{
+	int err, lnum, offs, len;
+
+	ubifs_assert(c, !c->ro_media && !c->ro_mount);
+	if (c->ro_error)
+		return -EROFS;
+
+	lnum = UBIFS_MST_LNUM;
+	offs = c->mst_offs + c->mst_node_alsz;
+	len = UBIFS_MST_NODE_SZ;
+
+	if (offs + UBIFS_MST_NODE_SZ > c->leb_size) {
+		err = ubifs_leb_unmap(c, lnum);
+		if (err)
+			return err;
+		offs = 0;
+	}
+
+	c->mst_offs = offs;
+	c->mst_node->highest_inum = cpu_to_le64(c->highest_inum);
+
+	ubifs_copy_hash(c, c->zroot.hash, c->mst_node->hash_root_idx);
+	err = ubifs_write_node_hmac(c, c->mst_node, len, lnum, offs,
+				    offsetof(struct ubifs_mst_node, hmac));
+	if (err)
+		return err;
+
+	lnum += 1;
+
+	if (offs == 0) {
+		err = ubifs_leb_unmap(c, lnum);
+		if (err)
+			return err;
+	}
+	err = ubifs_write_node_hmac(c, c->mst_node, len, lnum, offs,
+				    offsetof(struct ubifs_mst_node, hmac));
+
+	return err;
+}

linux-7.2.3/fs/ubifs/misc.c

@@ -0,0 +1,69 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/kernel.h>
+#include "ubifs.h"
+
+/* Normal UBIFS messages */
+void ubifs_msg(const struct ubifs_info *c, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	pr_notice("UBIFS (ubi%d:%d): %pV\n",
+		  c->vi.ubi_num, c->vi.vol_id, &vaf);
+
+	va_end(args);
+}								    \
+
+/* UBIFS error messages */
+void ubifs_err(const struct ubifs_info *c, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	pr_err("UBIFS error (ubi%d:%d pid %d): %ps: %pV\n",
+	       c->vi.ubi_num, c->vi.vol_id, current->pid,
+	       __builtin_return_address(0),
+	       &vaf);
+
+	va_end(args);
+}								    \
+
+/* UBIFS warning messages */
+void ubifs_warn(const struct ubifs_info *c, const char *fmt, ...)
+{
+	struct va_format vaf;
+	va_list args;
+
+	va_start(args, fmt);
+
+	vaf.fmt = fmt;
+	vaf.va = &args;
+
+	pr_warn("UBIFS warning (ubi%d:%d pid %d): %ps: %pV\n",
+		c->vi.ubi_num, c->vi.vol_id, current->pid,
+		__builtin_return_address(0),
+		&vaf);
+
+	va_end(args);
+}
+
+static char *assert_names[] = {
+	[ASSACT_REPORT] = "report",
+	[ASSACT_RO] = "read-only",
+	[ASSACT_PANIC] = "panic",
+};
+
+const char *ubifs_assert_action_name(struct ubifs_info *c)
+{
+	return assert_names[c->assert_action];
+}

linux-7.2.3/fs/ubifs/misc.h

@@ -0,0 +1,289 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation
+ *
+ * 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_zn_obsolete - check if znode is obsolete.
+ * @znode: znode to check
+ *
+ * This helper function returns %1 if @znode is obsolete and %0 otherwise.
+ */
+static inline int ubifs_zn_obsolete(const struct ubifs_znode *znode)
+{
+	return !!test_bit(OBSOLETE_ZNODE, &znode->flags);
+}
+
+/**
+ * ubifs_zn_cow - check if znode has to be copied on write.
+ * @znode: znode to check
+ *
+ * This helper function returns %1 if @znode is has COW flag set and %0
+ * otherwise.
+ */
+static inline int ubifs_zn_cow(const struct ubifs_znode *znode)
+{
+	return !!test_bit(COW_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
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns %1 of compressor of type @compr_type is present, and
+ * %0 if not.
+ */
+static inline int ubifs_compr_present(struct ubifs_info *c, int compr_type)
+{
+	ubifs_assert(c, 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
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns compressor type string.
+ */
+static inline const char *ubifs_compr_name(struct ubifs_info *c, int compr_type)
+{
+	ubifs_assert(c, 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 '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_encode_dev - encode device node IDs.
+ * @dev: UBIFS device node information
+ * @rdev: device IDs to encode
+ *
+ * This is a helper function which encodes major/minor numbers of a device node
+ * into UBIFS device node description. We use standard Linux "new" and "huge"
+ * encodings.
+ */
+static inline int ubifs_encode_dev(union ubifs_dev_desc *dev, dev_t rdev)
+{
+	dev->new = cpu_to_le32(new_encode_dev(rdev));
+	return sizeof(dev->new);
+}
+
+/**
+ * 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 + c->hash_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 + c->hash_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(c, mutex_is_locked(&c->lp_mutex));
+	ubifs_assert(c, c->lst.empty_lebs >= 0 &&
+		     c->lst.empty_lebs <= c->main_lebs);
+	mutex_unlock(&c->lp_mutex);
+}
+
+/**
+ * ubifs_next_log_lnum - switch to the next log LEB.
+ * @c: UBIFS file-system description object
+ * @lnum: current log LEB
+ *
+ * This helper function returns the log LEB number which goes next after LEB
+ * 'lnum'.
+ */
+static inline int ubifs_next_log_lnum(const struct ubifs_info *c, int lnum)
+{
+	lnum += 1;
+	if (lnum > c->log_last)
+		lnum = UBIFS_LOG_LNUM;
+
+	return lnum;
+}
+
+static inline int ubifs_xattr_max_cnt(struct ubifs_info *c)
+{
+	int max_xattrs = (c->leb_size / 2) / UBIFS_INO_NODE_SZ;
+
+	ubifs_assert(c, max_xattrs < c->max_orphans);
+	return max_xattrs;
+}
+
+const char *ubifs_assert_action_name(struct ubifs_info *c);
+
+#endif /* __UBIFS_MISC_H__ */

linux-7.2.3/fs/ubifs/orphan.c

@@ -0,0 +1,947 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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.
+ */
+
+static int dbg_check_orphans(struct ubifs_info *c);
+
+/**
+ * ubifs_add_orphan - add an orphan.
+ * @c: UBIFS file-system description object
+ * @inum: orphan inode number
+ *
+ * Add an orphan. This function is called when an inodes link count drops to
+ * zero.
+ */
+int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
+{
+	struct ubifs_orphan *orphan, *o;
+	struct rb_node **p, *parent = NULL;
+
+	orphan = kzalloc_obj(struct ubifs_orphan, GFP_NOFS);
+	if (!orphan)
+		return -ENOMEM;
+	orphan->inum = inum;
+	orphan->new = 1;
+
+	spin_lock(&c->orphan_lock);
+	if (c->tot_orphans >= c->max_orphans) {
+		spin_unlock(&c->orphan_lock);
+		kfree(orphan);
+		return -ENFILE;
+	}
+	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 {
+			ubifs_err(c, "ino %lu orphaned twice", (unsigned long)inum);
+			spin_unlock(&c->orphan_lock);
+			kfree(orphan);
+			return -EINVAL;
+		}
+	}
+	c->tot_orphans += 1;
+	c->new_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);
+	list_add_tail(&orphan->new_list, &c->orph_new);
+
+	spin_unlock(&c->orphan_lock);
+	dbg_gen("ino %lu", (unsigned long)inum);
+	return 0;
+}
+
+static struct ubifs_orphan *lookup_orphan(struct ubifs_info *c, ino_t inum)
+{
+	struct ubifs_orphan *o;
+	struct rb_node *p;
+
+	p = c->orph_tree.rb_node;
+	while (p) {
+		o = rb_entry(p, struct ubifs_orphan, rb);
+		if (inum < o->inum)
+			p = p->rb_left;
+		else if (inum > o->inum)
+			p = p->rb_right;
+		else {
+			return o;
+		}
+	}
+	return NULL;
+}
+
+static void __orphan_drop(struct ubifs_info *c, struct ubifs_orphan *o)
+{
+	rb_erase(&o->rb, &c->orph_tree);
+	list_del(&o->list);
+	c->tot_orphans -= 1;
+
+	if (o->new) {
+		list_del(&o->new_list);
+		c->new_orphans -= 1;
+	}
+
+	kfree(o);
+}
+
+static void orphan_delete(struct ubifs_info *c, struct ubifs_orphan *orph)
+{
+	if (orph->del) {
+		dbg_gen("deleted twice ino %lu", (unsigned long)orph->inum);
+		return;
+	}
+
+	if (orph->cmt) {
+		orph->del = 1;
+		rb_erase(&orph->rb, &c->orph_tree);
+		orph->dnext = c->orph_dnext;
+		c->orph_dnext = orph;
+		dbg_gen("delete later ino %lu", (unsigned long)orph->inum);
+		return;
+	}
+
+	__orphan_drop(c, orph);
+}
+
+/**
+ * ubifs_delete_orphan - delete an orphan.
+ * @c: UBIFS file-system description object
+ * @inum: orphan inode number
+ *
+ * Delete an orphan. This function is called when an inode is deleted.
+ */
+void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
+{
+	struct ubifs_orphan *orph;
+
+	spin_lock(&c->orphan_lock);
+
+	orph = lookup_orphan(c, inum);
+	if (!orph) {
+		spin_unlock(&c->orphan_lock);
+		ubifs_err(c, "missing orphan ino %lu", (unsigned long)inum);
+		dump_stack();
+
+		return;
+	}
+
+	orphan_delete(c, orph);
+
+	spin_unlock(&c->orphan_lock);
+}
+
+/**
+ * ubifs_orphan_start_commit - start commit of orphans.
+ * @c: UBIFS file-system description object
+ *
+ * Start commit of orphans.
+ */
+int ubifs_orphan_start_commit(struct ubifs_info *c)
+{
+	struct ubifs_orphan *orphan, **last;
+
+	spin_lock(&c->orphan_lock);
+	last = &c->orph_cnext;
+	list_for_each_entry(orphan, &c->orph_new, new_list) {
+		ubifs_assert(c, orphan->new);
+		ubifs_assert(c, !orphan->cmt);
+		orphan->new = 0;
+		orphan->cmt = 1;
+		*last = orphan;
+		last = &orphan->cnext;
+	}
+	*last = NULL;
+	c->cmt_orphans = c->new_orphans;
+	c->new_orphans = 0;
+	dbg_cmt("%d orphans to commit", c->cmt_orphans);
+	INIT_LIST_HEAD(&c->orph_new);
+	if (c->tot_orphans == 0)
+		c->no_orphs = 1;
+	else
+		c->no_orphs = 0;
+	spin_unlock(&c->orphan_lock);
+	return 0;
+}
+
+/**
+ * avail_orphs - calculate available space.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns the number of orphans that can be written in the
+ * available space.
+ */
+static int avail_orphs(struct ubifs_info *c)
+{
+	int avail_lebs, avail, gap;
+
+	avail_lebs = c->orph_lebs - (c->ohead_lnum - c->orph_first) - 1;
+	avail = avail_lebs *
+	       ((c->leb_size - UBIFS_ORPH_NODE_SZ) / sizeof(__le64));
+	gap = c->leb_size - c->ohead_offs;
+	if (gap >= UBIFS_ORPH_NODE_SZ + sizeof(__le64))
+		avail += (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
+	return avail;
+}
+
+/**
+ * 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;
+}
+
+/**
+ * do_write_orph_node - write a node to the orphan head.
+ * @c: UBIFS file-system description object
+ * @len: length of node
+ * @atomic: write atomically
+ *
+ * This function writes a node to the orphan head from the orphan buffer. If
+ * %atomic is not zero, then the write is done atomically. On success, %0 is
+ * returned, otherwise a negative error code is returned.
+ */
+static int do_write_orph_node(struct ubifs_info *c, int len, int atomic)
+{
+	int err = 0;
+
+	if (atomic) {
+		ubifs_assert(c, c->ohead_offs == 0);
+		ubifs_prepare_node(c, c->orph_buf, len, 1);
+		len = ALIGN(len, c->min_io_size);
+		err = ubifs_leb_change(c, c->ohead_lnum, c->orph_buf, len);
+	} else {
+		if (c->ohead_offs == 0) {
+			/* Ensure LEB has been unmapped */
+			err = ubifs_leb_unmap(c, c->ohead_lnum);
+			if (err)
+				return err;
+		}
+		err = ubifs_write_node(c, c->orph_buf, len, c->ohead_lnum,
+				       c->ohead_offs);
+	}
+	return err;
+}
+
+/**
+ * write_orph_node - write an orphan node.
+ * @c: UBIFS file-system description object
+ * @atomic: write atomically
+ *
+ * This function builds an orphan node from the cnext list and writes it to the
+ * orphan head. On success, %0 is returned, otherwise a negative error code
+ * is returned.
+ */
+static int write_orph_node(struct ubifs_info *c, int atomic)
+{
+	struct ubifs_orphan *orphan, *cnext;
+	struct ubifs_orph_node *orph;
+	int gap, err, len, cnt, i;
+
+	ubifs_assert(c, c->cmt_orphans > 0);
+	gap = c->leb_size - c->ohead_offs;
+	if (gap < UBIFS_ORPH_NODE_SZ + sizeof(__le64)) {
+		c->ohead_lnum += 1;
+		c->ohead_offs = 0;
+		gap = c->leb_size;
+		if (c->ohead_lnum > c->orph_last) {
+			/*
+			 * We limit the number of orphans so that this should
+			 * never happen.
+			 */
+			ubifs_err(c, "out of space in orphan area");
+			return -EINVAL;
+		}
+	}
+	cnt = (gap - UBIFS_ORPH_NODE_SZ) / sizeof(__le64);
+	if (cnt > c->cmt_orphans)
+		cnt = c->cmt_orphans;
+	len = UBIFS_ORPH_NODE_SZ + cnt * sizeof(__le64);
+	ubifs_assert(c, c->orph_buf);
+	orph = c->orph_buf;
+	orph->ch.node_type = UBIFS_ORPH_NODE;
+	spin_lock(&c->orphan_lock);
+	cnext = c->orph_cnext;
+	for (i = 0; i < cnt; i++) {
+		orphan = cnext;
+		ubifs_assert(c, orphan->cmt);
+		orph->inos[i] = cpu_to_le64(orphan->inum);
+		orphan->cmt = 0;
+		cnext = orphan->cnext;
+		orphan->cnext = NULL;
+	}
+	c->orph_cnext = cnext;
+	c->cmt_orphans -= cnt;
+	spin_unlock(&c->orphan_lock);
+	if (c->cmt_orphans)
+		orph->cmt_no = cpu_to_le64(c->cmt_no);
+	else
+		/* Mark the last node of the commit */
+		orph->cmt_no = cpu_to_le64((c->cmt_no) | (1ULL << 63));
+	ubifs_assert(c, c->ohead_offs + len <= c->leb_size);
+	ubifs_assert(c, c->ohead_lnum >= c->orph_first);
+	ubifs_assert(c, c->ohead_lnum <= c->orph_last);
+	err = do_write_orph_node(c, len, atomic);
+	c->ohead_offs += ALIGN(len, c->min_io_size);
+	c->ohead_offs = ALIGN(c->ohead_offs, 8);
+	return err;
+}
+
+/**
+ * write_orph_nodes - write orphan nodes until there are no more to commit.
+ * @c: UBIFS file-system description object
+ * @atomic: write atomically
+ *
+ * This function writes orphan nodes for all the orphans to commit. On success,
+ * %0 is returned, otherwise a negative error code is returned.
+ */
+static int write_orph_nodes(struct ubifs_info *c, int atomic)
+{
+	int err;
+
+	while (c->cmt_orphans > 0) {
+		err = write_orph_node(c, atomic);
+		if (err)
+			return err;
+	}
+	if (atomic) {
+		int lnum;
+
+		/* Unmap any unused LEBs after consolidation */
+		for (lnum = c->ohead_lnum + 1; lnum <= c->orph_last; lnum++) {
+			err = ubifs_leb_unmap(c, lnum);
+			if (err)
+				return err;
+		}
+	}
+	return 0;
+}
+
+/**
+ * consolidate - consolidate the orphan area.
+ * @c: UBIFS file-system description object
+ *
+ * This function enables consolidation by putting all the orphans into the list
+ * to commit. The list is in the order that the orphans were added, and the
+ * LEBs are written atomically in order, so at no time can orphans be lost by
+ * an unclean unmount.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int consolidate(struct ubifs_info *c)
+{
+	int tot_avail = tot_avail_orphs(c), err = 0;
+
+	spin_lock(&c->orphan_lock);
+	dbg_cmt("there is space for %d orphans and there are %d",
+		tot_avail, c->tot_orphans);
+	if (c->tot_orphans - c->new_orphans <= tot_avail) {
+		struct ubifs_orphan *orphan, **last;
+		int cnt = 0;
+
+		/* Change the cnext list to include all non-new orphans */
+		last = &c->orph_cnext;
+		list_for_each_entry(orphan, &c->orph_list, list) {
+			if (orphan->new)
+				continue;
+			orphan->cmt = 1;
+			*last = orphan;
+			last = &orphan->cnext;
+			cnt += 1;
+		}
+		*last = NULL;
+		ubifs_assert(c, cnt == c->tot_orphans - c->new_orphans);
+		c->cmt_orphans = cnt;
+		c->ohead_lnum = c->orph_first;
+		c->ohead_offs = 0;
+	} else {
+		/*
+		 * We limit the number of orphans so that this should
+		 * never happen.
+		 */
+		ubifs_err(c, "out of space in orphan area");
+		err = -EINVAL;
+	}
+	spin_unlock(&c->orphan_lock);
+	return err;
+}
+
+/**
+ * commit_orphans - commit orphans.
+ * @c: UBIFS file-system description object
+ *
+ * This function commits orphans to flash. On success, %0 is returned,
+ * otherwise a negative error code is returned.
+ */
+static int commit_orphans(struct ubifs_info *c)
+{
+	int avail, atomic = 0, err;
+
+	ubifs_assert(c, c->cmt_orphans > 0);
+	avail = avail_orphs(c);
+	if (avail < c->cmt_orphans) {
+		/* Not enough space to write new orphans, so consolidate */
+		err = consolidate(c);
+		if (err)
+			return err;
+		atomic = 1;
+	}
+	err = write_orph_nodes(c, atomic);
+	return err;
+}
+
+/**
+ * erase_deleted - erase the orphans marked for deletion.
+ * @c: UBIFS file-system description object
+ *
+ * During commit, the orphans being committed cannot be deleted, so they are
+ * marked for deletion and deleted by this function. Also, the recovery
+ * adds killed orphans to the deletion list, and therefore they are deleted
+ * here too.
+ */
+static void erase_deleted(struct ubifs_info *c)
+{
+	struct ubifs_orphan *orphan, *dnext;
+
+	spin_lock(&c->orphan_lock);
+	dnext = c->orph_dnext;
+	while (dnext) {
+		orphan = dnext;
+		dnext = orphan->dnext;
+		ubifs_assert(c, !orphan->new);
+		ubifs_assert(c, orphan->del);
+		list_del(&orphan->list);
+		c->tot_orphans -= 1;
+		dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
+		kfree(orphan);
+	}
+	c->orph_dnext = NULL;
+	spin_unlock(&c->orphan_lock);
+}
+
+/**
+ * ubifs_orphan_end_commit - end commit of orphans.
+ * @c: UBIFS file-system description object
+ *
+ * End commit of orphans.
+ */
+int ubifs_orphan_end_commit(struct ubifs_info *c)
+{
+	int err;
+
+	if (c->cmt_orphans != 0) {
+		err = commit_orphans(c);
+		if (err)
+			return err;
+	}
+	erase_deleted(c);
+	err = dbg_check_orphans(c);
+	return err;
+}
+
+/**
+ * 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;
+}
+
+/**
+ * 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;
+	struct ubifs_ino_node *ino = NULL;
+	unsigned long long cmt_no;
+	ino_t inum;
+	int i, n, err, first = 1;
+
+	ino = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
+	if (!ino)
+		return -ENOMEM;
+
+	list_for_each_entry(snod, &sleb->nodes, list) {
+		if (snod->type != UBIFS_ORPH_NODE) {
+			ubifs_err(c, "invalid node type %d in orphan area at %d:%d",
+				  snod->type, sleb->lnum, snod->offs);
+			ubifs_dump_node(c, snod->node,
+					c->leb_size - snod->offs);
+			err = -EINVAL;
+			goto out_free;
+		}
+
+		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(c, "out of order commit number %llu in orphan node at %d:%d",
+					  cmt_no, sleb->lnum, snod->offs);
+				ubifs_dump_node(c, snod->node,
+						c->leb_size - snod->offs);
+				err = -EINVAL;
+				goto out_free;
+			}
+			dbg_rcvry("out of date LEB %d", sleb->lnum);
+			*outofdate = 1;
+			err = 0;
+			goto out_free;
+		}
+
+		if (first)
+			first = 0;
+
+		n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
+		for (i = 0; i < n; i++) {
+			union ubifs_key key;
+
+			inum = le64_to_cpu(orph->inos[i]);
+
+			ino_key_init(c, &key, inum);
+			err = ubifs_tnc_lookup(c, &key, ino);
+			if (err && err != -ENOENT)
+				goto out_free;
+
+			/*
+			 * Check whether an inode can really get deleted.
+			 * linkat() with O_TMPFILE allows rebirth of an inode.
+			 */
+			if (err == 0 && ino->nlink == 0) {
+				dbg_rcvry("deleting orphaned inode %lu",
+					  (unsigned long)inum);
+
+				err = ubifs_tnc_remove_ino(c, inum);
+				if (err)
+					goto out_ro;
+			}
+		}
+
+		*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;
+	}
+
+	err = 0;
+out_free:
+	kfree(ino);
+	return err;
+
+out_ro:
+	ubifs_ro_mode(c, err);
+	kfree(ino);
+	return err;
+}
+
+/**
+ * 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, 1);
+		if (IS_ERR(sleb)) {
+			if (PTR_ERR(sleb) == -EUCLEAN)
+				sleb = ubifs_recover_leb(c, lnum, 0,
+							 c->sbuf, -1);
+			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;
+}
+
+/*
+ * Everything below is related to debugging.
+ */
+
+struct check_orphan {
+	struct rb_node rb;
+	ino_t inum;
+};
+
+struct check_info {
+	unsigned long last_ino;
+	unsigned long tot_inos;
+	unsigned long missing;
+	unsigned long long leaf_cnt;
+	struct ubifs_ino_node *node;
+	struct rb_root root;
+};
+
+static bool dbg_find_orphan(struct ubifs_info *c, ino_t inum)
+{
+	bool found = false;
+
+	spin_lock(&c->orphan_lock);
+	found = !!lookup_orphan(c, inum);
+	spin_unlock(&c->orphan_lock);
+
+	return found;
+}
+
+static int dbg_ins_check_orphan(struct rb_root *root, ino_t inum)
+{
+	struct check_orphan *orphan, *o;
+	struct rb_node **p, *parent = NULL;
+
+	orphan = kzalloc_obj(struct check_orphan, GFP_NOFS);
+	if (!orphan)
+		return -ENOMEM;
+	orphan->inum = inum;
+
+	p = &root->rb_node;
+	while (*p) {
+		parent = *p;
+		o = rb_entry(parent, struct check_orphan, rb);
+		if (inum < o->inum)
+			p = &(*p)->rb_left;
+		else if (inum > o->inum)
+			p = &(*p)->rb_right;
+		else {
+			kfree(orphan);
+			return 0;
+		}
+	}
+	rb_link_node(&orphan->rb, parent, p);
+	rb_insert_color(&orphan->rb, root);
+	return 0;
+}
+
+static int dbg_find_check_orphan(struct rb_root *root, ino_t inum)
+{
+	struct check_orphan *o;
+	struct rb_node *p;
+
+	p = root->rb_node;
+	while (p) {
+		o = rb_entry(p, struct check_orphan, rb);
+		if (inum < o->inum)
+			p = p->rb_left;
+		else if (inum > o->inum)
+			p = p->rb_right;
+		else
+			return 1;
+	}
+	return 0;
+}
+
+static void dbg_free_check_tree(struct rb_root *root)
+{
+	struct check_orphan *o, *n;
+
+	rbtree_postorder_for_each_entry_safe(o, n, root, rb)
+		kfree(o);
+}
+
+static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+			    void *priv)
+{
+	struct check_info *ci = priv;
+	ino_t inum;
+	int err;
+
+	inum = key_inum(c, &zbr->key);
+	if (inum != ci->last_ino) {
+		/*
+		 * Lowest node type is the inode node or xattr entry(when
+		 * selinux/encryption is enabled), so it comes first
+		 */
+		if (key_type(c, &zbr->key) != UBIFS_INO_KEY &&
+		    key_type(c, &zbr->key) != UBIFS_XENT_KEY)
+			ubifs_err(c, "found orphan node ino %lu, type %d",
+				  (unsigned long)inum, key_type(c, &zbr->key));
+		ci->last_ino = inum;
+		ci->tot_inos += 1;
+		err = ubifs_tnc_read_node(c, zbr, ci->node);
+		if (err) {
+			ubifs_err(c, "node read failed, error %d", err);
+			return err;
+		}
+		if (ci->node->nlink == 0)
+			/* Must be recorded as an orphan */
+			if (!dbg_find_check_orphan(&ci->root, inum) &&
+			    !dbg_find_orphan(c, inum)) {
+				ubifs_err(c, "missing orphan, ino %lu",
+					  (unsigned long)inum);
+				ci->missing += 1;
+			}
+	}
+	ci->leaf_cnt += 1;
+	return 0;
+}
+
+static int dbg_read_orphans(struct check_info *ci, struct ubifs_scan_leb *sleb)
+{
+	struct ubifs_scan_node *snod;
+	struct ubifs_orph_node *orph;
+	ino_t inum;
+	int i, n, err;
+
+	list_for_each_entry(snod, &sleb->nodes, list) {
+		cond_resched();
+		if (snod->type != UBIFS_ORPH_NODE)
+			continue;
+		orph = snod->node;
+		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]);
+			err = dbg_ins_check_orphan(&ci->root, inum);
+			if (err)
+				return err;
+		}
+	}
+	return 0;
+}
+
+static int dbg_scan_orphans(struct ubifs_info *c, struct check_info *ci)
+{
+	int lnum, err = 0;
+	void *buf;
+
+	/* Check no-orphans flag and skip this if no orphans */
+	if (c->no_orphs)
+		return 0;
+
+	buf = __vmalloc(c->leb_size, GFP_NOFS);
+	if (!buf) {
+		ubifs_err(c, "cannot allocate memory to check orphans");
+		return 0;
+	}
+
+	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
+		struct ubifs_scan_leb *sleb;
+
+		sleb = ubifs_scan(c, lnum, 0, buf, 0);
+		if (IS_ERR(sleb)) {
+			err = PTR_ERR(sleb);
+			break;
+		}
+
+		err = dbg_read_orphans(ci, sleb);
+		ubifs_scan_destroy(sleb);
+		if (err)
+			break;
+	}
+
+	vfree(buf);
+	return err;
+}
+
+static int dbg_check_orphans(struct ubifs_info *c)
+{
+	struct check_info ci;
+	int err;
+
+	if (!dbg_is_chk_orph(c))
+		return 0;
+
+	ci.last_ino = 0;
+	ci.tot_inos = 0;
+	ci.missing  = 0;
+	ci.leaf_cnt = 0;
+	ci.root = RB_ROOT;
+	ci.node = kmalloc(UBIFS_MAX_INO_NODE_SZ, GFP_NOFS);
+	if (!ci.node) {
+		ubifs_err(c, "out of memory");
+		return -ENOMEM;
+	}
+
+	err = dbg_scan_orphans(c, &ci);
+	if (err)
+		goto out;
+
+	err = dbg_walk_index(c, &dbg_orphan_check, NULL, &ci);
+	if (err) {
+		ubifs_err(c, "cannot scan TNC, error %d", err);
+		goto out;
+	}
+
+	if (ci.missing) {
+		ubifs_err(c, "%lu missing orphan(s)", ci.missing);
+		err = -EINVAL;
+		goto out;
+	}
+
+	dbg_cmt("last inode number is %lu", ci.last_ino);
+	dbg_cmt("total number of inodes is %lu", ci.tot_inos);
+	dbg_cmt("total number of leaf nodes is %llu", ci.leaf_cnt);
+
+out:
+	dbg_free_check_tree(&ci.root);
+	kfree(ci.node);
+	return err;
+}

linux-7.2.3/fs/ubifs/sb.c

@@ -0,0 +1,956 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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"
+#include <linux/slab.h>
+#include <linux/math64.h>
+#include <linux/uuid.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
+
+static int get_default_compressor(struct ubifs_info *c)
+{
+	if (ubifs_compr_present(c, UBIFS_COMPR_ZSTD))
+		return UBIFS_COMPR_ZSTD;
+
+	if (ubifs_compr_present(c, UBIFS_COMPR_LZO))
+		return UBIFS_COMPR_LZO;
+
+	if (ubifs_compr_present(c, UBIFS_COMPR_ZLIB))
+		return UBIFS_COMPR_ZLIB;
+
+	return UBIFS_COMPR_NONE;
+}
+
+/**
+ * create_default_filesystem - format empty UBI volume.
+ * @c: UBIFS file-system description object
+ *
+ * This function creates default empty file-system. Returns zero in case of
+ * success and a negative error code in case of failure.
+ */
+static int create_default_filesystem(struct ubifs_info *c)
+{
+	struct ubifs_sb_node *sup;
+	struct ubifs_mst_node *mst;
+	struct ubifs_idx_node *idx;
+	struct ubifs_branch *br;
+	struct ubifs_ino_node *ino;
+	struct ubifs_cs_node *cs;
+	union ubifs_key key;
+	int err, tmp, jnl_lebs, log_lebs, max_buds, main_lebs, main_first;
+	int lpt_lebs, lpt_first, orph_lebs, big_lpt, ino_waste, sup_flags = 0;
+	int min_leb_cnt = UBIFS_MIN_LEB_CNT;
+	int idx_node_size;
+	long long tmp64, main_bytes;
+	__le64 tmp_le64;
+	struct timespec64 ts;
+	u8 hash[UBIFS_HASH_ARR_SZ];
+	u8 hash_lpt[UBIFS_HASH_ARR_SZ];
+
+	/* Some functions called from here depend on the @c->key_len filed */
+	c->key_len = UBIFS_SK_LEN;
+
+	/*
+	 * First of all, we have to calculate default file-system geometry -
+	 * log size, journal size, etc.
+	 */
+	if (c->leb_cnt < 0x7FFFFFFF / DEFAULT_JNL_PERCENT)
+		/* We can first multiply then divide and have no overflow */
+		jnl_lebs = c->leb_cnt * DEFAULT_JNL_PERCENT / 100;
+	else
+		jnl_lebs = (c->leb_cnt / 100) * DEFAULT_JNL_PERCENT;
+
+	if (jnl_lebs < UBIFS_MIN_JNL_LEBS)
+		jnl_lebs = UBIFS_MIN_JNL_LEBS;
+	if (jnl_lebs * c->leb_size > DEFAULT_MAX_JNL)
+		jnl_lebs = DEFAULT_MAX_JNL / c->leb_size;
+
+	/*
+	 * The log should be large enough to fit reference nodes for all bud
+	 * LEBs. Because buds do not have to start from the beginning of LEBs
+	 * (half of the LEB may contain committed data), the log should
+	 * generally be larger, make it twice as large.
+	 */
+	tmp = 2 * (c->ref_node_alsz * jnl_lebs) + c->leb_size - 1;
+	log_lebs = tmp / c->leb_size;
+	/* Plus one LEB reserved for commit */
+	log_lebs += 1;
+	if (c->leb_cnt - min_leb_cnt > 8) {
+		/* And some extra space to allow writes while committing */
+		log_lebs += 1;
+		min_leb_cnt += 1;
+	}
+
+	max_buds = jnl_lebs - log_lebs;
+	if (max_buds < UBIFS_MIN_BUD_LEBS)
+		max_buds = UBIFS_MIN_BUD_LEBS;
+
+	/*
+	 * Orphan nodes are stored in a separate area. One node can store a lot
+	 * of orphan inode numbers, but when new orphan comes we just add a new
+	 * orphan node. At some point the nodes are consolidated into one
+	 * orphan node.
+	 */
+	orph_lebs = UBIFS_MIN_ORPH_LEBS;
+	if (c->leb_cnt - min_leb_cnt > 1)
+		/*
+		 * For debugging purposes it is better to have at least 2
+		 * orphan LEBs, because the orphan subsystem would need to do
+		 * consolidations and would be stressed more.
+		 */
+		orph_lebs += 1;
+
+	main_lebs = c->leb_cnt - UBIFS_SB_LEBS - UBIFS_MST_LEBS - log_lebs;
+	main_lebs -= orph_lebs;
+
+	lpt_first = UBIFS_LOG_LNUM + log_lebs;
+	c->lsave_cnt = DEFAULT_LSAVE_CNT;
+	c->max_leb_cnt = c->leb_cnt;
+	err = ubifs_create_dflt_lpt(c, &main_lebs, lpt_first, &lpt_lebs,
+				    &big_lpt, hash_lpt);
+	if (err)
+		return err;
+
+	dbg_gen("LEB Properties Tree created (LEBs %d-%d)", lpt_first,
+		lpt_first + lpt_lebs - 1);
+
+	main_first = c->leb_cnt - main_lebs;
+
+	sup = kzalloc(ALIGN(UBIFS_SB_NODE_SZ, c->min_io_size), GFP_KERNEL);
+	mst = kzalloc(c->mst_node_alsz, GFP_KERNEL);
+	idx_node_size = ubifs_idx_node_sz(c, 1);
+	idx = kzalloc(ALIGN(idx_node_size, c->min_io_size), GFP_KERNEL);
+	ino = kzalloc(ALIGN(UBIFS_INO_NODE_SZ, c->min_io_size), GFP_KERNEL);
+	cs = kzalloc(ALIGN(UBIFS_CS_NODE_SZ, c->min_io_size), GFP_KERNEL);
+
+	if (!sup || !mst || !idx || !ino || !cs) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	/* Create default superblock */
+
+	tmp64 = (long long)max_buds * c->leb_size;
+	if (big_lpt)
+		sup_flags |= UBIFS_FLG_BIGLPT;
+	if (ubifs_default_version > 4)
+		sup_flags |= UBIFS_FLG_DOUBLE_HASH;
+
+	if (ubifs_authenticated(c)) {
+		sup_flags |= UBIFS_FLG_AUTHENTICATION;
+		sup->hash_algo = cpu_to_le16(c->auth_hash_algo);
+		err = ubifs_hmac_wkm(c, sup->hmac_wkm);
+		if (err)
+			goto out;
+	} else {
+		sup->hash_algo = cpu_to_le16(0xffff);
+	}
+
+	sup->ch.node_type  = UBIFS_SB_NODE;
+	sup->key_hash      = UBIFS_KEY_HASH_R5;
+	sup->flags         = cpu_to_le32(sup_flags);
+	sup->min_io_size   = cpu_to_le32(c->min_io_size);
+	sup->leb_size      = cpu_to_le32(c->leb_size);
+	sup->leb_cnt       = cpu_to_le32(c->leb_cnt);
+	sup->max_leb_cnt   = cpu_to_le32(c->max_leb_cnt);
+	sup->max_bud_bytes = cpu_to_le64(tmp64);
+	sup->log_lebs      = cpu_to_le32(log_lebs);
+	sup->lpt_lebs      = cpu_to_le32(lpt_lebs);
+	sup->orph_lebs     = cpu_to_le32(orph_lebs);
+	sup->jhead_cnt     = cpu_to_le32(DEFAULT_JHEADS_CNT);
+	sup->fanout        = cpu_to_le32(DEFAULT_FANOUT);
+	sup->lsave_cnt     = cpu_to_le32(c->lsave_cnt);
+	sup->fmt_version   = cpu_to_le32(ubifs_default_version);
+	sup->time_gran     = cpu_to_le32(DEFAULT_TIME_GRAN);
+	if (c->mount_opts.override_compr)
+		sup->default_compr = cpu_to_le16(c->mount_opts.compr_type);
+	else
+		sup->default_compr = cpu_to_le16(get_default_compressor(c));
+
+	generate_random_uuid(sup->uuid);
+
+	main_bytes = (long long)main_lebs * c->leb_size;
+	tmp64 = div_u64(main_bytes * DEFAULT_RP_PERCENT, 100);
+	if (tmp64 > DEFAULT_MAX_RP_SIZE)
+		tmp64 = DEFAULT_MAX_RP_SIZE;
+	sup->rp_size = cpu_to_le64(tmp64);
+	sup->ro_compat_version = cpu_to_le32(UBIFS_RO_COMPAT_VERSION);
+
+	dbg_gen("default superblock created at LEB 0:0");
+
+	/* Create default master node */
+
+	mst->ch.node_type = UBIFS_MST_NODE;
+	mst->log_lnum     = cpu_to_le32(UBIFS_LOG_LNUM);
+	mst->highest_inum = cpu_to_le64(UBIFS_FIRST_INO);
+	mst->cmt_no       = 0;
+	mst->root_lnum    = cpu_to_le32(main_first + DEFAULT_IDX_LEB);
+	mst->root_offs    = 0;
+	tmp = ubifs_idx_node_sz(c, 1);
+	mst->root_len     = cpu_to_le32(tmp);
+	mst->gc_lnum      = cpu_to_le32(main_first + DEFAULT_GC_LEB);
+	mst->ihead_lnum   = cpu_to_le32(main_first + DEFAULT_IDX_LEB);
+	mst->ihead_offs   = cpu_to_le32(ALIGN(tmp, c->min_io_size));
+	mst->index_size   = cpu_to_le64(ALIGN(tmp, 8));
+	mst->lpt_lnum     = cpu_to_le32(c->lpt_lnum);
+	mst->lpt_offs     = cpu_to_le32(c->lpt_offs);
+	mst->nhead_lnum   = cpu_to_le32(c->nhead_lnum);
+	mst->nhead_offs   = cpu_to_le32(c->nhead_offs);
+	mst->ltab_lnum    = cpu_to_le32(c->ltab_lnum);
+	mst->ltab_offs    = cpu_to_le32(c->ltab_offs);
+	mst->lsave_lnum   = cpu_to_le32(c->lsave_lnum);
+	mst->lsave_offs   = cpu_to_le32(c->lsave_offs);
+	mst->lscan_lnum   = cpu_to_le32(main_first);
+	mst->empty_lebs   = cpu_to_le32(main_lebs - 2);
+	mst->idx_lebs     = cpu_to_le32(1);
+	mst->leb_cnt      = cpu_to_le32(c->leb_cnt);
+	ubifs_copy_hash(c, hash_lpt, mst->hash_lpt);
+
+	/* Calculate lprops statistics */
+	tmp64 = main_bytes;
+	tmp64 -= ALIGN(ubifs_idx_node_sz(c, 1), c->min_io_size);
+	tmp64 -= ALIGN(UBIFS_INO_NODE_SZ, c->min_io_size);
+	mst->total_free = cpu_to_le64(tmp64);
+
+	tmp64 = ALIGN(ubifs_idx_node_sz(c, 1), c->min_io_size);
+	ino_waste = ALIGN(UBIFS_INO_NODE_SZ, c->min_io_size) -
+			  UBIFS_INO_NODE_SZ;
+	tmp64 += ino_waste;
+	tmp64 -= ALIGN(ubifs_idx_node_sz(c, 1), 8);
+	mst->total_dirty = cpu_to_le64(tmp64);
+
+	/*  The indexing LEB does not contribute to dark space */
+	tmp64 = ((long long)(c->main_lebs - 1) * c->dark_wm);
+	mst->total_dark = cpu_to_le64(tmp64);
+
+	mst->total_used = cpu_to_le64(UBIFS_INO_NODE_SZ);
+
+	dbg_gen("default master node created at LEB %d:0", UBIFS_MST_LNUM);
+
+	/* Create the root indexing node */
+
+	c->key_fmt = UBIFS_SIMPLE_KEY_FMT;
+	c->key_hash = key_r5_hash;
+
+	idx->ch.node_type = UBIFS_IDX_NODE;
+	idx->child_cnt = cpu_to_le16(1);
+	ino_key_init(c, &key, UBIFS_ROOT_INO);
+	br = ubifs_idx_branch(c, idx, 0);
+	key_write_idx(c, &key, &br->key);
+	br->lnum = cpu_to_le32(main_first + DEFAULT_DATA_LEB);
+	br->len  = cpu_to_le32(UBIFS_INO_NODE_SZ);
+
+	dbg_gen("default root indexing node created LEB %d:0",
+		main_first + DEFAULT_IDX_LEB);
+
+	/* Create default root inode */
+
+	ino_key_init_flash(c, &ino->key, UBIFS_ROOT_INO);
+	ino->ch.node_type = UBIFS_INO_NODE;
+	ino->creat_sqnum = cpu_to_le64(++c->max_sqnum);
+	ino->nlink = cpu_to_le32(2);
+
+	ktime_get_coarse_real_ts64(&ts);
+	tmp_le64 = cpu_to_le64(ts.tv_sec);
+	ino->atime_sec   = tmp_le64;
+	ino->ctime_sec   = tmp_le64;
+	ino->mtime_sec   = tmp_le64;
+	ino->atime_nsec  = 0;
+	ino->ctime_nsec  = 0;
+	ino->mtime_nsec  = 0;
+	ino->mode = cpu_to_le32(S_IFDIR | S_IRUGO | S_IWUSR | S_IXUGO);
+	ino->size = cpu_to_le64(UBIFS_INO_NODE_SZ);
+
+	/* Set compression enabled by default */
+	ino->flags = cpu_to_le32(UBIFS_COMPR_FL);
+
+	dbg_gen("root inode created at LEB %d:0",
+		main_first + DEFAULT_DATA_LEB);
+
+	/*
+	 * The first node in the log has to be the commit start node. This is
+	 * always the case during normal file-system operation. Write a fake
+	 * commit start node to the log.
+	 */
+
+	cs->ch.node_type = UBIFS_CS_NODE;
+
+	err = ubifs_write_node_hmac(c, sup, UBIFS_SB_NODE_SZ, 0, 0,
+				    offsetof(struct ubifs_sb_node, hmac));
+	if (err)
+		goto out;
+
+	err = ubifs_write_node(c, ino, UBIFS_INO_NODE_SZ,
+			       main_first + DEFAULT_DATA_LEB, 0);
+	if (err)
+		goto out;
+
+	ubifs_node_calc_hash(c, ino, hash);
+	ubifs_copy_hash(c, hash, ubifs_branch_hash(c, br));
+
+	err = ubifs_write_node(c, idx, idx_node_size, main_first + DEFAULT_IDX_LEB, 0);
+	if (err)
+		goto out;
+
+	ubifs_node_calc_hash(c, idx, hash);
+	ubifs_copy_hash(c, hash, mst->hash_root_idx);
+
+	err = ubifs_write_node_hmac(c, mst, UBIFS_MST_NODE_SZ, UBIFS_MST_LNUM, 0,
+		offsetof(struct ubifs_mst_node, hmac));
+	if (err)
+		goto out;
+
+	err = ubifs_write_node_hmac(c, mst, UBIFS_MST_NODE_SZ, UBIFS_MST_LNUM + 1,
+			       0, offsetof(struct ubifs_mst_node, hmac));
+	if (err)
+		goto out;
+
+	err = ubifs_write_node(c, cs, UBIFS_CS_NODE_SZ, UBIFS_LOG_LNUM, 0);
+	if (err)
+		goto out;
+
+	ubifs_msg(c, "default file-system created");
+
+	err = 0;
+out:
+	kfree(sup);
+	kfree(mst);
+	kfree(idx);
+	kfree(ino);
+	kfree(cs);
+
+	return err;
+}
+
+/**
+ * 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(c, "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(c, "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(c, "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(c, "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) {
+		ubifs_err(c, "too few main LEBs count %d, must be at least %d",
+			  c->main_lebs, UBIFS_MIN_MAIN_LEBS);
+		goto failed;
+	}
+
+	max_bytes = (long long)c->leb_size * UBIFS_MIN_BUD_LEBS;
+	if (c->max_bud_bytes < max_bytes) {
+		ubifs_err(c, "too small journal (%lld bytes), must be at least %lld bytes",
+			  c->max_bud_bytes, max_bytes);
+		goto failed;
+	}
+
+	max_bytes = (long long)c->leb_size * c->main_lebs;
+	if (c->max_bud_bytes > max_bytes) {
+		ubifs_err(c, "too large journal size (%lld bytes), only %lld bytes available in the main area",
+			  c->max_bud_bytes, max_bytes);
+		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 >= UBIFS_COMPR_TYPES_CNT) {
+		err = 13;
+		goto failed;
+	}
+
+	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;
+	}
+
+	if (!c->double_hash && c->fmt_version >= 5) {
+		err = 16;
+		goto failed;
+	}
+
+	if (c->encrypted && c->fmt_version < 5) {
+		err = 17;
+		goto failed;
+	}
+
+	return 0;
+
+failed:
+	ubifs_err(c, "bad superblock, error %d", err);
+	ubifs_dump_node(c, sup, ALIGN(UBIFS_SB_NODE_SZ, c->min_io_size));
+	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. Note, the user of this function is responsible of kfree()'ing the
+ * returned superblock buffer.
+ */
+static 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;
+}
+
+static int authenticate_sb_node(struct ubifs_info *c,
+				const struct ubifs_sb_node *sup)
+{
+	unsigned int sup_flags = le32_to_cpu(sup->flags);
+	u8 hmac_wkm[UBIFS_HMAC_ARR_SZ];
+	int authenticated = !!(sup_flags & UBIFS_FLG_AUTHENTICATION);
+	int hash_algo;
+	int err;
+
+	if (c->authenticated && !authenticated) {
+		ubifs_err(c, "authenticated FS forced, but found FS without authentication");
+		return -EINVAL;
+	}
+
+	if (!c->authenticated && authenticated) {
+		ubifs_err(c, "authenticated FS found, but no key given");
+		return -EINVAL;
+	}
+
+	ubifs_msg(c, "Mounting in %sauthenticated mode",
+		  c->authenticated ? "" : "un");
+
+	if (!c->authenticated)
+		return 0;
+
+	if (!IS_ENABLED(CONFIG_UBIFS_FS_AUTHENTICATION))
+		return -EOPNOTSUPP;
+
+	hash_algo = le16_to_cpu(sup->hash_algo);
+	if (hash_algo >= HASH_ALGO__LAST) {
+		ubifs_err(c, "superblock uses unknown hash algo %d",
+			  hash_algo);
+		return -EINVAL;
+	}
+
+	if (strcmp(hash_algo_name[hash_algo], c->auth_hash_name)) {
+		ubifs_err(c, "This filesystem uses %s for hashing,"
+			     " but %s is specified", hash_algo_name[hash_algo],
+			     c->auth_hash_name);
+		return -EINVAL;
+	}
+
+	/*
+	 * The super block node can either be authenticated by a HMAC or
+	 * by a signature in a ubifs_sig_node directly following the
+	 * super block node to support offline image creation.
+	 */
+	if (ubifs_hmac_zero(c, sup->hmac)) {
+		err = ubifs_sb_verify_signature(c, sup);
+	} else {
+		err = ubifs_hmac_wkm(c, hmac_wkm);
+		if (err)
+			return err;
+		if (ubifs_check_hmac(c, hmac_wkm, sup->hmac_wkm)) {
+			ubifs_err(c, "provided key does not fit");
+			return -ENOKEY;
+		}
+		err = ubifs_node_verify_hmac(c, sup, sizeof(*sup),
+					     offsetof(struct ubifs_sb_node,
+						      hmac));
+	}
+
+	if (err)
+		ubifs_err(c, "Failed to authenticate superblock: %d", err);
+
+	return err;
+}
+
+/**
+ * ubifs_write_sb_node - write superblock node.
+ * @c: UBIFS file-system description object
+ * @sup: superblock node read with 'ubifs_read_sb_node()'
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_write_sb_node(struct ubifs_info *c, struct ubifs_sb_node *sup)
+{
+	int len = ALIGN(UBIFS_SB_NODE_SZ, c->min_io_size);
+	int err;
+
+	err = ubifs_prepare_node_hmac(c, sup, UBIFS_SB_NODE_SZ,
+				      offsetof(struct ubifs_sb_node, hmac), 1);
+	if (err)
+		return err;
+
+	return ubifs_leb_change(c, UBIFS_SB_LNUM, sup, len);
+}
+
+/**
+ * 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) {
+		err = create_default_filesystem(c);
+		if (err)
+			return err;
+	}
+
+	sup = ubifs_read_sb_node(c);
+	if (IS_ERR(sup))
+		return PTR_ERR(sup);
+
+	c->sup_node = sup;
+
+	c->fmt_version = le32_to_cpu(sup->fmt_version);
+	c->ro_compat_version = le32_to_cpu(sup->ro_compat_version);
+
+	/*
+	 * The software supports all previous versions but not future versions,
+	 * due to the unavailability of time-travelling equipment.
+	 */
+	if (c->fmt_version > UBIFS_FORMAT_VERSION) {
+		ubifs_assert(c, !c->ro_media || c->ro_mount);
+		if (!c->ro_mount ||
+		    c->ro_compat_version > UBIFS_RO_COMPAT_VERSION) {
+			ubifs_err(c, "on-flash format version is w%d/r%d, but software only supports up to version w%d/r%d",
+				  c->fmt_version, c->ro_compat_version,
+				  UBIFS_FORMAT_VERSION,
+				  UBIFS_RO_COMPAT_VERSION);
+			if (c->ro_compat_version <= UBIFS_RO_COMPAT_VERSION) {
+				ubifs_msg(c, "only R/O mounting is possible");
+				err = -EROFS;
+			} else
+				err = -EINVAL;
+			goto out;
+		}
+
+		/*
+		 * The FS is mounted R/O, and the media format is
+		 * R/O-compatible with the UBIFS implementation, so we can
+		 * mount.
+		 */
+		c->rw_incompat = 1;
+	}
+
+	if (c->fmt_version < 3) {
+		ubifs_err(c, "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(c, "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->rp_size       = le64_to_cpu(sup->rp_size);
+	c->rp_uid        = make_kuid(&init_user_ns, le32_to_cpu(sup->rp_uid));
+	c->rp_gid        = make_kgid(&init_user_ns, le32_to_cpu(sup->rp_gid));
+	sup_flags        = le32_to_cpu(sup->flags);
+	if (!c->mount_opts.override_compr)
+		c->default_compr = le16_to_cpu(sup->default_compr);
+
+	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);
+	c->space_fixup = !!(sup_flags & UBIFS_FLG_SPACE_FIXUP);
+	c->double_hash = !!(sup_flags & UBIFS_FLG_DOUBLE_HASH);
+	c->encrypted = !!(sup_flags & UBIFS_FLG_ENCRYPTION);
+
+	err = authenticate_sb_node(c, sup);
+	if (err)
+		goto out;
+
+	if ((sup_flags & ~UBIFS_FLG_MASK) != 0) {
+		ubifs_err(c, "Unknown feature flags found: %#x",
+			  sup_flags & ~UBIFS_FLG_MASK);
+		err = -EINVAL;
+		goto out;
+	}
+
+	if (!IS_ENABLED(CONFIG_FS_ENCRYPTION) && c->encrypted) {
+		ubifs_err(c, "file system contains encrypted files but UBIFS"
+			     " was built without crypto support.");
+		err = -EINVAL;
+		goto out;
+	}
+
+	/* Automatically increase file system size to the maximum size */
+	if (c->leb_cnt < c->vi.size && c->leb_cnt < c->max_leb_cnt) {
+		int old_leb_cnt = c->leb_cnt;
+
+		c->leb_cnt = min_t(int, c->max_leb_cnt, c->vi.size);
+		sup->leb_cnt = cpu_to_le32(c->leb_cnt);
+
+		c->superblock_need_write = 1;
+
+		dbg_mnt("Auto resizing from %d LEBs to %d LEBs",
+			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;
+
+	err = validate_sb(c, sup);
+out:
+	return err;
+}
+
+/**
+ * fixup_leb - fixup/unmap an LEB containing free space.
+ * @c: UBIFS file-system description object
+ * @lnum: the LEB number to fix up
+ * @len: number of used bytes in LEB (starting at offset 0)
+ *
+ * This function reads the contents of the given LEB number @lnum, then fixes
+ * it up, so that empty min. I/O units in the end of LEB are actually erased on
+ * flash (rather than being just all-0xff real data). If the LEB is completely
+ * empty, it is simply unmapped.
+ */
+static int fixup_leb(struct ubifs_info *c, int lnum, int len)
+{
+	int err;
+
+	ubifs_assert(c, len >= 0);
+	ubifs_assert(c, len % c->min_io_size == 0);
+	ubifs_assert(c, len < c->leb_size);
+
+	if (len == 0) {
+		dbg_mnt("unmap empty LEB %d", lnum);
+		return ubifs_leb_unmap(c, lnum);
+	}
+
+	dbg_mnt("fixup LEB %d, data len %d", lnum, len);
+	err = ubifs_leb_read(c, lnum, c->sbuf, 0, len, 1);
+	if (err)
+		return err;
+
+	return ubifs_leb_change(c, lnum, c->sbuf, len);
+}
+
+/**
+ * fixup_free_space - find & remap all LEBs containing free space.
+ * @c: UBIFS file-system description object
+ *
+ * This function walks through all LEBs in the filesystem and fiexes up those
+ * containing free/empty space.
+ */
+static int fixup_free_space(struct ubifs_info *c)
+{
+	int lnum, err = 0;
+	struct ubifs_lprops *lprops;
+
+	ubifs_get_lprops(c);
+
+	/* Fixup LEBs in the master area */
+	for (lnum = UBIFS_MST_LNUM; lnum < UBIFS_LOG_LNUM; lnum++) {
+		err = fixup_leb(c, lnum, c->mst_offs + c->mst_node_alsz);
+		if (err)
+			goto out;
+	}
+
+	/* Unmap unused log LEBs */
+	lnum = ubifs_next_log_lnum(c, c->lhead_lnum);
+	while (lnum != c->ltail_lnum) {
+		err = fixup_leb(c, lnum, 0);
+		if (err)
+			goto out;
+		lnum = ubifs_next_log_lnum(c, lnum);
+	}
+
+	/*
+	 * Fixup the log head which contains the only a CS node at the
+	 * beginning.
+	 */
+	err = fixup_leb(c, c->lhead_lnum,
+			ALIGN(UBIFS_CS_NODE_SZ, c->min_io_size));
+	if (err)
+		goto out;
+
+	/* Fixup LEBs in the LPT area */
+	for (lnum = c->lpt_first; lnum <= c->lpt_last; lnum++) {
+		int free = c->ltab[lnum - c->lpt_first].free;
+
+		if (free > 0) {
+			err = fixup_leb(c, lnum, c->leb_size - free);
+			if (err)
+				goto out;
+		}
+	}
+
+	/* Unmap LEBs in the orphans area */
+	for (lnum = c->orph_first; lnum <= c->orph_last; lnum++) {
+		err = fixup_leb(c, lnum, 0);
+		if (err)
+			goto out;
+	}
+
+	/* Fixup LEBs in the main area */
+	for (lnum = c->main_first; lnum < c->leb_cnt; lnum++) {
+		lprops = ubifs_lpt_lookup(c, lnum);
+		if (IS_ERR(lprops)) {
+			err = PTR_ERR(lprops);
+			goto out;
+		}
+
+		if (lprops->free > 0) {
+			err = fixup_leb(c, lnum, c->leb_size - lprops->free);
+			if (err)
+				goto out;
+		}
+	}
+
+out:
+	ubifs_release_lprops(c);
+	return err;
+}
+
+/**
+ * ubifs_fixup_free_space - find & fix all LEBs with free space.
+ * @c: UBIFS file-system description object
+ *
+ * This function fixes up LEBs containing free space on first mount, if the
+ * appropriate flag was set when the FS was created. Each LEB with one or more
+ * empty min. I/O unit (i.e. free-space-count > 0) is re-written, to make sure
+ * the free space is actually erased. E.g., this is necessary for some NAND
+ * chips, since the free space may have been programmed like real "0xff" data
+ * (generating a non-0xff ECC), causing future writes to the not-really-erased
+ * NAND pages to behave badly. After the space is fixed up, the superblock flag
+ * is cleared, so that this is skipped for all future mounts.
+ */
+int ubifs_fixup_free_space(struct ubifs_info *c)
+{
+	int err;
+	struct ubifs_sb_node *sup = c->sup_node;
+
+	ubifs_assert(c, c->space_fixup);
+	ubifs_assert(c, !c->ro_mount);
+
+	ubifs_msg(c, "start fixing up free space");
+
+	err = fixup_free_space(c);
+	if (err)
+		return err;
+
+	/* Free-space fixup is no longer required */
+	c->space_fixup = 0;
+	sup->flags &= cpu_to_le32(~UBIFS_FLG_SPACE_FIXUP);
+
+	c->superblock_need_write = 1;
+
+	ubifs_msg(c, "free space fixup complete");
+	return err;
+}
+
+int ubifs_enable_encryption(struct ubifs_info *c)
+{
+	int err;
+	struct ubifs_sb_node *sup = c->sup_node;
+
+	if (!IS_ENABLED(CONFIG_FS_ENCRYPTION))
+		return -EOPNOTSUPP;
+
+	if (c->encrypted)
+		return 0;
+
+	if (c->ro_mount || c->ro_media)
+		return -EROFS;
+
+	if (c->fmt_version < 5) {
+		ubifs_err(c, "on-flash format version 5 is needed for encryption");
+		return -EINVAL;
+	}
+
+	sup->flags |= cpu_to_le32(UBIFS_FLG_ENCRYPTION);
+
+	err = ubifs_write_sb_node(c, sup);
+	if (!err)
+		c->encrypted = 1;
+
+	return err;
+}

linux-7.2.3/fs/ubifs/scan.c

@@ -0,0 +1,366 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation
+ *
+ * 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 at LEB %d:%d", lnum, offs);
+		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 at LEB %d:%d",
+		 dbg_ntype(ch->node_type), lnum, offs);
+
+	if (ubifs_check_node(c, buf, len, 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(c, "bad pad node at LEB %d:%d",
+					  lnum, offs);
+				ubifs_dump_node(c, pad, len);
+			}
+			return SCANNED_A_BAD_PAD_NODE;
+		}
+
+		/* Make the node pads to 8-byte boundary */
+		if ((node_len + pad_len) & 7) {
+			if (!quiet)
+				ubifs_err(c, "bad padding length %d - %d",
+					  offs, offs + node_len + pad_len);
+			return SCANNED_A_BAD_PAD_NODE;
+		}
+
+		dbg_scan("%d bytes padded at LEB %d:%d, offset now %d", pad_len,
+			 lnum, offs, 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 the scanned information 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_obj(struct ubifs_scan_leb, GFP_NOFS);
+	if (!sleb)
+		return ERR_PTR(-ENOMEM);
+
+	sleb->lnum = lnum;
+	INIT_LIST_HEAD(&sleb->nodes);
+	sleb->buf = sbuf;
+
+	err = ubifs_leb_read(c, lnum, sbuf + offs, offs, c->leb_size - offs, 0);
+	if (err && err != -EBADMSG) {
+		ubifs_err(c, "cannot read %d bytes from LEB %d:%d, error %d",
+			  c->leb_size - offs, lnum, offs, err);
+		kfree(sleb);
+		return ERR_PTR(err);
+	}
+
+	/*
+	 * Note, we ignore integrity errors (EBASMSG) because all the nodes are
+	 * protected by CRC checksums.
+	 */
+	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)
+ */
+void ubifs_end_scan(const struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+		    int lnum, int offs)
+{
+	dbg_scan("stop scanning LEB %d at offset %d", lnum, offs);
+	ubifs_assert(c, 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 = kmalloc_obj(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:
+		/*
+		 * The key is in the same place in all keyed
+		 * nodes.
+		 */
+		key_read(c, &ino->key, &snod->key);
+		break;
+	default:
+		invalid_key_init(c, &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(c, "corruption at LEB %d:%d", lnum, offs);
+	len = c->leb_size - offs;
+	if (len > 8192)
+		len = 8192;
+	ubifs_err(c, "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 of @c->leb_size bytes in size)
+ * @quiet: print no messages
+ *
+ * This function scans LEB number @lnum and returns complete information about
+ * its contents. Returns the scanned information in case of success and,
+ * %-EUCLEAN if the LEB neads recovery, and other negative error codes in case
+ * of failure.
+ *
+ * If @quiet is non-zero, this function does not print large and scary
+ * error messages and flash dumps in case of errors.
+ */
+struct ubifs_scan_leb *ubifs_scan(const struct ubifs_info *c, int lnum,
+				  int offs, void *sbuf, int quiet)
+{
+	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, quiet);
+		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:
+			ubifs_err(c, "garbage");
+			goto corrupted;
+		case SCANNED_A_NODE:
+			break;
+		case SCANNED_A_CORRUPT_NODE:
+		case SCANNED_A_BAD_PAD_NODE:
+			ubifs_err(c, "bad node");
+			goto corrupted;
+		default:
+			ubifs_err(c, "unknown");
+			err = -EINVAL;
+			goto error;
+		}
+
+		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) {
+		if (!quiet)
+			ubifs_err(c, "empty space starts at non-aligned offset %d",
+				  offs);
+		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) {
+			if (!quiet)
+				ubifs_err(c, "corrupt empty space at LEB %d:%d",
+					  lnum, offs);
+			goto corrupted;
+		}
+
+	return sleb;
+
+corrupted:
+	if (!quiet) {
+		ubifs_scanned_corruption(c, lnum, offs, buf);
+		ubifs_err(c, "LEB %d scanning failed", lnum);
+	}
+	err = -EUCLEAN;
+	ubifs_scan_destroy(sleb);
+	return ERR_PTR(err);
+
+error:
+	ubifs_err(c, "LEB %d scanning failed, error %d", lnum, err);
+	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);
+}

linux-7.2.3/fs/ubifs/shrinker.c

@@ -0,0 +1,319 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+/*
+ * This file implements UBIFS shrinker which evicts clean znodes from the TNC
+ * tree when Linux VM needs more RAM.
+ *
+ * We do not implement any LRU lists to find oldest znodes to free because it
+ * would add additional overhead to the file system fast paths. So the shrinker
+ * just walks the TNC tree when searching for znodes to free.
+ *
+ * If the root of a TNC sub-tree is clean and old enough, then the children are
+ * also clean and old enough. So the shrinker walks the TNC in level order and
+ * dumps entire sub-trees.
+ *
+ * The age of znodes is just the time-stamp when they were last looked at.
+ * The current shrinker first tries to evict old znodes, then young ones.
+ *
+ * Since the shrinker is global, it has to protect against races with FS
+ * un-mounts, which is done by the 'ubifs_infos_lock' and 'c->umount_mutex'.
+ */
+
+#include "ubifs.h"
+
+/* List of all UBIFS file-system instances */
+LIST_HEAD(ubifs_infos);
+
+/*
+ * We number each shrinker run and record the number on the ubifs_info structure
+ * so that we can easily work out which ubifs_info structures have already been
+ * done by the current run.
+ */
+static unsigned int shrinker_run_no;
+
+/* Protects 'ubifs_infos' list */
+DEFINE_SPINLOCK(ubifs_infos_lock);
+
+/* Global clean znode counter (for all mounted UBIFS instances) */
+atomic_long_t ubifs_clean_zn_cnt;
+
+/**
+ * shrink_tnc - shrink TNC tree.
+ * @c: UBIFS file-system description object
+ * @nr: number of znodes to free
+ * @age: the age of znodes to free
+ * @contention: if any contention, this is set to %1
+ *
+ * This function traverses TNC tree and frees clean znodes. It does not free
+ * clean znodes which younger then @age. Returns number of freed znodes.
+ */
+static int shrink_tnc(struct ubifs_info *c, int nr, int age, int *contention)
+{
+	int total_freed = 0;
+	struct ubifs_znode *znode, *zprev;
+	time64_t time = ktime_get_seconds();
+
+	ubifs_assert(c, mutex_is_locked(&c->umount_mutex));
+	ubifs_assert(c, mutex_is_locked(&c->tnc_mutex));
+
+	if (!c->zroot.znode || atomic_long_read(&c->clean_zn_cnt) == 0)
+		return 0;
+
+	/*
+	 * Traverse the TNC tree in levelorder manner, so that it is possible
+	 * to destroy large sub-trees. Indeed, if a znode is old, then all its
+	 * children are older or of the same age.
+	 *
+	 * Note, we are holding 'c->tnc_mutex', so we do not have to lock the
+	 * 'c->space_lock' when _reading_ 'c->clean_zn_cnt', because it is
+	 * changed only when the 'c->tnc_mutex' is held.
+	 */
+	zprev = NULL;
+	znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, NULL);
+	while (znode && total_freed < nr &&
+	       atomic_long_read(&c->clean_zn_cnt) > 0) {
+		int freed;
+
+		/*
+		 * If the znode is clean, but it is in the 'c->cnext' list, this
+		 * means that this znode has just been written to flash as a
+		 * part of commit and was marked clean. They will be removed
+		 * from the list at end commit. We cannot change the list,
+		 * because it is not protected by any mutex (design decision to
+		 * make commit really independent and parallel to main I/O). So
+		 * we just skip these znodes.
+		 *
+		 * Note, the 'clean_zn_cnt' counters are not updated until
+		 * after the commit, so the UBIFS shrinker does not report
+		 * the znodes which are in the 'c->cnext' list as freeable.
+		 *
+		 * Also note, if the root of a sub-tree is not in 'c->cnext',
+		 * then the whole sub-tree is not in 'c->cnext' as well, so it
+		 * is safe to dump whole sub-tree.
+		 */
+
+		if (znode->cnext) {
+			/*
+			 * Very soon these znodes will be removed from the list
+			 * and become freeable.
+			 */
+			*contention = 1;
+		} else if (!ubifs_zn_dirty(znode) &&
+			   abs(time - znode->time) >= age) {
+			if (znode->parent)
+				znode->parent->zbranch[znode->iip].znode = NULL;
+			else
+				c->zroot.znode = NULL;
+
+			freed = ubifs_destroy_tnc_subtree(c, znode);
+			atomic_long_sub(freed, &ubifs_clean_zn_cnt);
+			atomic_long_sub(freed, &c->clean_zn_cnt);
+			total_freed += freed;
+			znode = zprev;
+		}
+
+		if (unlikely(!c->zroot.znode))
+			break;
+
+		zprev = znode;
+		znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, znode);
+		cond_resched();
+	}
+
+	return total_freed;
+}
+
+/**
+ * shrink_tnc_trees - shrink UBIFS TNC trees.
+ * @nr: number of znodes to free
+ * @age: the age of znodes to free
+ * @contention: if any contention, this is set to %1
+ *
+ * This function walks the list of mounted UBIFS file-systems and frees clean
+ * znodes which are older than @age, until at least @nr znodes are freed.
+ * Returns the number of freed znodes.
+ */
+static int shrink_tnc_trees(int nr, int age, int *contention)
+{
+	struct ubifs_info *c;
+	struct list_head *p;
+	unsigned int run_no;
+	int freed = 0;
+
+	spin_lock(&ubifs_infos_lock);
+	do {
+		run_no = ++shrinker_run_no;
+	} while (run_no == 0);
+	/* Iterate over all mounted UBIFS file-systems and try to shrink them */
+	p = ubifs_infos.next;
+	while (p != &ubifs_infos) {
+		c = list_entry(p, struct ubifs_info, infos_list);
+		/*
+		 * We move the ones we do to the end of the list, so we stop
+		 * when we see one we have already done.
+		 */
+		if (c->shrinker_run_no == run_no)
+			break;
+		if (!mutex_trylock(&c->umount_mutex)) {
+			/* Some un-mount is in progress, try next FS */
+			*contention = 1;
+			p = p->next;
+			continue;
+		}
+		/*
+		 * We're holding 'c->umount_mutex', so the file-system won't go
+		 * away.
+		 */
+		if (!mutex_trylock(&c->tnc_mutex)) {
+			mutex_unlock(&c->umount_mutex);
+			*contention = 1;
+			p = p->next;
+			continue;
+		}
+		spin_unlock(&ubifs_infos_lock);
+		/*
+		 * OK, now we have TNC locked, the file-system cannot go away -
+		 * it is safe to reap the cache.
+		 */
+		c->shrinker_run_no = run_no;
+		freed += shrink_tnc(c, nr, age, contention);
+		mutex_unlock(&c->tnc_mutex);
+		spin_lock(&ubifs_infos_lock);
+		/* Get the next list element before we move this one */
+		p = p->next;
+		/*
+		 * Move this one to the end of the list to provide some
+		 * fairness.
+		 */
+		list_move_tail(&c->infos_list, &ubifs_infos);
+		mutex_unlock(&c->umount_mutex);
+		if (freed >= nr)
+			break;
+	}
+	spin_unlock(&ubifs_infos_lock);
+	return freed;
+}
+
+/**
+ * kick_a_thread - kick a background thread to start commit.
+ *
+ * This function kicks a background thread to start background commit. Returns
+ * %-1 if a thread was kicked or there is another reason to assume the memory
+ * will soon be freed or become freeable. If there are no dirty znodes, returns
+ * %0.
+ */
+static int kick_a_thread(void)
+{
+	int i;
+	struct ubifs_info *c;
+
+	/*
+	 * Iterate over all mounted UBIFS file-systems and find out if there is
+	 * already an ongoing commit operation there. If no, then iterate for
+	 * the second time and initiate background commit.
+	 */
+	spin_lock(&ubifs_infos_lock);
+	for (i = 0; i < 2; i++) {
+		list_for_each_entry(c, &ubifs_infos, infos_list) {
+			long dirty_zn_cnt;
+
+			if (!mutex_trylock(&c->umount_mutex)) {
+				/*
+				 * Some un-mount is in progress, it will
+				 * certainly free memory, so just return.
+				 */
+				spin_unlock(&ubifs_infos_lock);
+				return -1;
+			}
+
+			dirty_zn_cnt = atomic_long_read(&c->dirty_zn_cnt);
+
+			if (!dirty_zn_cnt || c->cmt_state == COMMIT_BROKEN ||
+			    c->ro_mount || c->ro_error) {
+				mutex_unlock(&c->umount_mutex);
+				continue;
+			}
+
+			if (c->cmt_state != COMMIT_RESTING) {
+				spin_unlock(&ubifs_infos_lock);
+				mutex_unlock(&c->umount_mutex);
+				return -1;
+			}
+
+			if (i == 1) {
+				list_move_tail(&c->infos_list, &ubifs_infos);
+				spin_unlock(&ubifs_infos_lock);
+
+				ubifs_request_bg_commit(c);
+				mutex_unlock(&c->umount_mutex);
+				return -1;
+			}
+			mutex_unlock(&c->umount_mutex);
+		}
+	}
+	spin_unlock(&ubifs_infos_lock);
+
+	return 0;
+}
+
+unsigned long ubifs_shrink_count(struct shrinker *shrink,
+				 struct shrink_control *sc)
+{
+	long clean_zn_cnt = atomic_long_read(&ubifs_clean_zn_cnt);
+
+	/*
+	 * Due to the way UBIFS updates the clean znode counter it may
+	 * temporarily be negative.
+	 */
+	return clean_zn_cnt >= 0 ? clean_zn_cnt : 1;
+}
+
+unsigned long ubifs_shrink_scan(struct shrinker *shrink,
+				struct shrink_control *sc)
+{
+	unsigned long nr = sc->nr_to_scan;
+	int contention = 0;
+	unsigned long freed;
+	long clean_zn_cnt = atomic_long_read(&ubifs_clean_zn_cnt);
+
+	if (!clean_zn_cnt) {
+		/*
+		 * No clean znodes, nothing to reap. All we can do in this case
+		 * is to kick background threads to start commit, which will
+		 * probably make clean znodes which, in turn, will be freeable.
+		 * And we return -1 which means will make VM call us again
+		 * later.
+		 */
+		dbg_tnc("no clean znodes, kick a thread");
+		return kick_a_thread();
+	}
+
+	freed = shrink_tnc_trees(nr, OLD_ZNODE_AGE, &contention);
+	if (freed >= nr)
+		goto out;
+
+	dbg_tnc("not enough old znodes, try to free young ones");
+	freed += shrink_tnc_trees(nr - freed, YOUNG_ZNODE_AGE, &contention);
+	if (freed >= nr)
+		goto out;
+
+	dbg_tnc("not enough young znodes, free all");
+	freed += shrink_tnc_trees(nr - freed, 0, &contention);
+
+	if (!freed && contention) {
+		dbg_tnc("freed nothing, but contention");
+		return SHRINK_STOP;
+	}
+
+out:
+	dbg_tnc("%lu znodes were freed, requested %lu", freed, nr);
+	return freed;
+}

linux-7.2.3/fs/ubifs/sysfs.c

@@ -0,0 +1,156 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2021 Cisco Systems
+ *
+ * Author: Stefan Schaeckeler
+ */
+
+
+#include <linux/fs.h>
+#include "ubifs.h"
+
+enum attr_id_t {
+	attr_errors_magic,
+	attr_errors_node,
+	attr_errors_crc,
+};
+
+struct ubifs_attr {
+	struct attribute attr;
+	enum attr_id_t attr_id;
+};
+
+#define UBIFS_ATTR(_name, _mode, _id)					\
+static struct ubifs_attr ubifs_attr_##_name = {				\
+	.attr = {.name = __stringify(_name), .mode = _mode },		\
+	.attr_id = attr_##_id,						\
+}
+
+#define UBIFS_ATTR_FUNC(_name, _mode) UBIFS_ATTR(_name, _mode, _name)
+
+UBIFS_ATTR_FUNC(errors_magic, 0444);
+UBIFS_ATTR_FUNC(errors_crc, 0444);
+UBIFS_ATTR_FUNC(errors_node, 0444);
+
+#define ATTR_LIST(name) (&ubifs_attr_##name.attr)
+
+static struct attribute *ubifs_attrs[] = {
+	ATTR_LIST(errors_magic),
+	ATTR_LIST(errors_node),
+	ATTR_LIST(errors_crc),
+	NULL,
+};
+ATTRIBUTE_GROUPS(ubifs);
+
+static ssize_t ubifs_attr_show(struct kobject *kobj,
+			       struct attribute *attr, char *buf)
+{
+	struct ubifs_info *sbi = container_of(kobj, struct ubifs_info,
+					      kobj);
+
+	struct ubifs_attr *a = container_of(attr, struct ubifs_attr, attr);
+
+	switch (a->attr_id) {
+	case attr_errors_magic:
+		return sysfs_emit(buf, "%u\n", sbi->stats->magic_errors);
+	case attr_errors_node:
+		return sysfs_emit(buf, "%u\n", sbi->stats->node_errors);
+	case attr_errors_crc:
+		return sysfs_emit(buf, "%u\n", sbi->stats->crc_errors);
+	}
+	return 0;
+};
+
+static void ubifs_sb_release(struct kobject *kobj)
+{
+	struct ubifs_info *c = container_of(kobj, struct ubifs_info, kobj);
+
+	complete(&c->kobj_unregister);
+}
+
+static const struct sysfs_ops ubifs_attr_ops = {
+	.show	= ubifs_attr_show,
+};
+
+static const struct kobj_type ubifs_sb_ktype = {
+	.default_groups	= ubifs_groups,
+	.sysfs_ops	= &ubifs_attr_ops,
+	.release	= ubifs_sb_release,
+};
+
+static const struct kobj_type ubifs_ktype = {
+	.sysfs_ops	= &ubifs_attr_ops,
+};
+
+static struct kset ubifs_kset = {
+	.kobj	= {.ktype = &ubifs_ktype},
+};
+
+int ubifs_sysfs_register(struct ubifs_info *c)
+{
+	int ret, n;
+	char dfs_dir_name[UBIFS_DFS_DIR_LEN];
+
+	c->stats = kzalloc_obj(struct ubifs_stats_info);
+	if (!c->stats) {
+		ret = -ENOMEM;
+		goto out_last;
+	}
+	n = snprintf(dfs_dir_name, UBIFS_DFS_DIR_LEN, UBIFS_DFS_DIR_NAME,
+		     c->vi.ubi_num, c->vi.vol_id);
+
+	if (n >= UBIFS_DFS_DIR_LEN) {
+		/* The array size is too small */
+		ret = -EINVAL;
+		goto out_free;
+	}
+
+	c->kobj.kset = &ubifs_kset;
+	init_completion(&c->kobj_unregister);
+
+	ret = kobject_init_and_add(&c->kobj, &ubifs_sb_ktype, NULL,
+				   "%s", dfs_dir_name);
+	if (ret)
+		goto out_put;
+
+	return 0;
+
+out_put:
+	kobject_put(&c->kobj);
+	wait_for_completion(&c->kobj_unregister);
+out_free:
+	kfree(c->stats);
+out_last:
+	ubifs_err(c, "cannot create sysfs entry for ubifs%d_%d, error %d\n",
+		  c->vi.ubi_num, c->vi.vol_id, ret);
+	return ret;
+}
+
+void ubifs_sysfs_unregister(struct ubifs_info *c)
+{
+	kobject_del(&c->kobj);
+	kobject_put(&c->kobj);
+	wait_for_completion(&c->kobj_unregister);
+
+	kfree(c->stats);
+}
+
+int __init ubifs_sysfs_init(void)
+{
+	int ret;
+
+	kobject_set_name(&ubifs_kset.kobj, "ubifs");
+	ubifs_kset.kobj.parent = fs_kobj;
+	ret = kset_register(&ubifs_kset);
+	if (ret)
+		kset_put(&ubifs_kset);
+
+	return ret;
+}
+
+void ubifs_sysfs_exit(void)
+{
+	kset_unregister(&ubifs_kset);
+}

linux-7.2.3/fs/ubifs/tnc_misc.c

@@ -0,0 +1,517 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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.
+ * @c: UBIFS file-system description object
+ * @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(const struct ubifs_info *c,
+					      struct ubifs_znode *zr,
+					      struct ubifs_znode *znode)
+{
+	int level, iip, level_search = 0;
+	struct ubifs_znode *zn;
+
+	ubifs_assert(c, 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(c, 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(c, 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(c, 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, mid;
+	int cmp;
+	const struct ubifs_zbranch *zbr = &znode->zbranch[0];
+
+	ubifs_assert(c, 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(c, *n >= -1 && *n < znode->child_cnt);
+	if (*n == -1)
+		ubifs_assert(c, keys_cmp(c, key, &zbr[0].key) < 0);
+	else
+		ubifs_assert(c, keys_cmp(c, key, &zbr[*n].key) > 0);
+	if (*n + 1 < znode->child_cnt)
+		ubifs_assert(c, 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.
+ * @c: UBIFS file-system description object
+ * @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(const struct ubifs_info *c,
+					     struct ubifs_znode *znode)
+{
+	struct ubifs_znode *zn;
+
+	ubifs_assert(c, 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);
+}
+
+/**
+ * ubifs_destroy_tnc_subtree - destroy all znodes connected to a subtree.
+ * @c: UBIFS file-system description object
+ * @znode: znode defining subtree to destroy
+ *
+ * This function destroys subtree of the TNC tree. Returns number of clean
+ * znodes in the subtree.
+ */
+long ubifs_destroy_tnc_subtree(const struct ubifs_info *c,
+			       struct ubifs_znode *znode)
+{
+	struct ubifs_znode *zn = ubifs_tnc_postorder_first(znode);
+	long clean_freed = 0;
+	int n;
+
+	ubifs_assert(c, zn);
+	while (1) {
+		for (n = 0; n < zn->child_cnt; n++) {
+			if (!zn->zbranch[n].znode)
+				continue;
+
+			if (zn->level > 0 &&
+			    !ubifs_zn_dirty(zn->zbranch[n].znode))
+				clean_freed += 1;
+
+			cond_resched();
+			kfree(zn->zbranch[n].znode);
+		}
+
+		if (zn == znode) {
+			if (!ubifs_zn_dirty(zn))
+				clean_freed += 1;
+			kfree(zn);
+			return clean_freed;
+		}
+
+		zn = ubifs_tnc_postorder_next(c, zn);
+	}
+}
+
+/**
+ * ubifs_destroy_tnc_tree - destroy all znodes connected to the TNC tree.
+ * @c: UBIFS file-system description object
+ *
+ * This function destroys the whole TNC tree and updates clean global znode
+ * count.
+ */
+void ubifs_destroy_tnc_tree(struct ubifs_info *c)
+{
+	long n, freed;
+
+	if (!c->zroot.znode)
+		return;
+
+	n = atomic_long_read(&c->clean_zn_cnt);
+	freed = ubifs_destroy_tnc_subtree(c, c->zroot.znode);
+	ubifs_assert(c, freed == n);
+	atomic_long_sub(n, &ubifs_clean_zn_cnt);
+
+	c->zroot.znode = NULL;
+}
+
+/**
+ * read_znode - read an indexing node from flash and fill znode.
+ * @c: UBIFS file-system description object
+ * @zzbr: the zbranch describing the node to read
+ * @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, struct ubifs_zbranch *zzbr,
+		      struct ubifs_znode *znode)
+{
+	int lnum = zzbr->lnum;
+	int offs = zzbr->offs;
+	int len = zzbr->len;
+	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;
+	}
+
+	err = ubifs_node_check_hash(c, idx, zzbr->hash);
+	if (err) {
+		ubifs_bad_hash(c, idx, zzbr->hash, lnum, offs);
+		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) {
+		ubifs_err(c, "current fanout %d, branch count %d",
+			  c->fanout, znode->child_cnt);
+		ubifs_err(c, "max levels %d, znode level %d",
+			  UBIFS_MAX_LEVELS, znode->level);
+		goto out_dump;
+	}
+
+	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_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);
+		ubifs_copy_hash(c, ubifs_branch_hash(c, br), zbr->hash);
+		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) {
+			ubifs_err(c, "bad branch %d", i);
+			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:
+			ubifs_err(c, "bad key type at slot %d: %d",
+				  i, key_type(c, &zbr->key));
+			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) {
+				ubifs_err(c, "bad target node (type %d) length (%d)",
+					  type, zbr->len);
+				ubifs_err(c, "have to be %d", c->ranges[type].len);
+				goto out_dump;
+			}
+		} else if (zbr->len < c->ranges[type].min_len ||
+			   zbr->len > c->ranges[type].max_len) {
+			ubifs_err(c, "bad target node (type %d) length (%d)",
+				  type, zbr->len);
+			ubifs_err(c, "have to be in range of %d-%d",
+				  c->ranges[type].min_len,
+				  c->ranges[type].max_len);
+			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) {
+			ubifs_err(c, "bad key order (keys %d and %d)", i, i + 1);
+			goto out_dump;
+		} else if (cmp == 0 && !is_hash_key(c, key1)) {
+			/* These can only be keys with colliding hash */
+			ubifs_err(c, "keys %d and %d are not hashed but equivalent",
+				  i, i + 1);
+			goto out_dump;
+		}
+	}
+
+	kfree(idx);
+	return 0;
+
+out_dump:
+	ubifs_err(c, "bad indexing node at LEB %d:%d", lnum, offs);
+	ubifs_dump_node(c, idx, c->max_idx_node_sz);
+	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(c, !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, znode);
+	if (err)
+		goto out;
+
+	atomic_long_inc(&c->clean_zn_cnt);
+
+	/*
+	 * Increment the global clean znode counter as well. It is OK that
+	 * global and per-FS clean znode counters may be inconsistent for some
+	 * short time (because we might be preempted at this point), the global
+	 * one is only used in shrinker.
+	 */
+	atomic_long_inc(&ubifs_clean_zn_cnt);
+
+	zbr->znode = znode;
+	znode->parent = parent;
+	znode->time = ktime_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 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);
+	struct ubifs_wbuf *wbuf;
+
+	/*
+	 * 'zbr' has to point to on-flash node. The node may sit in a bud and
+	 * may even be in a write buffer, so we have to take care about this.
+	 */
+	wbuf = ubifs_get_wbuf(c, zbr->lnum);
+	if (wbuf)
+		err = ubifs_read_node_wbuf(wbuf, node, type, zbr->len,
+					   zbr->lnum, zbr->offs);
+	else
+		err = ubifs_read_node(c, node, type, zbr->len, zbr->lnum,
+				      zbr->offs);
+
+	if (err) {
+		dbg_tnck(key, "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(c, "bad key in node at LEB %d:%d",
+			  zbr->lnum, zbr->offs);
+		dbg_tnck(key, "looked for key ");
+		dbg_tnck(&key1, "but found node's key ");
+		ubifs_dump_node(c, node, zbr->len);
+		return -EINVAL;
+	}
+
+	err = ubifs_node_check_hash(c, node, zbr->hash);
+	if (err) {
+		ubifs_bad_hash(c, node, zbr->hash, zbr->lnum, zbr->offs);
+		return err;
+	}
+
+	return 0;
+}

linux-7.2.3/fs/ubifs/ubifs-media.h

@@ -0,0 +1,855 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * 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. This version is increased when the on-flash
+ * format is changing. If this happens, UBIFS is will support older versions as
+ * well. But older UBIFS code will not support newer formats. Format changes
+ * will be rare and only when absolutely necessary, e.g. to fix a bug or to add
+ * a new feature.
+ *
+ * UBIFS went into mainline kernel with format version 4. The older formats
+ * were development formats.
+ */
+#define UBIFS_FORMAT_VERSION 5
+
+/*
+ * Read-only compatibility version. If the UBIFS format is changed, older UBIFS
+ * implementations will not be able to mount newer formats in read-write mode.
+ * However, depending on the change, it may be possible to mount newer formats
+ * in R/O mode. This is indicated by the R/O compatibility version which is
+ * stored in the super-block.
+ *
+ * This is needed to support boot-loaders which only need R/O mounting. With
+ * this flag it is possible to do UBIFS format changes without a need to update
+ * boot-loaders.
+ */
+#define UBIFS_RO_COMPAT_VERSION 0
+
+/* 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 prefers 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)
+
+/* Garbage collector journal head number */
+#define UBIFS_GC_HEAD   0
+/* Base journal head number */
+#define UBIFS_BASE_HEAD 1
+/* Data journal head number */
+#define UBIFS_DATA_HEAD 2
+
+/*
+ * 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)
+#define UBIFS_AUTH_NODE_SZ sizeof(struct ubifs_auth_node)
+#define UBIFS_SIG_NODE_SZ  sizeof(struct ubifs_sig_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
+
+/* The maxmimum size of a hash, enough for sha512 */
+#define UBIFS_MAX_HASH_LEN 64
+
+/* The maxmimum size of a hmac, enough for hmac(sha512) */
+#define UBIFS_MAX_HMAC_LEN 64
+
+/*
+ * xattr name of UBIFS encryption context, we don't use a prefix
+ * nor a long name to not waste space on the flash.
+ */
+#define UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT "c"
+
+/* Type field in ubifs_sig_node */
+#define UBIFS_SIGNATURE_TYPE_PKCS7	1
+
+/*
+ * 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
+ * UBIFS_CRYPT_FL: use encryption for this inode
+ *
+ * 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,
+	UBIFS_CRYPT_FL     = 0x40,
+};
+
+/* 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_ZSTD: ZSTD compression
+ * UBIFS_COMPR_TYPES_CNT: count of supported compression types
+ */
+enum {
+	UBIFS_COMPR_NONE,
+	UBIFS_COMPR_LZO,
+	UBIFS_COMPR_ZLIB,
+	UBIFS_COMPR_ZSTD,
+	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_AUTH_NODE: authentication node
+ * UBIFS_SIG_NODE: signature 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_AUTH_NODE,
+	UBIFS_SIG_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
+ * UBIFS_FLG_SPACE_FIXUP: first-mount "fixup" of free space within LEBs needed
+ * UBIFS_FLG_DOUBLE_HASH: store a 32bit cookie in directory entry nodes to
+ *			  support 64bit cookies for lookups by hash
+ * UBIFS_FLG_ENCRYPTION: this filesystem contains encrypted files
+ * UBIFS_FLG_AUTHENTICATION: this filesystem contains hashes for authentication
+ */
+enum {
+	UBIFS_FLG_BIGLPT = 0x02,
+	UBIFS_FLG_SPACE_FIXUP = 0x04,
+	UBIFS_FLG_DOUBLE_HASH = 0x08,
+	UBIFS_FLG_ENCRYPTION = 0x10,
+	UBIFS_FLG_AUTHENTICATION = 0x20,
+};
+
+#define UBIFS_FLG_MASK (UBIFS_FLG_BIGLPT | UBIFS_FLG_SPACE_FIXUP | \
+		UBIFS_FLG_DOUBLE_HASH | UBIFS_FLG_ENCRYPTION | \
+		UBIFS_FLG_AUTHENTICATION)
+
+/**
+ * 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];
+} __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;
+} __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[];
+} __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
+ * @cookie: A 32bits random number, used to construct a 64bits
+ *          identifier.
+ * @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;
+	__le32 cookie;
+	__u8 name[];
+} __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)
+ * @compr_size: compressed data size in bytes, only valid when data is encrypted
+ * @data: data
+ *
+ */
+struct ubifs_data_node {
+	struct ubifs_ch ch;
+	__u8 key[UBIFS_MAX_KEY_LEN];
+	__le32 size;
+	__le16 compr_type;
+	__le16 compr_size;
+	__u8 data[];
+} __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;
+} __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;
+} __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
+ * @ro_compat_version: UBIFS R/O compatibility version
+ * @hmac: HMAC to authenticate the superblock node
+ * @hmac_wkm: HMAC of a well known message (the string "UBIFS") as a convenience
+ *            to the user to check if the correct key is passed.
+ * @hash_algo: The hash algo used for this filesystem (one of enum hash_algo)
+ * @hash_mst: hash of the master node, only valid for signed images in which the
+ *            master node does not contain a hmac
+ */
+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];
+	__le32 ro_compat_version;
+	__u8 hmac[UBIFS_MAX_HMAC_LEN];
+	__u8 hmac_wkm[UBIFS_MAX_HMAC_LEN];
+	__le16 hash_algo;
+	__u8 hash_mst[UBIFS_MAX_HASH_LEN];
+	__u8 padding2[3774];
+} __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
+ * @hash_root_idx: the hash of the root index node
+ * @hash_lpt: the hash of the LPT
+ * @hmac: HMAC to authenticate the master node
+ * @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 hash_root_idx[UBIFS_MAX_HASH_LEN];
+	__u8 hash_lpt[UBIFS_MAX_HASH_LEN];
+	__u8 hmac[UBIFS_MAX_HMAC_LEN];
+	__u8 padding[152];
+} __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];
+} __packed;
+
+/**
+ * struct ubifs_auth_node - node for authenticating other nodes
+ * @ch: common header
+ * @hmac: The HMAC
+ */
+struct ubifs_auth_node {
+	struct ubifs_ch ch;
+	__u8 hmac[];
+} __packed;
+
+/**
+ * struct ubifs_sig_node - node for signing other nodes
+ * @ch: common header
+ * @type: type of the signature, currently only UBIFS_SIGNATURE_TYPE_PKCS7
+ * supported
+ * @len: The length of the signature data
+ * @padding: reserved for future, zeroes
+ * @sig: The signature data
+ */
+struct ubifs_sig_node {
+	struct ubifs_ch ch;
+	__le32 type;
+	__le32 len;
+	__u8 padding[32];
+	__u8 sig[];
+} __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
+ *
+ * In an authenticated UBIFS we have the hash of the referenced node after @key.
+ * This can't be added to the struct type definition because @key is a
+ * dynamically sized element already.
+ */
+struct ubifs_branch {
+	__le32 lnum;
+	__le32 offs;
+	__le32 len;
+	__u8 key[];
+} __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[];
+} __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;
+} __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[];
+} __packed;
+
+#endif /* __UBIFS_MEDIA_H__ */

linux-7.2.3/fs/ubifs/xattr.c

@@ -0,0 +1,706 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * This file is part of UBIFS.
+ *
+ * Copyright (C) 2006-2008 Nokia Corporation.
+ *
+ * Authors: Artem Bityutskiy (Битюцкий Артём)
+ *          Adrian Hunter
+ */
+
+/*
+ * This file implements UBIFS extended attributes support.
+ *
+ * Extended attributes are implemented as regular inodes with attached data,
+ * which limits extended attribute size to UBIFS block size (4KiB). Names of
+ * extended attributes are described by extended attribute entries (xentries),
+ * which are almost identical to directory entries, but have different key type.
+ *
+ * In other words, the situation with extended attributes is very similar to
+ * directories. Indeed, any inode (but of course not xattr inodes) may have a
+ * number of associated xentries, just like directory inodes have associated
+ * directory entries. Extended attribute entries store the name of the extended
+ * attribute, the host inode number, and the extended attribute inode number.
+ * Similarly, direntries store the name, the parent and the target inode
+ * numbers. Thus, most of the common UBIFS mechanisms may be re-used for
+ * extended attributes.
+ *
+ * The number of extended attributes is not limited, but there is Linux
+ * limitation on the maximum possible size of the list of all extended
+ * attributes associated with an inode (%XATTR_LIST_MAX), so UBIFS makes sure
+ * the sum of all extended attribute names of the inode does not exceed that
+ * limit.
+ *
+ * Extended attributes are synchronous, which means they are written to the
+ * flash media synchronously and there is no write-back for extended attribute
+ * inodes. The extended attribute values are not stored in compressed form on
+ * the media.
+ *
+ * Since extended attributes are represented by regular inodes, they are cached
+ * in the VFS inode cache. The xentries are cached in the LNC cache (see
+ * tnc.c).
+ *
+ * ACL support is not implemented.
+ */
+
+#include "ubifs.h"
+#include <linux/fs.h>
+#include <linux/slab.h>
+#include <linux/xattr.h>
+
+static const struct inode_operations empty_iops;
+static const struct file_operations empty_fops;
+
+/**
+ * create_xattr - create an extended attribute.
+ * @c: UBIFS file-system description object
+ * @host: host inode
+ * @nm: extended attribute name
+ * @value: extended attribute value
+ * @size: size of extended attribute value
+ *
+ * This is a helper function which creates an extended attribute of name @nm
+ * and value @value for inode @host. The host inode is also updated on flash
+ * because the ctime and extended attribute accounting data changes. This
+ * function returns zero in case of success and a negative error code in case
+ * of failure.
+ */
+static int create_xattr(struct ubifs_info *c, struct inode *host,
+			const struct fscrypt_name *nm, const void *value, int size)
+{
+	int err, names_len;
+	struct inode *inode;
+	struct ubifs_inode *ui, *host_ui = ubifs_inode(host);
+	struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+				.new_ino_d = ALIGN(size, 8), .dirtied_ino = 1,
+				.dirtied_ino_d = ALIGN(host_ui->data_len, 8) };
+
+	if (host_ui->xattr_cnt >= ubifs_xattr_max_cnt(c)) {
+		ubifs_err(c, "inode %llu already has too many xattrs (%d), cannot create more",
+			  host->i_ino, host_ui->xattr_cnt);
+		return -ENOSPC;
+	}
+	/*
+	 * Linux limits the maximum size of the extended attribute names list
+	 * to %XATTR_LIST_MAX. This means we should not allow creating more
+	 * extended attributes if the name list becomes larger. This limitation
+	 * is artificial for UBIFS, though.
+	 */
+	names_len = host_ui->xattr_names + host_ui->xattr_cnt + fname_len(nm) + 1;
+	if (names_len > XATTR_LIST_MAX) {
+		ubifs_err(c, "cannot add one more xattr name to inode %llu, total names length would become %d, max. is %d",
+			  host->i_ino, names_len, XATTR_LIST_MAX);
+		return -ENOSPC;
+	}
+
+	err = ubifs_budget_space(c, &req);
+	if (err)
+		return err;
+
+	inode = ubifs_new_inode(c, host, S_IFREG | S_IRWXUGO, true);
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out_budg;
+	}
+
+	/* Re-define all operations to be "nothing" */
+	inode->i_mapping->a_ops = &empty_aops;
+	inode->i_op = &empty_iops;
+	inode->i_fop = &empty_fops;
+
+	inode->i_flags |= S_SYNC | S_NOATIME | S_NOCMTIME;
+	ui = ubifs_inode(inode);
+	ui->xattr = 1;
+	ui->flags |= UBIFS_XATTR_FL;
+	ui->data = kmemdup(value, size, GFP_NOFS);
+	if (!ui->data) {
+		err = -ENOMEM;
+		goto out_free;
+	}
+	inode->i_size = ui->ui_size = size;
+	ui->data_len = size;
+
+	mutex_lock(&host_ui->ui_mutex);
+	inode_set_ctime_current(host);
+	host_ui->xattr_cnt += 1;
+	host_ui->xattr_size += CALC_DENT_SIZE(fname_len(nm));
+	host_ui->xattr_size += CALC_XATTR_BYTES(size);
+	host_ui->xattr_names += fname_len(nm);
+
+	/*
+	 * We handle UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT here because we
+	 * have to set the UBIFS_CRYPT_FL flag on the host inode.
+	 * To avoid multiple updates of the same inode in the same operation,
+	 * let's do it here.
+	 */
+	if (strcmp(fname_name(nm), UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT) == 0)
+		host_ui->flags |= UBIFS_CRYPT_FL;
+
+	err = ubifs_jnl_update(c, host, nm, inode, 0, 1, 0);
+	if (err)
+		goto out_cancel;
+	ubifs_set_inode_flags(host);
+	mutex_unlock(&host_ui->ui_mutex);
+
+	ubifs_release_budget(c, &req);
+	insert_inode_hash(inode);
+	iput(inode);
+	return 0;
+
+out_cancel:
+	host_ui->xattr_cnt -= 1;
+	host_ui->xattr_size -= CALC_DENT_SIZE(fname_len(nm));
+	host_ui->xattr_size -= CALC_XATTR_BYTES(size);
+	host_ui->xattr_names -= fname_len(nm);
+	host_ui->flags &= ~UBIFS_CRYPT_FL;
+	mutex_unlock(&host_ui->ui_mutex);
+out_free:
+	make_bad_inode(inode);
+	iput(inode);
+out_budg:
+	ubifs_release_budget(c, &req);
+	return err;
+}
+
+/**
+ * change_xattr - change an extended attribute.
+ * @c: UBIFS file-system description object
+ * @host: host inode
+ * @inode: extended attribute inode
+ * @value: extended attribute value
+ * @size: size of extended attribute value
+ *
+ * This helper function changes the value of extended attribute @inode with new
+ * data from @value. Returns zero in case of success and a negative error code
+ * in case of failure.
+ */
+static int change_xattr(struct ubifs_info *c, struct inode *host,
+			struct inode *inode, const void *value, int size)
+{
+	int err;
+	struct ubifs_inode *host_ui = ubifs_inode(host);
+	struct ubifs_inode *ui = ubifs_inode(inode);
+	void *buf = NULL;
+	int old_size;
+	struct ubifs_budget_req req = { .dirtied_ino = 2,
+		.dirtied_ino_d = ALIGN(size, 8) + ALIGN(host_ui->data_len, 8) };
+
+	ubifs_assert(c, ui->data_len == inode->i_size);
+	err = ubifs_budget_space(c, &req);
+	if (err)
+		return err;
+
+	buf = kmemdup(value, size, GFP_NOFS);
+	if (!buf) {
+		err = -ENOMEM;
+		goto out_free;
+	}
+	kfree(ui->data);
+	ui->data = buf;
+	inode->i_size = ui->ui_size = size;
+	old_size = ui->data_len;
+	ui->data_len = size;
+
+	mutex_lock(&host_ui->ui_mutex);
+	inode_set_ctime_current(host);
+	host_ui->xattr_size -= CALC_XATTR_BYTES(old_size);
+	host_ui->xattr_size += CALC_XATTR_BYTES(size);
+
+	/*
+	 * It is important to write the host inode after the xattr inode
+	 * because if the host inode gets synchronized (via 'fsync()'), then
+	 * the extended attribute inode gets synchronized, because it goes
+	 * before the host inode in the write-buffer.
+	 */
+	err = ubifs_jnl_change_xattr(c, inode, host);
+	if (err)
+		goto out_cancel;
+	mutex_unlock(&host_ui->ui_mutex);
+
+	ubifs_release_budget(c, &req);
+	return 0;
+
+out_cancel:
+	host_ui->xattr_size -= CALC_XATTR_BYTES(size);
+	host_ui->xattr_size += CALC_XATTR_BYTES(old_size);
+	mutex_unlock(&host_ui->ui_mutex);
+	make_bad_inode(inode);
+out_free:
+	ubifs_release_budget(c, &req);
+	return err;
+}
+
+static struct inode *iget_xattr(struct ubifs_info *c, ino_t inum)
+{
+	struct inode *inode;
+
+	inode = ubifs_iget(c->vfs_sb, inum);
+	if (IS_ERR(inode)) {
+		ubifs_err(c, "dead extended attribute entry, error %d",
+			  (int)PTR_ERR(inode));
+		return inode;
+	}
+	if (ubifs_inode(inode)->xattr)
+		return inode;
+	ubifs_err(c, "corrupt extended attribute entry");
+	iput(inode);
+	return ERR_PTR(-EINVAL);
+}
+
+int ubifs_xattr_set(struct inode *host, const char *name, const void *value,
+		    size_t size, int flags, bool check_lock)
+{
+	struct inode *inode;
+	struct ubifs_info *c = host->i_sb->s_fs_info;
+	struct fscrypt_name nm = { .disk_name = FSTR_INIT((char *)name, strlen(name))};
+	struct ubifs_dent_node *xent;
+	union ubifs_key key;
+	int err;
+
+	if (check_lock)
+		ubifs_assert(c, inode_is_locked(host));
+
+	if (size > UBIFS_MAX_INO_DATA)
+		return -ERANGE;
+
+	if (fname_len(&nm) > UBIFS_MAX_NLEN)
+		return -ENAMETOOLONG;
+
+	xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS);
+	if (!xent)
+		return -ENOMEM;
+
+	down_write(&ubifs_inode(host)->xattr_sem);
+	/*
+	 * The extended attribute entries are stored in LNC, so multiple
+	 * look-ups do not involve reading the flash.
+	 */
+	xent_key_init(c, &key, host->i_ino, &nm);
+	err = ubifs_tnc_lookup_nm(c, &key, xent, &nm);
+	if (err) {
+		if (err != -ENOENT)
+			goto out_free;
+
+		if (flags & XATTR_REPLACE)
+			/* We are asked not to create the xattr */
+			err = -ENODATA;
+		else
+			err = create_xattr(c, host, &nm, value, size);
+		goto out_free;
+	}
+
+	if (flags & XATTR_CREATE) {
+		/* We are asked not to replace the xattr */
+		err = -EEXIST;
+		goto out_free;
+	}
+
+	inode = iget_xattr(c, le64_to_cpu(xent->inum));
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out_free;
+	}
+
+	err = change_xattr(c, host, inode, value, size);
+	iput(inode);
+
+out_free:
+	up_write(&ubifs_inode(host)->xattr_sem);
+	kfree(xent);
+	return err;
+}
+
+ssize_t ubifs_xattr_get(struct inode *host, const char *name, void *buf,
+			size_t size)
+{
+	struct inode *inode;
+	struct ubifs_info *c = host->i_sb->s_fs_info;
+	struct fscrypt_name nm = { .disk_name = FSTR_INIT((char *)name, strlen(name))};
+	struct ubifs_inode *ui;
+	struct ubifs_dent_node *xent;
+	union ubifs_key key;
+	int err;
+
+	if (fname_len(&nm) > UBIFS_MAX_NLEN)
+		return -ENAMETOOLONG;
+
+	xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS);
+	if (!xent)
+		return -ENOMEM;
+
+	down_read(&ubifs_inode(host)->xattr_sem);
+	xent_key_init(c, &key, host->i_ino, &nm);
+	err = ubifs_tnc_lookup_nm(c, &key, xent, &nm);
+	if (err) {
+		if (err == -ENOENT)
+			err = -ENODATA;
+		goto out_cleanup;
+	}
+
+	inode = iget_xattr(c, le64_to_cpu(xent->inum));
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out_cleanup;
+	}
+
+	ui = ubifs_inode(inode);
+	ubifs_assert(c, inode->i_size == ui->data_len);
+	ubifs_assert(c, ubifs_inode(host)->xattr_size > ui->data_len);
+
+	if (buf) {
+		/* If @buf is %NULL we are supposed to return the length */
+		if (ui->data_len > size) {
+			err = -ERANGE;
+			goto out_iput;
+		}
+
+		memcpy(buf, ui->data, ui->data_len);
+	}
+	err = ui->data_len;
+
+out_iput:
+	iput(inode);
+out_cleanup:
+	up_read(&ubifs_inode(host)->xattr_sem);
+	kfree(xent);
+	return err;
+}
+
+static bool xattr_visible(const char *name)
+{
+	/* File encryption related xattrs are for internal use only */
+	if (strcmp(name, UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT) == 0)
+		return false;
+
+	/* Show trusted namespace only for "power" users */
+	if (strncmp(name, XATTR_TRUSTED_PREFIX,
+		    XATTR_TRUSTED_PREFIX_LEN) == 0 && !capable(CAP_SYS_ADMIN))
+		return false;
+
+	return true;
+}
+
+ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size)
+{
+	union ubifs_key key;
+	struct inode *host = d_inode(dentry);
+	struct ubifs_info *c = host->i_sb->s_fs_info;
+	struct ubifs_inode *host_ui = ubifs_inode(host);
+	struct ubifs_dent_node *xent, *pxent = NULL;
+	int err, len, written = 0;
+	struct fscrypt_name nm = {0};
+
+	dbg_gen("ino %llu ('%pd'), buffer size %zd", host->i_ino,
+		dentry, size);
+
+	down_read(&host_ui->xattr_sem);
+	len = host_ui->xattr_names + host_ui->xattr_cnt;
+	if (!buffer) {
+		/*
+		 * We should return the minimum buffer size which will fit a
+		 * null-terminated list of all the extended attribute names.
+		 */
+		err = len;
+		goto out_err;
+	}
+
+	if (len > size) {
+		err = -ERANGE;
+		goto out_err;
+	}
+
+	lowest_xent_key(c, &key, host->i_ino);
+	while (1) {
+		xent = ubifs_tnc_next_ent(c, &key, &nm);
+		if (IS_ERR(xent)) {
+			err = PTR_ERR(xent);
+			break;
+		}
+
+		fname_name(&nm) = xent->name;
+		fname_len(&nm) = le16_to_cpu(xent->nlen);
+
+		if (xattr_visible(xent->name)) {
+			memcpy(buffer + written, fname_name(&nm), fname_len(&nm) + 1);
+			written += fname_len(&nm) + 1;
+		}
+
+		kfree(pxent);
+		pxent = xent;
+		key_read(c, &xent->key, &key);
+	}
+	kfree(pxent);
+	up_read(&host_ui->xattr_sem);
+
+	if (err != -ENOENT) {
+		ubifs_err(c, "cannot find next direntry, error %d", err);
+		return err;
+	}
+
+	ubifs_assert(c, written <= size);
+	return written;
+
+out_err:
+	up_read(&host_ui->xattr_sem);
+	return err;
+}
+
+static int remove_xattr(struct ubifs_info *c, struct inode *host,
+			struct inode *inode, const struct fscrypt_name *nm)
+{
+	int err;
+	struct ubifs_inode *host_ui = ubifs_inode(host);
+	struct ubifs_inode *ui = ubifs_inode(inode);
+	struct ubifs_budget_req req = { .dirtied_ino = 2, .mod_dent = 1,
+				.dirtied_ino_d = ALIGN(host_ui->data_len, 8) };
+
+	ubifs_assert(c, ui->data_len == inode->i_size);
+
+	err = ubifs_budget_space(c, &req);
+	if (err)
+		return err;
+
+	mutex_lock(&host_ui->ui_mutex);
+	inode_set_ctime_current(host);
+	host_ui->xattr_cnt -= 1;
+	host_ui->xattr_size -= CALC_DENT_SIZE(fname_len(nm));
+	host_ui->xattr_size -= CALC_XATTR_BYTES(ui->data_len);
+	host_ui->xattr_names -= fname_len(nm);
+
+	err = ubifs_jnl_delete_xattr(c, host, inode, nm);
+	if (err)
+		goto out_cancel;
+	mutex_unlock(&host_ui->ui_mutex);
+
+	ubifs_release_budget(c, &req);
+	return 0;
+
+out_cancel:
+	host_ui->xattr_cnt += 1;
+	host_ui->xattr_size += CALC_DENT_SIZE(fname_len(nm));
+	host_ui->xattr_size += CALC_XATTR_BYTES(ui->data_len);
+	host_ui->xattr_names += fname_len(nm);
+	mutex_unlock(&host_ui->ui_mutex);
+	ubifs_release_budget(c, &req);
+	make_bad_inode(inode);
+	return err;
+}
+
+int ubifs_purge_xattrs(struct inode *host)
+{
+	union ubifs_key key;
+	struct ubifs_info *c = host->i_sb->s_fs_info;
+	struct ubifs_dent_node *xent, *pxent = NULL;
+	struct inode *xino;
+	struct fscrypt_name nm = {0};
+	int err;
+
+	if (ubifs_inode(host)->xattr_cnt <= ubifs_xattr_max_cnt(c))
+		return 0;
+
+	ubifs_warn(c, "inode %llu has too many xattrs, doing a non-atomic deletion",
+		   host->i_ino);
+
+	down_write(&ubifs_inode(host)->xattr_sem);
+	lowest_xent_key(c, &key, host->i_ino);
+	while (1) {
+		xent = ubifs_tnc_next_ent(c, &key, &nm);
+		if (IS_ERR(xent)) {
+			err = PTR_ERR(xent);
+			break;
+		}
+
+		fname_name(&nm) = xent->name;
+		fname_len(&nm) = le16_to_cpu(xent->nlen);
+
+		xino = ubifs_iget(c->vfs_sb, le64_to_cpu(xent->inum));
+		if (IS_ERR(xino)) {
+			err = PTR_ERR(xino);
+			ubifs_err(c, "dead directory entry '%s', error %d",
+				  xent->name, err);
+			ubifs_ro_mode(c, err);
+			goto out_err;
+		}
+
+		ubifs_assert(c, ubifs_inode(xino)->xattr);
+
+		clear_nlink(xino);
+		err = remove_xattr(c, host, xino, &nm);
+		iput(xino);
+		if (err) {
+			ubifs_err(c, "cannot remove xattr, error %d", err);
+			goto out_err;
+		}
+
+		kfree(pxent);
+		pxent = xent;
+		key_read(c, &xent->key, &key);
+	}
+	kfree(pxent);
+	up_write(&ubifs_inode(host)->xattr_sem);
+
+	if (err != -ENOENT) {
+		ubifs_err(c, "cannot find next direntry, error %d", err);
+		return err;
+	}
+
+	return 0;
+
+out_err:
+	kfree(pxent);
+	kfree(xent);
+	up_write(&ubifs_inode(host)->xattr_sem);
+	return err;
+}
+
+static int ubifs_xattr_remove(struct inode *host, const char *name)
+{
+	struct inode *inode;
+	struct ubifs_info *c = host->i_sb->s_fs_info;
+	struct fscrypt_name nm = { .disk_name = FSTR_INIT((char *)name, strlen(name))};
+	struct ubifs_dent_node *xent;
+	union ubifs_key key;
+	int err;
+
+	ubifs_assert(c, inode_is_locked(host));
+
+	if (fname_len(&nm) > UBIFS_MAX_NLEN)
+		return -ENAMETOOLONG;
+
+	xent = kmalloc(UBIFS_MAX_XENT_NODE_SZ, GFP_NOFS);
+	if (!xent)
+		return -ENOMEM;
+
+	down_write(&ubifs_inode(host)->xattr_sem);
+	xent_key_init(c, &key, host->i_ino, &nm);
+	err = ubifs_tnc_lookup_nm(c, &key, xent, &nm);
+	if (err) {
+		if (err == -ENOENT)
+			err = -ENODATA;
+		goto out_free;
+	}
+
+	inode = iget_xattr(c, le64_to_cpu(xent->inum));
+	if (IS_ERR(inode)) {
+		err = PTR_ERR(inode);
+		goto out_free;
+	}
+
+	ubifs_assert(c, inode->i_nlink == 1);
+	clear_nlink(inode);
+	err = remove_xattr(c, host, inode, &nm);
+	if (err)
+		set_nlink(inode, 1);
+
+	/* If @i_nlink is 0, 'iput()' will delete the inode */
+	iput(inode);
+
+out_free:
+	up_write(&ubifs_inode(host)->xattr_sem);
+	kfree(xent);
+	return err;
+}
+
+#ifdef CONFIG_UBIFS_FS_SECURITY
+static int init_xattrs(struct inode *inode, const struct xattr *xattr_array,
+		      void *fs_info)
+{
+	const struct xattr *xattr;
+	char *name;
+	int err = 0;
+
+	for (xattr = xattr_array; xattr->name != NULL; xattr++) {
+		name = kmalloc(XATTR_SECURITY_PREFIX_LEN +
+			       strlen(xattr->name) + 1, GFP_NOFS);
+		if (!name) {
+			err = -ENOMEM;
+			break;
+		}
+		strcpy(name, XATTR_SECURITY_PREFIX);
+		strcpy(name + XATTR_SECURITY_PREFIX_LEN, xattr->name);
+		/*
+		 * creating a new inode without holding the inode rwsem,
+		 * no need to check whether inode is locked.
+		 */
+		err = ubifs_xattr_set(inode, name, xattr->value,
+				      xattr->value_len, 0, false);
+		kfree(name);
+		if (err < 0)
+			break;
+	}
+
+	return err;
+}
+
+int ubifs_init_security(struct inode *dentry, struct inode *inode,
+			const struct qstr *qstr)
+{
+	int err;
+
+	err = security_inode_init_security(inode, dentry, qstr,
+					   &init_xattrs, NULL);
+	if (err) {
+		struct ubifs_info *c = dentry->i_sb->s_fs_info;
+		ubifs_err(c, "cannot initialize security for inode %llu, error %d",
+			  inode->i_ino, err);
+	}
+	return err;
+}
+#endif
+
+static int xattr_get(const struct xattr_handler *handler,
+			   struct dentry *dentry, struct inode *inode,
+			   const char *name, void *buffer, size_t size)
+{
+	dbg_gen("xattr '%s', ino %llu ('%pd'), buf size %zd", name,
+		inode->i_ino, dentry, size);
+
+	name = xattr_full_name(handler, name);
+	return ubifs_xattr_get(inode, name, buffer, size);
+}
+
+static int xattr_set(const struct xattr_handler *handler,
+			   struct mnt_idmap *idmap,
+			   struct dentry *dentry, struct inode *inode,
+			   const char *name, const void *value,
+			   size_t size, int flags)
+{
+	dbg_gen("xattr '%s', host ino %llu ('%pd'), size %zd",
+		name, inode->i_ino, dentry, size);
+
+	name = xattr_full_name(handler, name);
+
+	if (value)
+		return ubifs_xattr_set(inode, name, value, size, flags, true);
+	else
+		return ubifs_xattr_remove(inode, name);
+}
+
+static const struct xattr_handler ubifs_user_xattr_handler = {
+	.prefix = XATTR_USER_PREFIX,
+	.get = xattr_get,
+	.set = xattr_set,
+};
+
+static const struct xattr_handler ubifs_trusted_xattr_handler = {
+	.prefix = XATTR_TRUSTED_PREFIX,
+	.get = xattr_get,
+	.set = xattr_set,
+};
+
+#ifdef CONFIG_UBIFS_FS_SECURITY
+static const struct xattr_handler ubifs_security_xattr_handler = {
+	.prefix = XATTR_SECURITY_PREFIX,
+	.get = xattr_get,
+	.set = xattr_set,
+};
+#endif
+
+const struct xattr_handler * const ubifs_xattr_handlers[] = {
+	&ubifs_user_xattr_handler,
+	&ubifs_trusted_xattr_handler,
+#ifdef CONFIG_UBIFS_FS_SECURITY
+	&ubifs_security_xattr_handler,
+#endif
+	NULL
+};