new file mode 100644
@@ -0,0 +1,545 @@
+// 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)
+{
+ SHASH_DESC_ON_STACK(shash, c->hmac_tfm);
+ int err;
+ const char well_known_message[] = "UBIFS";
+
+ if (!ubifs_authenticated(c))
+ return 0;
+
+ shash->tfm = c->hmac_tfm;
+
+ err = crypto_shash_init(shash);
+ if (err)
+ return err;
+
+ err = crypto_shash_update(shash, well_known_message,
+ sizeof(well_known_message) - 1);
+ if (err < 0)
+ return err;
+
+ err = crypto_shash_final(shash, hmac);
+ if (err)
+ return err;
+ return 0;
+}
+
+/*
+ * 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);
+}
new file mode 100644
@@ -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;
+}
new file mode 100644
@@ -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;
+}
new file mode 100644
@@ -0,0 +1,3051 @@
+// 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 most of the debugging stuff which is compiled in only
+ * when it is enabled. But some debugging check functions are implemented in
+ * corresponding subsystem, just because they are closely related and utilize
+ * various local functions of those subsystems.
+ */
+
+#include <linux/module.h>
+#include <linux/debugfs.h>
+#include <linux/math64.h>
+#include <linux/uaccess.h>
+#include <linux/random.h>
+#include <linux/ctype.h>
+#include "ubifs.h"
+
+static DEFINE_SPINLOCK(dbg_lock);
+
+static const char *get_key_fmt(int fmt)
+{
+ switch (fmt) {
+ case UBIFS_SIMPLE_KEY_FMT:
+ return "simple";
+ default:
+ return "unknown/invalid format";
+ }
+}
+
+static const char *get_key_hash(int hash)
+{
+ switch (hash) {
+ case UBIFS_KEY_HASH_R5:
+ return "R5";
+ case UBIFS_KEY_HASH_TEST:
+ return "test";
+ default:
+ return "unknown/invalid name hash";
+ }
+}
+
+static const char *get_key_type(int type)
+{
+ switch (type) {
+ case UBIFS_INO_KEY:
+ return "inode";
+ case UBIFS_DENT_KEY:
+ return "direntry";
+ case UBIFS_XENT_KEY:
+ return "xentry";
+ case UBIFS_DATA_KEY:
+ return "data";
+ case UBIFS_TRUN_KEY:
+ return "truncate";
+ default:
+ return "unknown/invalid key";
+ }
+}
+
+static const char *get_dent_type(int type)
+{
+ switch (type) {
+ case UBIFS_ITYPE_REG:
+ return "file";
+ case UBIFS_ITYPE_DIR:
+ return "dir";
+ case UBIFS_ITYPE_LNK:
+ return "symlink";
+ case UBIFS_ITYPE_BLK:
+ return "blkdev";
+ case UBIFS_ITYPE_CHR:
+ return "char dev";
+ case UBIFS_ITYPE_FIFO:
+ return "fifo";
+ case UBIFS_ITYPE_SOCK:
+ return "socket";
+ default:
+ return "unknown/invalid type";
+ }
+}
+
+const char *dbg_snprintf_key(const struct ubifs_info *c,
+ const union ubifs_key *key, char *buffer, int len)
+{
+ char *p = buffer;
+ int type = key_type(c, key);
+
+ if (c->key_fmt == UBIFS_SIMPLE_KEY_FMT) {
+ switch (type) {
+ case UBIFS_INO_KEY:
+ len -= snprintf(p, len, "(%lu, %s)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type));
+ break;
+ case UBIFS_DENT_KEY:
+ case UBIFS_XENT_KEY:
+ len -= snprintf(p, len, "(%lu, %s, %#08x)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type), key_hash(c, key));
+ break;
+ case UBIFS_DATA_KEY:
+ len -= snprintf(p, len, "(%lu, %s, %u)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type), key_block(c, key));
+ break;
+ case UBIFS_TRUN_KEY:
+ len -= snprintf(p, len, "(%lu, %s)",
+ (unsigned long)key_inum(c, key),
+ get_key_type(type));
+ break;
+ default:
+ len -= snprintf(p, len, "(bad key type: %#08x, %#08x)",
+ key->u32[0], key->u32[1]);
+ }
+ } else
+ len -= snprintf(p, len, "bad key format %d", c->key_fmt);
+ ubifs_assert(c, len > 0);
+ return p;
+}
+
+const char *dbg_ntype(int type)
+{
+ switch (type) {
+ case UBIFS_PAD_NODE:
+ return "padding node";
+ case UBIFS_SB_NODE:
+ return "superblock node";
+ case UBIFS_MST_NODE:
+ return "master node";
+ case UBIFS_REF_NODE:
+ return "reference node";
+ case UBIFS_INO_NODE:
+ return "inode node";
+ case UBIFS_DENT_NODE:
+ return "direntry node";
+ case UBIFS_XENT_NODE:
+ return "xentry node";
+ case UBIFS_DATA_NODE:
+ return "data node";
+ case UBIFS_TRUN_NODE:
+ return "truncate node";
+ case UBIFS_IDX_NODE:
+ return "indexing node";
+ case UBIFS_CS_NODE:
+ return "commit start node";
+ case UBIFS_ORPH_NODE:
+ return "orphan node";
+ case UBIFS_AUTH_NODE:
+ return "auth node";
+ default:
+ return "unknown node";
+ }
+}
+
+static const char *dbg_gtype(int type)
+{
+ switch (type) {
+ case UBIFS_NO_NODE_GROUP:
+ return "no node group";
+ case UBIFS_IN_NODE_GROUP:
+ return "in node group";
+ case UBIFS_LAST_OF_NODE_GROUP:
+ return "last of node group";
+ default:
+ return "unknown";
+ }
+}
+
+const char *dbg_cstate(int cmt_state)
+{
+ switch (cmt_state) {
+ case COMMIT_RESTING:
+ return "commit resting";
+ case COMMIT_BACKGROUND:
+ return "background commit requested";
+ case COMMIT_REQUIRED:
+ return "commit required";
+ case COMMIT_RUNNING_BACKGROUND:
+ return "BACKGROUND commit running";
+ case COMMIT_RUNNING_REQUIRED:
+ return "commit running and required";
+ case COMMIT_BROKEN:
+ return "broken commit";
+ default:
+ return "unknown commit state";
+ }
+}
+
+const char *dbg_jhead(int jhead)
+{
+ switch (jhead) {
+ case GCHD:
+ return "0 (GC)";
+ case BASEHD:
+ return "1 (base)";
+ case DATAHD:
+ return "2 (data)";
+ default:
+ return "unknown journal head";
+ }
+}
+
+static void dump_ch(const struct ubifs_ch *ch)
+{
+ pr_err("\tmagic %#x\n", le32_to_cpu(ch->magic));
+ pr_err("\tcrc %#x\n", le32_to_cpu(ch->crc));
+ pr_err("\tnode_type %d (%s)\n", ch->node_type,
+ dbg_ntype(ch->node_type));
+ pr_err("\tgroup_type %d (%s)\n", ch->group_type,
+ dbg_gtype(ch->group_type));
+ pr_err("\tsqnum %llu\n",
+ (unsigned long long)le64_to_cpu(ch->sqnum));
+ pr_err("\tlen %u\n", le32_to_cpu(ch->len));
+}
+
+void ubifs_dump_inode(struct ubifs_info *c, const struct inode *inode)
+{
+ const struct ubifs_inode *ui = ubifs_inode(inode);
+ struct fscrypt_name nm = {0};
+ union ubifs_key key;
+ struct ubifs_dent_node *dent, *pdent = NULL;
+ int count = 2;
+
+ pr_err("Dump in-memory inode:");
+ pr_err("\tinode %lu\n", inode->i_ino);
+ pr_err("\tsize %llu\n",
+ (unsigned long long)i_size_read(inode));
+ pr_err("\tnlink %u\n", inode->i_nlink);
+ pr_err("\tuid %u\n", (unsigned int)i_uid_read(inode));
+ pr_err("\tgid %u\n", (unsigned int)i_gid_read(inode));
+ pr_err("\tatime %u.%u\n",
+ (unsigned int) inode_get_atime_sec(inode),
+ (unsigned int) inode_get_atime_nsec(inode));
+ pr_err("\tmtime %u.%u\n",
+ (unsigned int) inode_get_mtime_sec(inode),
+ (unsigned int) inode_get_mtime_nsec(inode));
+ pr_err("\tctime %u.%u\n",
+ (unsigned int) inode_get_ctime_sec(inode),
+ (unsigned int) inode_get_ctime_nsec(inode));
+ pr_err("\tcreat_sqnum %llu\n", ui->creat_sqnum);
+ pr_err("\txattr_size %u\n", ui->xattr_size);
+ pr_err("\txattr_cnt %u\n", ui->xattr_cnt);
+ pr_err("\txattr_names %u\n", ui->xattr_names);
+ pr_err("\tdirty %u\n", ui->dirty);
+ pr_err("\txattr %u\n", ui->xattr);
+ pr_err("\tbulk_read %u\n", ui->bulk_read);
+ pr_err("\tsynced_i_size %llu\n",
+ (unsigned long long)ui->synced_i_size);
+ pr_err("\tui_size %llu\n",
+ (unsigned long long)ui->ui_size);
+ pr_err("\tflags %d\n", ui->flags);
+ pr_err("\tcompr_type %d\n", ui->compr_type);
+ pr_err("\tlast_page_read %lu\n", ui->last_page_read);
+ pr_err("\tread_in_a_row %lu\n", ui->read_in_a_row);
+ pr_err("\tdata_len %d\n", ui->data_len);
+
+ if (!S_ISDIR(inode->i_mode))
+ return;
+
+ pr_err("List of directory entries:\n");
+ ubifs_assert(c, !mutex_is_locked(&c->tnc_mutex));
+
+ lowest_dent_key(c, &key, inode->i_ino);
+ while (1) {
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ if (PTR_ERR(dent) != -ENOENT)
+ pr_err("error %ld\n", PTR_ERR(dent));
+ break;
+ }
+
+ pr_err("\t%d: inode %llu, type %s, len %d\n",
+ count++, (unsigned long long) le64_to_cpu(dent->inum),
+ get_dent_type(dent->type),
+ le16_to_cpu(dent->nlen));
+
+ fname_name(&nm) = dent->name;
+ fname_len(&nm) = le16_to_cpu(dent->nlen);
+ kfree(pdent);
+ pdent = dent;
+ key_read(c, &dent->key, &key);
+ }
+ kfree(pdent);
+}
+
+void ubifs_dump_node(const struct ubifs_info *c, const void *node, int node_len)
+{
+ int i, n, type, safe_len, max_node_len, min_node_len;
+ union ubifs_key key;
+ const struct ubifs_ch *ch = node;
+ char key_buf[DBG_KEY_BUF_LEN];
+
+ /* If the magic is incorrect, just hexdump the first bytes */
+ if (le32_to_cpu(ch->magic) != UBIFS_NODE_MAGIC) {
+ pr_err("Not a node, first %zu bytes:", UBIFS_CH_SZ);
+ print_hex_dump(KERN_ERR, "", DUMP_PREFIX_OFFSET, 32, 1,
+ (void *)node, UBIFS_CH_SZ, 1);
+ return;
+ }
+
+ /* Skip dumping unknown type node */
+ type = ch->node_type;
+ if (type < 0 || type >= UBIFS_NODE_TYPES_CNT) {
+ pr_err("node type %d was not recognized\n", type);
+ return;
+ }
+
+ spin_lock(&dbg_lock);
+ dump_ch(node);
+
+ if (c->ranges[type].max_len == 0) {
+ max_node_len = min_node_len = c->ranges[type].len;
+ } else {
+ max_node_len = c->ranges[type].max_len;
+ min_node_len = c->ranges[type].min_len;
+ }
+ safe_len = le32_to_cpu(ch->len);
+ safe_len = safe_len > 0 ? safe_len : 0;
+ safe_len = min3(safe_len, max_node_len, node_len);
+ if (safe_len < min_node_len) {
+ pr_err("node len(%d) is too short for %s, left %d bytes:\n",
+ safe_len, dbg_ntype(type),
+ safe_len > UBIFS_CH_SZ ?
+ safe_len - (int)UBIFS_CH_SZ : 0);
+ if (safe_len > UBIFS_CH_SZ)
+ print_hex_dump(KERN_ERR, "", DUMP_PREFIX_OFFSET, 32, 1,
+ (void *)node + UBIFS_CH_SZ,
+ safe_len - UBIFS_CH_SZ, 0);
+ goto out_unlock;
+ }
+ if (safe_len != le32_to_cpu(ch->len))
+ pr_err("\ttruncated node length %d\n", safe_len);
+
+ switch (type) {
+ case UBIFS_PAD_NODE:
+ {
+ const struct ubifs_pad_node *pad = node;
+
+ pr_err("\tpad_len %u\n", le32_to_cpu(pad->pad_len));
+ break;
+ }
+ case UBIFS_SB_NODE:
+ {
+ const struct ubifs_sb_node *sup = node;
+ unsigned int sup_flags = le32_to_cpu(sup->flags);
+
+ pr_err("\tkey_hash %d (%s)\n",
+ (int)sup->key_hash, get_key_hash(sup->key_hash));
+ pr_err("\tkey_fmt %d (%s)\n",
+ (int)sup->key_fmt, get_key_fmt(sup->key_fmt));
+ pr_err("\tflags %#x\n", sup_flags);
+ pr_err("\tbig_lpt %u\n",
+ !!(sup_flags & UBIFS_FLG_BIGLPT));
+ pr_err("\tspace_fixup %u\n",
+ !!(sup_flags & UBIFS_FLG_SPACE_FIXUP));
+ pr_err("\tmin_io_size %u\n", le32_to_cpu(sup->min_io_size));
+ pr_err("\tleb_size %u\n", le32_to_cpu(sup->leb_size));
+ pr_err("\tleb_cnt %u\n", le32_to_cpu(sup->leb_cnt));
+ pr_err("\tmax_leb_cnt %u\n", le32_to_cpu(sup->max_leb_cnt));
+ pr_err("\tmax_bud_bytes %llu\n",
+ (unsigned long long)le64_to_cpu(sup->max_bud_bytes));
+ pr_err("\tlog_lebs %u\n", le32_to_cpu(sup->log_lebs));
+ pr_err("\tlpt_lebs %u\n", le32_to_cpu(sup->lpt_lebs));
+ pr_err("\torph_lebs %u\n", le32_to_cpu(sup->orph_lebs));
+ pr_err("\tjhead_cnt %u\n", le32_to_cpu(sup->jhead_cnt));
+ pr_err("\tfanout %u\n", le32_to_cpu(sup->fanout));
+ pr_err("\tlsave_cnt %u\n", le32_to_cpu(sup->lsave_cnt));
+ pr_err("\tdefault_compr %u\n",
+ (int)le16_to_cpu(sup->default_compr));
+ pr_err("\trp_size %llu\n",
+ (unsigned long long)le64_to_cpu(sup->rp_size));
+ pr_err("\trp_uid %u\n", le32_to_cpu(sup->rp_uid));
+ pr_err("\trp_gid %u\n", le32_to_cpu(sup->rp_gid));
+ pr_err("\tfmt_version %u\n", le32_to_cpu(sup->fmt_version));
+ pr_err("\ttime_gran %u\n", le32_to_cpu(sup->time_gran));
+ pr_err("\tUUID %pUB\n", sup->uuid);
+ break;
+ }
+ case UBIFS_MST_NODE:
+ {
+ const struct ubifs_mst_node *mst = node;
+
+ pr_err("\thighest_inum %llu\n",
+ (unsigned long long)le64_to_cpu(mst->highest_inum));
+ pr_err("\tcommit number %llu\n",
+ (unsigned long long)le64_to_cpu(mst->cmt_no));
+ pr_err("\tflags %#x\n", le32_to_cpu(mst->flags));
+ pr_err("\tlog_lnum %u\n", le32_to_cpu(mst->log_lnum));
+ pr_err("\troot_lnum %u\n", le32_to_cpu(mst->root_lnum));
+ pr_err("\troot_offs %u\n", le32_to_cpu(mst->root_offs));
+ pr_err("\troot_len %u\n", le32_to_cpu(mst->root_len));
+ pr_err("\tgc_lnum %u\n", le32_to_cpu(mst->gc_lnum));
+ pr_err("\tihead_lnum %u\n", le32_to_cpu(mst->ihead_lnum));
+ pr_err("\tihead_offs %u\n", le32_to_cpu(mst->ihead_offs));
+ pr_err("\tindex_size %llu\n",
+ (unsigned long long)le64_to_cpu(mst->index_size));
+ pr_err("\tlpt_lnum %u\n", le32_to_cpu(mst->lpt_lnum));
+ pr_err("\tlpt_offs %u\n", le32_to_cpu(mst->lpt_offs));
+ pr_err("\tnhead_lnum %u\n", le32_to_cpu(mst->nhead_lnum));
+ pr_err("\tnhead_offs %u\n", le32_to_cpu(mst->nhead_offs));
+ pr_err("\tltab_lnum %u\n", le32_to_cpu(mst->ltab_lnum));
+ pr_err("\tltab_offs %u\n", le32_to_cpu(mst->ltab_offs));
+ pr_err("\tlsave_lnum %u\n", le32_to_cpu(mst->lsave_lnum));
+ pr_err("\tlsave_offs %u\n", le32_to_cpu(mst->lsave_offs));
+ pr_err("\tlscan_lnum %u\n", le32_to_cpu(mst->lscan_lnum));
+ pr_err("\tleb_cnt %u\n", le32_to_cpu(mst->leb_cnt));
+ pr_err("\tempty_lebs %u\n", le32_to_cpu(mst->empty_lebs));
+ pr_err("\tidx_lebs %u\n", le32_to_cpu(mst->idx_lebs));
+ pr_err("\ttotal_free %llu\n",
+ (unsigned long long)le64_to_cpu(mst->total_free));
+ pr_err("\ttotal_dirty %llu\n",
+ (unsigned long long)le64_to_cpu(mst->total_dirty));
+ pr_err("\ttotal_used %llu\n",
+ (unsigned long long)le64_to_cpu(mst->total_used));
+ pr_err("\ttotal_dead %llu\n",
+ (unsigned long long)le64_to_cpu(mst->total_dead));
+ pr_err("\ttotal_dark %llu\n",
+ (unsigned long long)le64_to_cpu(mst->total_dark));
+ break;
+ }
+ case UBIFS_REF_NODE:
+ {
+ const struct ubifs_ref_node *ref = node;
+
+ pr_err("\tlnum %u\n", le32_to_cpu(ref->lnum));
+ pr_err("\toffs %u\n", le32_to_cpu(ref->offs));
+ pr_err("\tjhead %u\n", le32_to_cpu(ref->jhead));
+ break;
+ }
+ case UBIFS_INO_NODE:
+ {
+ const struct ubifs_ino_node *ino = node;
+
+ key_read(c, &ino->key, &key);
+ pr_err("\tkey %s\n",
+ dbg_snprintf_key(c, &key, key_buf, DBG_KEY_BUF_LEN));
+ pr_err("\tcreat_sqnum %llu\n",
+ (unsigned long long)le64_to_cpu(ino->creat_sqnum));
+ pr_err("\tsize %llu\n",
+ (unsigned long long)le64_to_cpu(ino->size));
+ pr_err("\tnlink %u\n", le32_to_cpu(ino->nlink));
+ pr_err("\tatime %lld.%u\n",
+ (long long)le64_to_cpu(ino->atime_sec),
+ le32_to_cpu(ino->atime_nsec));
+ pr_err("\tmtime %lld.%u\n",
+ (long long)le64_to_cpu(ino->mtime_sec),
+ le32_to_cpu(ino->mtime_nsec));
+ pr_err("\tctime %lld.%u\n",
+ (long long)le64_to_cpu(ino->ctime_sec),
+ le32_to_cpu(ino->ctime_nsec));
+ pr_err("\tuid %u\n", le32_to_cpu(ino->uid));
+ pr_err("\tgid %u\n", le32_to_cpu(ino->gid));
+ pr_err("\tmode %u\n", le32_to_cpu(ino->mode));
+ pr_err("\tflags %#x\n", le32_to_cpu(ino->flags));
+ pr_err("\txattr_cnt %u\n", le32_to_cpu(ino->xattr_cnt));
+ pr_err("\txattr_size %u\n", le32_to_cpu(ino->xattr_size));
+ pr_err("\txattr_names %u\n", le32_to_cpu(ino->xattr_names));
+ pr_err("\tcompr_type %#x\n",
+ (int)le16_to_cpu(ino->compr_type));
+ pr_err("\tdata len %u\n", le32_to_cpu(ino->data_len));
+ break;
+ }
+ case UBIFS_DENT_NODE:
+ case UBIFS_XENT_NODE:
+ {
+ const struct ubifs_dent_node *dent = node;
+ int nlen = le16_to_cpu(dent->nlen);
+
+ key_read(c, &dent->key, &key);
+ pr_err("\tkey %s\n",
+ dbg_snprintf_key(c, &key, key_buf, DBG_KEY_BUF_LEN));
+ pr_err("\tinum %llu\n",
+ (unsigned long long)le64_to_cpu(dent->inum));
+ pr_err("\ttype %d\n", (int)dent->type);
+ pr_err("\tnlen %d\n", nlen);
+ pr_err("\tname ");
+
+ if (nlen > UBIFS_MAX_NLEN ||
+ nlen > safe_len - UBIFS_DENT_NODE_SZ)
+ pr_err("(bad name length, not printing, bad or corrupted node)");
+ else {
+ for (i = 0; i < nlen && dent->name[i]; i++)
+ pr_cont("%c", isprint(dent->name[i]) ?
+ dent->name[i] : '?');
+ }
+ pr_cont("\n");
+
+ break;
+ }
+ case UBIFS_DATA_NODE:
+ {
+ const struct ubifs_data_node *dn = node;
+
+ key_read(c, &dn->key, &key);
+ pr_err("\tkey %s\n",
+ dbg_snprintf_key(c, &key, key_buf, DBG_KEY_BUF_LEN));
+ pr_err("\tsize %u\n", le32_to_cpu(dn->size));
+ pr_err("\tcompr_typ %d\n",
+ (int)le16_to_cpu(dn->compr_type));
+ pr_err("\tdata size %u\n",
+ le32_to_cpu(ch->len) - (unsigned int)UBIFS_DATA_NODE_SZ);
+ pr_err("\tdata (length = %d):\n",
+ safe_len - (int)UBIFS_DATA_NODE_SZ);
+ print_hex_dump(KERN_ERR, "\t", DUMP_PREFIX_OFFSET, 32, 1,
+ (void *)&dn->data,
+ safe_len - (int)UBIFS_DATA_NODE_SZ, 0);
+ break;
+ }
+ case UBIFS_TRUN_NODE:
+ {
+ const struct ubifs_trun_node *trun = node;
+
+ pr_err("\tinum %u\n", le32_to_cpu(trun->inum));
+ pr_err("\told_size %llu\n",
+ (unsigned long long)le64_to_cpu(trun->old_size));
+ pr_err("\tnew_size %llu\n",
+ (unsigned long long)le64_to_cpu(trun->new_size));
+ break;
+ }
+ case UBIFS_IDX_NODE:
+ {
+ const struct ubifs_idx_node *idx = node;
+ int max_child_cnt = (safe_len - UBIFS_IDX_NODE_SZ) /
+ (ubifs_idx_node_sz(c, 1) -
+ UBIFS_IDX_NODE_SZ);
+
+ n = min_t(int, le16_to_cpu(idx->child_cnt), max_child_cnt);
+ pr_err("\tchild_cnt %d\n", (int)le16_to_cpu(idx->child_cnt));
+ pr_err("\tlevel %d\n", (int)le16_to_cpu(idx->level));
+ pr_err("\tBranches:\n");
+
+ for (i = 0; i < n && i < c->fanout; i++) {
+ const struct ubifs_branch *br;
+
+ br = ubifs_idx_branch(c, idx, i);
+ key_read(c, &br->key, &key);
+ pr_err("\t%d: LEB %d:%d len %d key %s\n",
+ i, le32_to_cpu(br->lnum), le32_to_cpu(br->offs),
+ le32_to_cpu(br->len),
+ dbg_snprintf_key(c, &key, key_buf,
+ DBG_KEY_BUF_LEN));
+ }
+ break;
+ }
+ case UBIFS_CS_NODE:
+ break;
+ case UBIFS_ORPH_NODE:
+ {
+ const struct ubifs_orph_node *orph = node;
+
+ pr_err("\tcommit number %llu\n",
+ (unsigned long long)
+ le64_to_cpu(orph->cmt_no) & LLONG_MAX);
+ pr_err("\tlast node flag %llu\n",
+ (unsigned long long)(le64_to_cpu(orph->cmt_no)) >> 63);
+ n = (safe_len - UBIFS_ORPH_NODE_SZ) >> 3;
+ pr_err("\t%d orphan inode numbers:\n", n);
+ for (i = 0; i < n; i++)
+ pr_err("\t ino %llu\n",
+ (unsigned long long)le64_to_cpu(orph->inos[i]));
+ break;
+ }
+ case UBIFS_AUTH_NODE:
+ {
+ break;
+ }
+ default:
+ pr_err("node type %d was not recognized\n", type);
+ }
+
+out_unlock:
+ spin_unlock(&dbg_lock);
+}
+
+void ubifs_dump_budget_req(const struct ubifs_budget_req *req)
+{
+ spin_lock(&dbg_lock);
+ pr_err("Budgeting request: new_ino %d, dirtied_ino %d\n",
+ req->new_ino, req->dirtied_ino);
+ pr_err("\tnew_ino_d %d, dirtied_ino_d %d\n",
+ req->new_ino_d, req->dirtied_ino_d);
+ pr_err("\tnew_page %d, dirtied_page %d\n",
+ req->new_page, req->dirtied_page);
+ pr_err("\tnew_dent %d, mod_dent %d\n",
+ req->new_dent, req->mod_dent);
+ pr_err("\tidx_growth %d\n", req->idx_growth);
+ pr_err("\tdata_growth %d dd_growth %d\n",
+ req->data_growth, req->dd_growth);
+ spin_unlock(&dbg_lock);
+}
+
+void ubifs_dump_lstats(const struct ubifs_lp_stats *lst)
+{
+ spin_lock(&dbg_lock);
+ pr_err("(pid %d) Lprops statistics: empty_lebs %d, idx_lebs %d\n",
+ current->pid, lst->empty_lebs, lst->idx_lebs);
+ pr_err("\ttaken_empty_lebs %d, total_free %lld, total_dirty %lld\n",
+ lst->taken_empty_lebs, lst->total_free, lst->total_dirty);
+ pr_err("\ttotal_used %lld, total_dark %lld, total_dead %lld\n",
+ lst->total_used, lst->total_dark, lst->total_dead);
+ spin_unlock(&dbg_lock);
+}
+
+void ubifs_dump_budg(struct ubifs_info *c, const struct ubifs_budg_info *bi)
+{
+ int i;
+ struct rb_node *rb;
+ struct ubifs_bud *bud;
+ struct ubifs_gced_idx_leb *idx_gc;
+ long long available, outstanding, free;
+
+ spin_lock(&c->space_lock);
+ spin_lock(&dbg_lock);
+ pr_err("(pid %d) Budgeting info: data budget sum %lld, total budget sum %lld\n",
+ current->pid, bi->data_growth + bi->dd_growth,
+ bi->data_growth + bi->dd_growth + bi->idx_growth);
+ pr_err("\tbudg_data_growth %lld, budg_dd_growth %lld, budg_idx_growth %lld\n",
+ bi->data_growth, bi->dd_growth, bi->idx_growth);
+ pr_err("\tmin_idx_lebs %d, old_idx_sz %llu, uncommitted_idx %lld\n",
+ bi->min_idx_lebs, bi->old_idx_sz, bi->uncommitted_idx);
+ pr_err("\tpage_budget %d, inode_budget %d, dent_budget %d\n",
+ bi->page_budget, bi->inode_budget, bi->dent_budget);
+ pr_err("\tnospace %u, nospace_rp %u\n", bi->nospace, bi->nospace_rp);
+ pr_err("\tdark_wm %d, dead_wm %d, max_idx_node_sz %d\n",
+ c->dark_wm, c->dead_wm, c->max_idx_node_sz);
+
+ if (bi != &c->bi)
+ /*
+ * If we are dumping saved budgeting data, do not print
+ * additional information which is about the current state, not
+ * the old one which corresponded to the saved budgeting data.
+ */
+ goto out_unlock;
+
+ pr_err("\tfreeable_cnt %d, calc_idx_sz %lld, idx_gc_cnt %d\n",
+ c->freeable_cnt, c->calc_idx_sz, c->idx_gc_cnt);
+ pr_err("\tdirty_pg_cnt %ld, dirty_zn_cnt %ld, clean_zn_cnt %ld\n",
+ atomic_long_read(&c->dirty_pg_cnt),
+ atomic_long_read(&c->dirty_zn_cnt),
+ atomic_long_read(&c->clean_zn_cnt));
+ pr_err("\tgc_lnum %d, ihead_lnum %d\n", c->gc_lnum, c->ihead_lnum);
+
+ /* If we are in R/O mode, journal heads do not exist */
+ if (c->jheads)
+ for (i = 0; i < c->jhead_cnt; i++)
+ pr_err("\tjhead %s\t LEB %d\n",
+ dbg_jhead(c->jheads[i].wbuf.jhead),
+ c->jheads[i].wbuf.lnum);
+ for (rb = rb_first(&c->buds); rb; rb = rb_next(rb)) {
+ bud = rb_entry(rb, struct ubifs_bud, rb);
+ pr_err("\tbud LEB %d\n", bud->lnum);
+ }
+ list_for_each_entry(bud, &c->old_buds, list)
+ pr_err("\told bud LEB %d\n", bud->lnum);
+ list_for_each_entry(idx_gc, &c->idx_gc, list)
+ pr_err("\tGC'ed idx LEB %d unmap %d\n",
+ idx_gc->lnum, idx_gc->unmap);
+ pr_err("\tcommit state %d\n", c->cmt_state);
+
+ /* Print budgeting predictions */
+ available = ubifs_calc_available(c, c->bi.min_idx_lebs);
+ outstanding = c->bi.data_growth + c->bi.dd_growth;
+ free = ubifs_get_free_space_nolock(c);
+ pr_err("Budgeting predictions:\n");
+ pr_err("\tavailable: %lld, outstanding %lld, free %lld\n",
+ available, outstanding, free);
+out_unlock:
+ spin_unlock(&dbg_lock);
+ spin_unlock(&c->space_lock);
+}
+
+void ubifs_dump_lprop(const struct ubifs_info *c, const struct ubifs_lprops *lp)
+{
+ int i, spc, dark = 0, dead = 0;
+ struct rb_node *rb;
+ struct ubifs_bud *bud;
+
+ spc = lp->free + lp->dirty;
+ if (spc < c->dead_wm)
+ dead = spc;
+ else
+ dark = ubifs_calc_dark(c, spc);
+
+ if (lp->flags & LPROPS_INDEX)
+ pr_err("LEB %-7d free %-8d dirty %-8d used %-8d free + dirty %-8d flags %#x (",
+ lp->lnum, lp->free, lp->dirty, c->leb_size - spc, spc,
+ lp->flags);
+ else
+ pr_err("LEB %-7d free %-8d dirty %-8d used %-8d free + dirty %-8d dark %-4d dead %-4d nodes fit %-3d flags %#-4x (",
+ lp->lnum, lp->free, lp->dirty, c->leb_size - spc, spc,
+ dark, dead, (int)(spc / UBIFS_MAX_NODE_SZ), lp->flags);
+
+ if (lp->flags & LPROPS_TAKEN) {
+ if (lp->flags & LPROPS_INDEX)
+ pr_cont("index, taken");
+ else
+ pr_cont("taken");
+ } else {
+ const char *s;
+
+ if (lp->flags & LPROPS_INDEX) {
+ switch (lp->flags & LPROPS_CAT_MASK) {
+ case LPROPS_DIRTY_IDX:
+ s = "dirty index";
+ break;
+ case LPROPS_FRDI_IDX:
+ s = "freeable index";
+ break;
+ default:
+ s = "index";
+ }
+ } else {
+ switch (lp->flags & LPROPS_CAT_MASK) {
+ case LPROPS_UNCAT:
+ s = "not categorized";
+ break;
+ case LPROPS_DIRTY:
+ s = "dirty";
+ break;
+ case LPROPS_FREE:
+ s = "free";
+ break;
+ case LPROPS_EMPTY:
+ s = "empty";
+ break;
+ case LPROPS_FREEABLE:
+ s = "freeable";
+ break;
+ default:
+ s = NULL;
+ break;
+ }
+ }
+ pr_cont("%s", s);
+ }
+
+ for (rb = rb_first((struct rb_root *)&c->buds); rb; rb = rb_next(rb)) {
+ bud = rb_entry(rb, struct ubifs_bud, rb);
+ if (bud->lnum == lp->lnum) {
+ int head = 0;
+ for (i = 0; i < c->jhead_cnt; i++) {
+ /*
+ * Note, if we are in R/O mode or in the middle
+ * of mounting/re-mounting, the write-buffers do
+ * not exist.
+ */
+ if (c->jheads &&
+ lp->lnum == c->jheads[i].wbuf.lnum) {
+ pr_cont(", jhead %s", dbg_jhead(i));
+ head = 1;
+ }
+ }
+ if (!head)
+ pr_cont(", bud of jhead %s",
+ dbg_jhead(bud->jhead));
+ }
+ }
+ if (lp->lnum == c->gc_lnum)
+ pr_cont(", GC LEB");
+ pr_cont(")\n");
+}
+
+void ubifs_dump_lprops(struct ubifs_info *c)
+{
+ int lnum, err;
+ struct ubifs_lprops lp;
+ struct ubifs_lp_stats lst;
+
+ pr_err("(pid %d) start dumping LEB properties\n", current->pid);
+ ubifs_get_lp_stats(c, &lst);
+ ubifs_dump_lstats(&lst);
+
+ for (lnum = c->main_first; lnum < c->leb_cnt; lnum++) {
+ err = ubifs_read_one_lp(c, lnum, &lp);
+ if (err) {
+ ubifs_err(c, "cannot read lprops for LEB %d", lnum);
+ continue;
+ }
+
+ ubifs_dump_lprop(c, &lp);
+ }
+ pr_err("(pid %d) finish dumping LEB properties\n", current->pid);
+}
+
+void ubifs_dump_lpt_info(struct ubifs_info *c)
+{
+ int i;
+
+ spin_lock(&dbg_lock);
+ pr_err("(pid %d) dumping LPT information\n", current->pid);
+ pr_err("\tlpt_sz: %lld\n", c->lpt_sz);
+ pr_err("\tpnode_sz: %d\n", c->pnode_sz);
+ pr_err("\tnnode_sz: %d\n", c->nnode_sz);
+ pr_err("\tltab_sz: %d\n", c->ltab_sz);
+ pr_err("\tlsave_sz: %d\n", c->lsave_sz);
+ pr_err("\tbig_lpt: %u\n", c->big_lpt);
+ pr_err("\tlpt_hght: %d\n", c->lpt_hght);
+ pr_err("\tpnode_cnt: %d\n", c->pnode_cnt);
+ pr_err("\tnnode_cnt: %d\n", c->nnode_cnt);
+ pr_err("\tdirty_pn_cnt: %d\n", c->dirty_pn_cnt);
+ pr_err("\tdirty_nn_cnt: %d\n", c->dirty_nn_cnt);
+ pr_err("\tlsave_cnt: %d\n", c->lsave_cnt);
+ pr_err("\tspace_bits: %d\n", c->space_bits);
+ pr_err("\tlpt_lnum_bits: %d\n", c->lpt_lnum_bits);
+ pr_err("\tlpt_offs_bits: %d\n", c->lpt_offs_bits);
+ pr_err("\tlpt_spc_bits: %d\n", c->lpt_spc_bits);
+ pr_err("\tpcnt_bits: %d\n", c->pcnt_bits);
+ pr_err("\tlnum_bits: %d\n", c->lnum_bits);
+ pr_err("\tLPT root is at %d:%d\n", c->lpt_lnum, c->lpt_offs);
+ pr_err("\tLPT head is at %d:%d\n",
+ c->nhead_lnum, c->nhead_offs);
+ pr_err("\tLPT ltab is at %d:%d\n", c->ltab_lnum, c->ltab_offs);
+ if (c->big_lpt)
+ pr_err("\tLPT lsave is at %d:%d\n",
+ c->lsave_lnum, c->lsave_offs);
+ for (i = 0; i < c->lpt_lebs; i++)
+ pr_err("\tLPT LEB %d free %d dirty %d tgc %d cmt %d\n",
+ i + c->lpt_first, c->ltab[i].free, c->ltab[i].dirty,
+ c->ltab[i].tgc, c->ltab[i].cmt);
+ spin_unlock(&dbg_lock);
+}
+
+void ubifs_dump_leb(const struct ubifs_info *c, int lnum)
+{
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ void *buf;
+
+ pr_err("(pid %d) start dumping LEB %d\n", current->pid, lnum);
+
+ buf = __vmalloc(c->leb_size, GFP_NOFS);
+ if (!buf) {
+ ubifs_err(c, "cannot allocate memory for dumping LEB %d", lnum);
+ return;
+ }
+
+ sleb = ubifs_scan(c, lnum, 0, buf, 0);
+ if (IS_ERR(sleb)) {
+ ubifs_err(c, "scan error %d", (int)PTR_ERR(sleb));
+ goto out;
+ }
+
+ pr_err("LEB %d has %d nodes ending at %d\n", lnum,
+ sleb->nodes_cnt, sleb->endpt);
+
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ cond_resched();
+ pr_err("Dumping node at LEB %d:%d len %d\n", lnum,
+ snod->offs, snod->len);
+ ubifs_dump_node(c, snod->node, c->leb_size - snod->offs);
+ }
+
+ pr_err("(pid %d) finish dumping LEB %d\n", current->pid, lnum);
+ ubifs_scan_destroy(sleb);
+
+out:
+ vfree(buf);
+ return;
+}
+
+void ubifs_dump_znode(const struct ubifs_info *c,
+ const struct ubifs_znode *znode)
+{
+ int n;
+ const struct ubifs_zbranch *zbr;
+ char key_buf[DBG_KEY_BUF_LEN];
+
+ spin_lock(&dbg_lock);
+ if (znode->parent)
+ zbr = &znode->parent->zbranch[znode->iip];
+ else
+ zbr = &c->zroot;
+
+ pr_err("znode %p, LEB %d:%d len %d parent %p iip %d level %d child_cnt %d flags %lx\n",
+ znode, zbr->lnum, zbr->offs, zbr->len, znode->parent, znode->iip,
+ znode->level, znode->child_cnt, znode->flags);
+
+ if (znode->child_cnt <= 0 || znode->child_cnt > c->fanout) {
+ spin_unlock(&dbg_lock);
+ return;
+ }
+
+ pr_err("zbranches:\n");
+ for (n = 0; n < znode->child_cnt; n++) {
+ zbr = &znode->zbranch[n];
+ if (znode->level > 0)
+ pr_err("\t%d: znode %p LEB %d:%d len %d key %s\n",
+ n, zbr->znode, zbr->lnum, zbr->offs, zbr->len,
+ dbg_snprintf_key(c, &zbr->key, key_buf,
+ DBG_KEY_BUF_LEN));
+ else
+ pr_err("\t%d: LNC %p LEB %d:%d len %d key %s\n",
+ n, zbr->znode, zbr->lnum, zbr->offs, zbr->len,
+ dbg_snprintf_key(c, &zbr->key, key_buf,
+ DBG_KEY_BUF_LEN));
+ }
+ spin_unlock(&dbg_lock);
+}
+
+void ubifs_dump_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat)
+{
+ int i;
+
+ pr_err("(pid %d) start dumping heap cat %d (%d elements)\n",
+ current->pid, cat, heap->cnt);
+ for (i = 0; i < heap->cnt; i++) {
+ struct ubifs_lprops *lprops = heap->arr[i];
+
+ pr_err("\t%d. LEB %d hpos %d free %d dirty %d flags %d\n",
+ i, lprops->lnum, lprops->hpos, lprops->free,
+ lprops->dirty, lprops->flags);
+ }
+ pr_err("(pid %d) finish dumping heap\n", current->pid);
+}
+
+void ubifs_dump_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i;
+
+ pr_err("(pid %d) dumping pnode:\n", current->pid);
+ pr_err("\taddress %zx parent %zx cnext %zx\n",
+ (size_t)pnode, (size_t)parent, (size_t)pnode->cnext);
+ pr_err("\tflags %lu iip %d level %d num %d\n",
+ pnode->flags, iip, pnode->level, pnode->num);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops *lp = &pnode->lprops[i];
+
+ pr_err("\t%d: free %d dirty %d flags %d lnum %d\n",
+ i, lp->free, lp->dirty, lp->flags, lp->lnum);
+ }
+}
+
+void ubifs_dump_tnc(struct ubifs_info *c)
+{
+ struct ubifs_znode *znode;
+ int level;
+
+ pr_err("\n");
+ pr_err("(pid %d) start dumping TNC tree\n", current->pid);
+ znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, NULL);
+ level = znode->level;
+ pr_err("== Level %d ==\n", level);
+ while (znode) {
+ if (level != znode->level) {
+ level = znode->level;
+ pr_err("== Level %d ==\n", level);
+ }
+ ubifs_dump_znode(c, znode);
+ znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, znode);
+ }
+ pr_err("(pid %d) finish dumping TNC tree\n", current->pid);
+}
+
+static int dump_znode(struct ubifs_info *c, struct ubifs_znode *znode,
+ void *priv)
+{
+ ubifs_dump_znode(c, znode);
+ return 0;
+}
+
+/**
+ * ubifs_dump_index - dump the on-flash index.
+ * @c: UBIFS file-system description object
+ *
+ * This function dumps whole UBIFS indexing B-tree, unlike 'ubifs_dump_tnc()'
+ * which dumps only in-memory znodes and does not read znodes which from flash.
+ */
+void ubifs_dump_index(struct ubifs_info *c)
+{
+ dbg_walk_index(c, NULL, dump_znode, NULL);
+}
+
+/**
+ * dbg_save_space_info - save information about flash space.
+ * @c: UBIFS file-system description object
+ *
+ * This function saves information about UBIFS free space, dirty space, etc, in
+ * order to check it later.
+ */
+void dbg_save_space_info(struct ubifs_info *c)
+{
+ struct ubifs_debug_info *d = c->dbg;
+ int freeable_cnt;
+
+ spin_lock(&c->space_lock);
+ memcpy(&d->saved_lst, &c->lst, sizeof(struct ubifs_lp_stats));
+ memcpy(&d->saved_bi, &c->bi, sizeof(struct ubifs_budg_info));
+ d->saved_idx_gc_cnt = c->idx_gc_cnt;
+
+ /*
+ * We use a dirty hack here and zero out @c->freeable_cnt, because it
+ * affects the free space calculations, and UBIFS might not know about
+ * all freeable eraseblocks. Indeed, we know about freeable eraseblocks
+ * only when we read their lprops, and we do this only lazily, upon the
+ * need. So at any given point of time @c->freeable_cnt might be not
+ * exactly accurate.
+ *
+ * Just one example about the issue we hit when we did not zero
+ * @c->freeable_cnt.
+ * 1. The file-system is mounted R/O, c->freeable_cnt is %0. We save the
+ * amount of free space in @d->saved_free
+ * 2. We re-mount R/W, which makes UBIFS to read the "lsave"
+ * information from flash, where we cache LEBs from various
+ * categories ('ubifs_remount_fs()' -> 'ubifs_lpt_init()'
+ * -> 'lpt_init_wr()' -> 'read_lsave()' -> 'ubifs_lpt_lookup()'
+ * -> 'ubifs_get_pnode()' -> 'update_cats()'
+ * -> 'ubifs_add_to_cat()').
+ * 3. Lsave contains a freeable eraseblock, and @c->freeable_cnt
+ * becomes %1.
+ * 4. We calculate the amount of free space when the re-mount is
+ * finished in 'dbg_check_space_info()' and it does not match
+ * @d->saved_free.
+ */
+ freeable_cnt = c->freeable_cnt;
+ c->freeable_cnt = 0;
+ d->saved_free = ubifs_get_free_space_nolock(c);
+ c->freeable_cnt = freeable_cnt;
+ spin_unlock(&c->space_lock);
+}
+
+/**
+ * dbg_check_space_info - check flash space information.
+ * @c: UBIFS file-system description object
+ *
+ * This function compares current flash space information with the information
+ * which was saved when the 'dbg_save_space_info()' function was called.
+ * Returns zero if the information has not changed, and %-EINVAL if it has
+ * changed.
+ */
+int dbg_check_space_info(struct ubifs_info *c)
+{
+ struct ubifs_debug_info *d = c->dbg;
+ struct ubifs_lp_stats lst;
+ long long free;
+ int freeable_cnt;
+
+ spin_lock(&c->space_lock);
+ freeable_cnt = c->freeable_cnt;
+ c->freeable_cnt = 0;
+ free = ubifs_get_free_space_nolock(c);
+ c->freeable_cnt = freeable_cnt;
+ spin_unlock(&c->space_lock);
+
+ if (free != d->saved_free) {
+ ubifs_err(c, "free space changed from %lld to %lld",
+ d->saved_free, free);
+ goto out;
+ }
+
+ return 0;
+
+out:
+ ubifs_msg(c, "saved lprops statistics dump");
+ ubifs_dump_lstats(&d->saved_lst);
+ ubifs_msg(c, "saved budgeting info dump");
+ ubifs_dump_budg(c, &d->saved_bi);
+ ubifs_msg(c, "saved idx_gc_cnt %d", d->saved_idx_gc_cnt);
+ ubifs_msg(c, "current lprops statistics dump");
+ ubifs_get_lp_stats(c, &lst);
+ ubifs_dump_lstats(&lst);
+ ubifs_msg(c, "current budgeting info dump");
+ ubifs_dump_budg(c, &c->bi);
+ dump_stack();
+ return -EINVAL;
+}
+
+/**
+ * dbg_check_synced_i_size - check synchronized inode size.
+ * @c: UBIFS file-system description object
+ * @inode: inode to check
+ *
+ * If inode is clean, synchronized inode size has to be equivalent to current
+ * inode size. This function has to be called only for locked inodes (@i_mutex
+ * has to be locked). Returns %0 if synchronized inode size if correct, and
+ * %-EINVAL if not.
+ */
+int dbg_check_synced_i_size(const struct ubifs_info *c, struct inode *inode)
+{
+ int err = 0;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ if (!dbg_is_chk_gen(c))
+ return 0;
+ if (!S_ISREG(inode->i_mode))
+ return 0;
+
+ mutex_lock(&ui->ui_mutex);
+ spin_lock(&ui->ui_lock);
+ if (ui->ui_size != ui->synced_i_size && !ui->dirty) {
+ ubifs_err(c, "ui_size is %lld, synced_i_size is %lld, but inode is clean",
+ ui->ui_size, ui->synced_i_size);
+ ubifs_err(c, "i_ino %lu, i_mode %#x, i_size %lld", inode->i_ino,
+ inode->i_mode, i_size_read(inode));
+ dump_stack();
+ err = -EINVAL;
+ }
+ spin_unlock(&ui->ui_lock);
+ mutex_unlock(&ui->ui_mutex);
+ return err;
+}
+
+/*
+ * dbg_check_dir - check directory inode size and link count.
+ * @c: UBIFS file-system description object
+ * @dir: the directory to calculate size for
+ * @size: the result is returned here
+ *
+ * This function makes sure that directory size and link count are correct.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ *
+ * Note, it is good idea to make sure the @dir->i_mutex is locked before
+ * calling this function.
+ */
+int dbg_check_dir(struct ubifs_info *c, const struct inode *dir)
+{
+ unsigned int nlink = 2;
+ union ubifs_key key;
+ struct ubifs_dent_node *dent, *pdent = NULL;
+ struct fscrypt_name nm = {0};
+ loff_t size = UBIFS_INO_NODE_SZ;
+
+ if (!dbg_is_chk_gen(c))
+ return 0;
+
+ if (!S_ISDIR(dir->i_mode))
+ return 0;
+
+ lowest_dent_key(c, &key, dir->i_ino);
+ while (1) {
+ int err;
+
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ if (err == -ENOENT)
+ break;
+ kfree(pdent);
+ return err;
+ }
+
+ fname_name(&nm) = dent->name;
+ fname_len(&nm) = le16_to_cpu(dent->nlen);
+ size += CALC_DENT_SIZE(fname_len(&nm));
+ if (dent->type == UBIFS_ITYPE_DIR)
+ nlink += 1;
+ kfree(pdent);
+ pdent = dent;
+ key_read(c, &dent->key, &key);
+ }
+ kfree(pdent);
+
+ if (i_size_read(dir) != size) {
+ ubifs_err(c, "directory inode %lu has size %llu, but calculated size is %llu",
+ dir->i_ino, (unsigned long long)i_size_read(dir),
+ (unsigned long long)size);
+ ubifs_dump_inode(c, dir);
+ dump_stack();
+ return -EINVAL;
+ }
+ if (dir->i_nlink != nlink) {
+ ubifs_err(c, "directory inode %lu has nlink %u, but calculated nlink is %u",
+ dir->i_ino, dir->i_nlink, nlink);
+ ubifs_dump_inode(c, dir);
+ dump_stack();
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+/**
+ * dbg_check_key_order - make sure that colliding keys are properly ordered.
+ * @c: UBIFS file-system description object
+ * @zbr1: first zbranch
+ * @zbr2: following zbranch
+ *
+ * In UBIFS indexing B-tree colliding keys has to be sorted in binary order of
+ * names of the direntries/xentries which are referred by the keys. This
+ * function reads direntries/xentries referred by @zbr1 and @zbr2 and makes
+ * sure the name of direntry/xentry referred by @zbr1 is less than
+ * direntry/xentry referred by @zbr2. Returns zero if this is true, %1 if not,
+ * and a negative error code in case of failure.
+ */
+static int dbg_check_key_order(struct ubifs_info *c, struct ubifs_zbranch *zbr1,
+ struct ubifs_zbranch *zbr2)
+{
+ int err, nlen1, nlen2, cmp;
+ struct ubifs_dent_node *dent1, *dent2;
+ union ubifs_key key;
+ char key_buf[DBG_KEY_BUF_LEN];
+
+ ubifs_assert(c, !keys_cmp(c, &zbr1->key, &zbr2->key));
+ dent1 = kmalloc(UBIFS_MAX_DENT_NODE_SZ, GFP_NOFS);
+ if (!dent1)
+ return -ENOMEM;
+ dent2 = kmalloc(UBIFS_MAX_DENT_NODE_SZ, GFP_NOFS);
+ if (!dent2) {
+ err = -ENOMEM;
+ goto out_free;
+ }
+
+ err = ubifs_tnc_read_node(c, zbr1, dent1);
+ if (err)
+ goto out_free;
+ err = ubifs_validate_entry(c, dent1);
+ if (err)
+ goto out_free;
+
+ err = ubifs_tnc_read_node(c, zbr2, dent2);
+ if (err)
+ goto out_free;
+ err = ubifs_validate_entry(c, dent2);
+ if (err)
+ goto out_free;
+
+ /* Make sure node keys are the same as in zbranch */
+ err = 1;
+ key_read(c, &dent1->key, &key);
+ if (keys_cmp(c, &zbr1->key, &key)) {
+ ubifs_err(c, "1st entry at %d:%d has key %s", zbr1->lnum,
+ zbr1->offs, dbg_snprintf_key(c, &key, key_buf,
+ DBG_KEY_BUF_LEN));
+ ubifs_err(c, "but it should have key %s according to tnc",
+ dbg_snprintf_key(c, &zbr1->key, key_buf,
+ DBG_KEY_BUF_LEN));
+ ubifs_dump_node(c, dent1, UBIFS_MAX_DENT_NODE_SZ);
+ goto out_free;
+ }
+
+ key_read(c, &dent2->key, &key);
+ if (keys_cmp(c, &zbr2->key, &key)) {
+ ubifs_err(c, "2nd entry at %d:%d has key %s", zbr1->lnum,
+ zbr1->offs, dbg_snprintf_key(c, &key, key_buf,
+ DBG_KEY_BUF_LEN));
+ ubifs_err(c, "but it should have key %s according to tnc",
+ dbg_snprintf_key(c, &zbr2->key, key_buf,
+ DBG_KEY_BUF_LEN));
+ ubifs_dump_node(c, dent2, UBIFS_MAX_DENT_NODE_SZ);
+ goto out_free;
+ }
+
+ nlen1 = le16_to_cpu(dent1->nlen);
+ nlen2 = le16_to_cpu(dent2->nlen);
+
+ cmp = memcmp(dent1->name, dent2->name, min_t(int, nlen1, nlen2));
+ if (cmp < 0 || (cmp == 0 && nlen1 < nlen2)) {
+ err = 0;
+ goto out_free;
+ }
+ if (cmp == 0 && nlen1 == nlen2)
+ ubifs_err(c, "2 xent/dent nodes with the same name");
+ else
+ ubifs_err(c, "bad order of colliding key %s",
+ dbg_snprintf_key(c, &key, key_buf, DBG_KEY_BUF_LEN));
+
+ ubifs_msg(c, "first node at %d:%d\n", zbr1->lnum, zbr1->offs);
+ ubifs_dump_node(c, dent1, UBIFS_MAX_DENT_NODE_SZ);
+ ubifs_msg(c, "second node at %d:%d\n", zbr2->lnum, zbr2->offs);
+ ubifs_dump_node(c, dent2, UBIFS_MAX_DENT_NODE_SZ);
+
+out_free:
+ kfree(dent2);
+ kfree(dent1);
+ return err;
+}
+
+/**
+ * dbg_check_znode - check if znode is all right.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch which points to this znode
+ *
+ * This function makes sure that znode referred to by @zbr is all right.
+ * Returns zero if it is, and %-EINVAL if it is not.
+ */
+static int dbg_check_znode(struct ubifs_info *c, struct ubifs_zbranch *zbr)
+{
+ struct ubifs_znode *znode = zbr->znode;
+ struct ubifs_znode *zp = znode->parent;
+ int n, err, cmp;
+
+ if (znode->child_cnt <= 0 || znode->child_cnt > c->fanout) {
+ err = 1;
+ goto out;
+ }
+ if (znode->level < 0) {
+ err = 2;
+ goto out;
+ }
+ if (znode->iip < 0 || znode->iip >= c->fanout) {
+ err = 3;
+ goto out;
+ }
+
+ if (zbr->len == 0)
+ /* Only dirty zbranch may have no on-flash nodes */
+ if (!ubifs_zn_dirty(znode)) {
+ err = 4;
+ goto out;
+ }
+
+ if (ubifs_zn_dirty(znode)) {
+ /*
+ * If znode is dirty, its parent has to be dirty as well. The
+ * order of the operation is important, so we have to have
+ * memory barriers.
+ */
+ smp_mb();
+ if (zp && !ubifs_zn_dirty(zp)) {
+ /*
+ * The dirty flag is atomic and is cleared outside the
+ * TNC mutex, so znode's dirty flag may now have
+ * been cleared. The child is always cleared before the
+ * parent, so we just need to check again.
+ */
+ smp_mb();
+ if (ubifs_zn_dirty(znode)) {
+ err = 5;
+ goto out;
+ }
+ }
+ }
+
+ if (zp) {
+ const union ubifs_key *min, *max;
+
+ if (znode->level != zp->level - 1) {
+ err = 6;
+ goto out;
+ }
+
+ /* Make sure the 'parent' pointer in our znode is correct */
+ err = ubifs_search_zbranch(c, zp, &zbr->key, &n);
+ if (!err) {
+ /* This zbranch does not exist in the parent */
+ err = 7;
+ goto out;
+ }
+
+ if (znode->iip >= zp->child_cnt) {
+ err = 8;
+ goto out;
+ }
+
+ if (znode->iip != n) {
+ /* This may happen only in case of collisions */
+ if (keys_cmp(c, &zp->zbranch[n].key,
+ &zp->zbranch[znode->iip].key)) {
+ err = 9;
+ goto out;
+ }
+ n = znode->iip;
+ }
+
+ /*
+ * Make sure that the first key in our znode is greater than or
+ * equal to the key in the pointing zbranch.
+ */
+ min = &zbr->key;
+ cmp = keys_cmp(c, min, &znode->zbranch[0].key);
+ if (cmp == 1) {
+ err = 10;
+ goto out;
+ }
+
+ if (n + 1 < zp->child_cnt) {
+ max = &zp->zbranch[n + 1].key;
+
+ /*
+ * Make sure the last key in our znode is less or
+ * equivalent than the key in the zbranch which goes
+ * after our pointing zbranch.
+ */
+ cmp = keys_cmp(c, max,
+ &znode->zbranch[znode->child_cnt - 1].key);
+ if (cmp == -1) {
+ err = 11;
+ goto out;
+ }
+ }
+ } else {
+ /* This may only be root znode */
+ if (zbr != &c->zroot) {
+ err = 12;
+ goto out;
+ }
+ }
+
+ /*
+ * Make sure that next key is greater or equivalent then the previous
+ * one.
+ */
+ for (n = 1; n < znode->child_cnt; n++) {
+ cmp = keys_cmp(c, &znode->zbranch[n - 1].key,
+ &znode->zbranch[n].key);
+ if (cmp > 0) {
+ err = 13;
+ goto out;
+ }
+ if (cmp == 0) {
+ /* This can only be keys with colliding hash */
+ if (!is_hash_key(c, &znode->zbranch[n].key)) {
+ err = 14;
+ goto out;
+ }
+
+ if (znode->level != 0 || c->replaying)
+ continue;
+
+ /*
+ * Colliding keys should follow binary order of
+ * corresponding xentry/dentry names.
+ */
+ err = dbg_check_key_order(c, &znode->zbranch[n - 1],
+ &znode->zbranch[n]);
+ if (err < 0)
+ return err;
+ if (err) {
+ err = 15;
+ goto out;
+ }
+ }
+ }
+
+ for (n = 0; n < znode->child_cnt; n++) {
+ if (!znode->zbranch[n].znode &&
+ (znode->zbranch[n].lnum == 0 ||
+ znode->zbranch[n].len == 0)) {
+ err = 16;
+ goto out;
+ }
+
+ if (znode->zbranch[n].lnum != 0 &&
+ znode->zbranch[n].len == 0) {
+ err = 17;
+ goto out;
+ }
+
+ if (znode->zbranch[n].lnum == 0 &&
+ znode->zbranch[n].len != 0) {
+ err = 18;
+ goto out;
+ }
+
+ if (znode->zbranch[n].lnum == 0 &&
+ znode->zbranch[n].offs != 0) {
+ err = 19;
+ goto out;
+ }
+
+ if (znode->level != 0 && znode->zbranch[n].znode)
+ if (znode->zbranch[n].znode->parent != znode) {
+ err = 20;
+ goto out;
+ }
+ }
+
+ return 0;
+
+out:
+ ubifs_err(c, "failed, error %d", err);
+ ubifs_msg(c, "dump of the znode");
+ ubifs_dump_znode(c, znode);
+ if (zp) {
+ ubifs_msg(c, "dump of the parent znode");
+ ubifs_dump_znode(c, zp);
+ }
+ dump_stack();
+ return -EINVAL;
+}
+
+/**
+ * dbg_check_tnc - check TNC tree.
+ * @c: UBIFS file-system description object
+ * @extra: do extra checks that are possible at start commit
+ *
+ * This function traverses whole TNC tree and checks every znode. Returns zero
+ * if everything is all right and %-EINVAL if something is wrong with TNC.
+ */
+int dbg_check_tnc(struct ubifs_info *c, int extra)
+{
+ struct ubifs_znode *znode;
+ long clean_cnt = 0, dirty_cnt = 0;
+ int err, last;
+
+ if (!dbg_is_chk_index(c))
+ return 0;
+
+ ubifs_assert(c, mutex_is_locked(&c->tnc_mutex));
+ if (!c->zroot.znode)
+ return 0;
+
+ znode = ubifs_tnc_postorder_first(c->zroot.znode);
+ while (1) {
+ struct ubifs_znode *prev;
+ struct ubifs_zbranch *zbr;
+
+ if (!znode->parent)
+ zbr = &c->zroot;
+ else
+ zbr = &znode->parent->zbranch[znode->iip];
+
+ err = dbg_check_znode(c, zbr);
+ if (err)
+ return err;
+
+ if (extra) {
+ if (ubifs_zn_dirty(znode))
+ dirty_cnt += 1;
+ else
+ clean_cnt += 1;
+ }
+
+ prev = znode;
+ znode = ubifs_tnc_postorder_next(c, znode);
+ if (!znode)
+ break;
+
+ /*
+ * If the last key of this znode is equivalent to the first key
+ * of the next znode (collision), then check order of the keys.
+ */
+ last = prev->child_cnt - 1;
+ if (prev->level == 0 && znode->level == 0 && !c->replaying &&
+ !keys_cmp(c, &prev->zbranch[last].key,
+ &znode->zbranch[0].key)) {
+ err = dbg_check_key_order(c, &prev->zbranch[last],
+ &znode->zbranch[0]);
+ if (err < 0)
+ return err;
+ if (err) {
+ ubifs_msg(c, "first znode");
+ ubifs_dump_znode(c, prev);
+ ubifs_msg(c, "second znode");
+ ubifs_dump_znode(c, znode);
+ return -EINVAL;
+ }
+ }
+ }
+
+ if (extra) {
+ if (clean_cnt != atomic_long_read(&c->clean_zn_cnt)) {
+ ubifs_err(c, "incorrect clean_zn_cnt %ld, calculated %ld",
+ atomic_long_read(&c->clean_zn_cnt),
+ clean_cnt);
+ return -EINVAL;
+ }
+ if (dirty_cnt != atomic_long_read(&c->dirty_zn_cnt)) {
+ ubifs_err(c, "incorrect dirty_zn_cnt %ld, calculated %ld",
+ atomic_long_read(&c->dirty_zn_cnt),
+ dirty_cnt);
+ return -EINVAL;
+ }
+ }
+
+ return 0;
+}
+
+/**
+ * dbg_walk_index - walk the on-flash index.
+ * @c: UBIFS file-system description object
+ * @leaf_cb: called for each leaf node
+ * @znode_cb: called for each indexing node
+ * @priv: private data which is passed to callbacks
+ *
+ * This function walks the UBIFS index and calls the @leaf_cb for each leaf
+ * node and @znode_cb for each indexing node. Returns zero in case of success
+ * and a negative error code in case of failure.
+ *
+ * It would be better if this function removed every znode it pulled to into
+ * the TNC, so that the behavior more closely matched the non-debugging
+ * behavior.
+ */
+int dbg_walk_index(struct ubifs_info *c, dbg_leaf_callback leaf_cb,
+ dbg_znode_callback znode_cb, void *priv)
+{
+ int err;
+ struct ubifs_zbranch *zbr;
+ struct ubifs_znode *znode, *child;
+
+ mutex_lock(&c->tnc_mutex);
+ /* If the root indexing node is not in TNC - pull it */
+ if (!c->zroot.znode) {
+ c->zroot.znode = ubifs_load_znode(c, &c->zroot, NULL, 0);
+ if (IS_ERR(c->zroot.znode)) {
+ err = PTR_ERR(c->zroot.znode);
+ c->zroot.znode = NULL;
+ goto out_unlock;
+ }
+ }
+
+ /*
+ * We are going to traverse the indexing tree in the postorder manner.
+ * Go down and find the leftmost indexing node where we are going to
+ * start from.
+ */
+ znode = c->zroot.znode;
+ while (znode->level > 0) {
+ zbr = &znode->zbranch[0];
+ child = zbr->znode;
+ if (!child) {
+ child = ubifs_load_znode(c, zbr, znode, 0);
+ if (IS_ERR(child)) {
+ err = PTR_ERR(child);
+ goto out_unlock;
+ }
+ }
+
+ znode = child;
+ }
+
+ /* Iterate over all indexing nodes */
+ while (1) {
+ int idx;
+
+ cond_resched();
+
+ if (znode_cb) {
+ err = znode_cb(c, znode, priv);
+ if (err) {
+ ubifs_err(c, "znode checking function returned error %d",
+ err);
+ ubifs_dump_znode(c, znode);
+ goto out_dump;
+ }
+ }
+ if (leaf_cb && znode->level == 0) {
+ for (idx = 0; idx < znode->child_cnt; idx++) {
+ zbr = &znode->zbranch[idx];
+ err = leaf_cb(c, zbr, priv);
+ if (err) {
+ ubifs_err(c, "leaf checking function returned error %d, for leaf at LEB %d:%d",
+ err, zbr->lnum, zbr->offs);
+ goto out_dump;
+ }
+ }
+ }
+
+ if (!znode->parent)
+ break;
+
+ idx = znode->iip + 1;
+ znode = znode->parent;
+ if (idx < znode->child_cnt) {
+ /* Switch to the next index in the parent */
+ zbr = &znode->zbranch[idx];
+ child = zbr->znode;
+ if (!child) {
+ child = ubifs_load_znode(c, zbr, znode, idx);
+ if (IS_ERR(child)) {
+ err = PTR_ERR(child);
+ goto out_unlock;
+ }
+ zbr->znode = child;
+ }
+ znode = child;
+ } else
+ /*
+ * This is the last child, switch to the parent and
+ * continue.
+ */
+ continue;
+
+ /* Go to the lowest leftmost znode in the new sub-tree */
+ while (znode->level > 0) {
+ zbr = &znode->zbranch[0];
+ child = zbr->znode;
+ if (!child) {
+ child = ubifs_load_znode(c, zbr, znode, 0);
+ if (IS_ERR(child)) {
+ err = PTR_ERR(child);
+ goto out_unlock;
+ }
+ zbr->znode = child;
+ }
+ znode = child;
+ }
+ }
+
+ mutex_unlock(&c->tnc_mutex);
+ return 0;
+
+out_dump:
+ if (znode->parent)
+ zbr = &znode->parent->zbranch[znode->iip];
+ else
+ zbr = &c->zroot;
+ ubifs_msg(c, "dump of znode at LEB %d:%d", zbr->lnum, zbr->offs);
+ ubifs_dump_znode(c, znode);
+out_unlock:
+ mutex_unlock(&c->tnc_mutex);
+ return err;
+}
+
+/**
+ * add_size - add znode size to partially calculated index size.
+ * @c: UBIFS file-system description object
+ * @znode: znode to add size for
+ * @priv: partially calculated index size
+ *
+ * This is a helper function for 'dbg_check_idx_size()' which is called for
+ * every indexing node and adds its size to the 'long long' variable pointed to
+ * by @priv.
+ */
+static int add_size(struct ubifs_info *c, struct ubifs_znode *znode, void *priv)
+{
+ long long *idx_size = priv;
+ int add;
+
+ add = ubifs_idx_node_sz(c, znode->child_cnt);
+ add = ALIGN(add, 8);
+ *idx_size += add;
+ return 0;
+}
+
+/**
+ * dbg_check_idx_size - check index size.
+ * @c: UBIFS file-system description object
+ * @idx_size: size to check
+ *
+ * This function walks the UBIFS index, calculates its size and checks that the
+ * size is equivalent to @idx_size. Returns zero in case of success and a
+ * negative error code in case of failure.
+ */
+int dbg_check_idx_size(struct ubifs_info *c, long long idx_size)
+{
+ int err;
+ long long calc = 0;
+
+ if (!dbg_is_chk_index(c))
+ return 0;
+
+ err = dbg_walk_index(c, NULL, add_size, &calc);
+ if (err) {
+ ubifs_err(c, "error %d while walking the index", err);
+ goto out_err;
+ }
+
+ if (calc != idx_size) {
+ ubifs_err(c, "index size check failed: calculated size is %lld, should be %lld",
+ calc, idx_size);
+ dump_stack();
+ err = -EINVAL;
+ goto out_err;
+ }
+
+ return 0;
+
+out_err:
+ ubifs_destroy_tnc_tree(c);
+ return err;
+}
+
+/**
+ * struct fsck_inode - information about an inode used when checking the file-system.
+ * @rb: link in the RB-tree of inodes
+ * @inum: inode number
+ * @mode: inode type, permissions, etc
+ * @nlink: inode link count
+ * @xattr_cnt: count of extended attributes
+ * @references: how many directory/xattr entries refer this inode (calculated
+ * while walking the index)
+ * @calc_cnt: for directory inode count of child directories
+ * @size: inode size (read from on-flash inode)
+ * @xattr_sz: summary size of all extended attributes (read from on-flash
+ * inode)
+ * @calc_sz: for directories calculated directory size
+ * @calc_xcnt: count of extended attributes
+ * @calc_xsz: calculated summary size of all extended attributes
+ * @xattr_nms: sum of lengths of all extended attribute names belonging to this
+ * inode (read from on-flash inode)
+ * @calc_xnms: calculated sum of lengths of all extended attribute names
+ */
+struct fsck_inode {
+ struct rb_node rb;
+ ino_t inum;
+ umode_t mode;
+ unsigned int nlink;
+ unsigned int xattr_cnt;
+ int references;
+ int calc_cnt;
+ long long size;
+ unsigned int xattr_sz;
+ long long calc_sz;
+ long long calc_xcnt;
+ long long calc_xsz;
+ unsigned int xattr_nms;
+ long long calc_xnms;
+};
+
+/**
+ * struct fsck_data - private FS checking information.
+ * @inodes: RB-tree of all inodes (contains @struct fsck_inode objects)
+ */
+struct fsck_data {
+ struct rb_root inodes;
+};
+
+/**
+ * add_inode - add inode information to RB-tree of inodes.
+ * @c: UBIFS file-system description object
+ * @fsckd: FS checking information
+ * @ino: raw UBIFS inode to add
+ *
+ * This is a helper function for 'check_leaf()' which adds information about
+ * inode @ino to the RB-tree of inodes. Returns inode information pointer in
+ * case of success and a negative error code in case of failure.
+ */
+static struct fsck_inode *add_inode(struct ubifs_info *c,
+ struct fsck_data *fsckd,
+ struct ubifs_ino_node *ino)
+{
+ struct rb_node **p, *parent = NULL;
+ struct fsck_inode *fscki;
+ ino_t inum = key_inum_flash(c, &ino->key);
+ struct inode *inode;
+ struct ubifs_inode *ui;
+
+ p = &fsckd->inodes.rb_node;
+ while (*p) {
+ parent = *p;
+ fscki = rb_entry(parent, struct fsck_inode, rb);
+ if (inum < fscki->inum)
+ p = &(*p)->rb_left;
+ else if (inum > fscki->inum)
+ p = &(*p)->rb_right;
+ else
+ return fscki;
+ }
+
+ if (inum > c->highest_inum) {
+ ubifs_err(c, "too high inode number, max. is %lu",
+ (unsigned long)c->highest_inum);
+ return ERR_PTR(-EINVAL);
+ }
+
+ fscki = kzalloc(sizeof(struct fsck_inode), GFP_NOFS);
+ if (!fscki)
+ return ERR_PTR(-ENOMEM);
+
+ inode = ilookup(c->vfs_sb, inum);
+
+ fscki->inum = inum;
+ /*
+ * If the inode is present in the VFS inode cache, use it instead of
+ * the on-flash inode which might be out-of-date. E.g., the size might
+ * be out-of-date. If we do not do this, the following may happen, for
+ * example:
+ * 1. A power cut happens
+ * 2. We mount the file-system R/O, the replay process fixes up the
+ * inode size in the VFS cache, but on on-flash.
+ * 3. 'check_leaf()' fails because it hits a data node beyond inode
+ * size.
+ */
+ if (!inode) {
+ fscki->nlink = le32_to_cpu(ino->nlink);
+ fscki->size = le64_to_cpu(ino->size);
+ fscki->xattr_cnt = le32_to_cpu(ino->xattr_cnt);
+ fscki->xattr_sz = le32_to_cpu(ino->xattr_size);
+ fscki->xattr_nms = le32_to_cpu(ino->xattr_names);
+ fscki->mode = le32_to_cpu(ino->mode);
+ } else {
+ ui = ubifs_inode(inode);
+ fscki->nlink = inode->i_nlink;
+ fscki->size = inode->i_size;
+ fscki->xattr_cnt = ui->xattr_cnt;
+ fscki->xattr_sz = ui->xattr_size;
+ fscki->xattr_nms = ui->xattr_names;
+ fscki->mode = inode->i_mode;
+ iput(inode);
+ }
+
+ if (S_ISDIR(fscki->mode)) {
+ fscki->calc_sz = UBIFS_INO_NODE_SZ;
+ fscki->calc_cnt = 2;
+ }
+
+ rb_link_node(&fscki->rb, parent, p);
+ rb_insert_color(&fscki->rb, &fsckd->inodes);
+
+ return fscki;
+}
+
+/**
+ * search_inode - search inode in the RB-tree of inodes.
+ * @fsckd: FS checking information
+ * @inum: inode number to search
+ *
+ * This is a helper function for 'check_leaf()' which searches inode @inum in
+ * the RB-tree of inodes and returns an inode information pointer or %NULL if
+ * the inode was not found.
+ */
+static struct fsck_inode *search_inode(struct fsck_data *fsckd, ino_t inum)
+{
+ struct rb_node *p;
+ struct fsck_inode *fscki;
+
+ p = fsckd->inodes.rb_node;
+ while (p) {
+ fscki = rb_entry(p, struct fsck_inode, rb);
+ if (inum < fscki->inum)
+ p = p->rb_left;
+ else if (inum > fscki->inum)
+ p = p->rb_right;
+ else
+ return fscki;
+ }
+ return NULL;
+}
+
+/**
+ * read_add_inode - read inode node and add it to RB-tree of inodes.
+ * @c: UBIFS file-system description object
+ * @fsckd: FS checking information
+ * @inum: inode number to read
+ *
+ * This is a helper function for 'check_leaf()' which finds inode node @inum in
+ * the index, reads it, and adds it to the RB-tree of inodes. Returns inode
+ * information pointer in case of success and a negative error code in case of
+ * failure.
+ */
+static struct fsck_inode *read_add_inode(struct ubifs_info *c,
+ struct fsck_data *fsckd, ino_t inum)
+{
+ int n, err;
+ union ubifs_key key;
+ struct ubifs_znode *znode;
+ struct ubifs_zbranch *zbr;
+ struct ubifs_ino_node *ino;
+ struct fsck_inode *fscki;
+
+ fscki = search_inode(fsckd, inum);
+ if (fscki)
+ return fscki;
+
+ ino_key_init(c, &key, inum);
+ err = ubifs_lookup_level0(c, &key, &znode, &n);
+ if (!err) {
+ ubifs_err(c, "inode %lu not found in index", (unsigned long)inum);
+ return ERR_PTR(-ENOENT);
+ } else if (err < 0) {
+ ubifs_err(c, "error %d while looking up inode %lu",
+ err, (unsigned long)inum);
+ return ERR_PTR(err);
+ }
+
+ zbr = &znode->zbranch[n];
+ if (zbr->len < UBIFS_INO_NODE_SZ) {
+ ubifs_err(c, "bad node %lu node length %d",
+ (unsigned long)inum, zbr->len);
+ return ERR_PTR(-EINVAL);
+ }
+
+ ino = kmalloc(zbr->len, GFP_NOFS);
+ if (!ino)
+ return ERR_PTR(-ENOMEM);
+
+ err = ubifs_tnc_read_node(c, zbr, ino);
+ if (err) {
+ ubifs_err(c, "cannot read inode node at LEB %d:%d, error %d",
+ zbr->lnum, zbr->offs, err);
+ kfree(ino);
+ return ERR_PTR(err);
+ }
+
+ fscki = add_inode(c, fsckd, ino);
+ kfree(ino);
+ if (IS_ERR(fscki)) {
+ ubifs_err(c, "error %ld while adding inode %lu node",
+ PTR_ERR(fscki), (unsigned long)inum);
+ return fscki;
+ }
+
+ return fscki;
+}
+
+/**
+ * check_leaf - check leaf node.
+ * @c: UBIFS file-system description object
+ * @zbr: zbranch of the leaf node to check
+ * @priv: FS checking information
+ *
+ * This is a helper function for 'dbg_check_filesystem()' which is called for
+ * every single leaf node while walking the indexing tree. It checks that the
+ * leaf node referred from the indexing tree exists, has correct CRC, and does
+ * some other basic validation. This function is also responsible for building
+ * an RB-tree of inodes - it adds all inodes into the RB-tree. It also
+ * calculates reference count, size, etc for each inode in order to later
+ * compare them to the information stored inside the inodes and detect possible
+ * inconsistencies. Returns zero in case of success and a negative error code
+ * in case of failure.
+ */
+static int check_leaf(struct ubifs_info *c, struct ubifs_zbranch *zbr,
+ void *priv)
+{
+ ino_t inum;
+ void *node;
+ struct ubifs_ch *ch;
+ int err, type = key_type(c, &zbr->key);
+ struct fsck_inode *fscki;
+
+ if (zbr->len < UBIFS_CH_SZ) {
+ ubifs_err(c, "bad leaf length %d (LEB %d:%d)",
+ zbr->len, zbr->lnum, zbr->offs);
+ return -EINVAL;
+ }
+
+ node = kmalloc(zbr->len, GFP_NOFS);
+ if (!node)
+ return -ENOMEM;
+
+ err = ubifs_tnc_read_node(c, zbr, node);
+ if (err) {
+ ubifs_err(c, "cannot read leaf node at LEB %d:%d, error %d",
+ zbr->lnum, zbr->offs, err);
+ goto out_free;
+ }
+
+ /* If this is an inode node, add it to RB-tree of inodes */
+ if (type == UBIFS_INO_KEY) {
+ fscki = add_inode(c, priv, node);
+ if (IS_ERR(fscki)) {
+ err = PTR_ERR(fscki);
+ ubifs_err(c, "error %d while adding inode node", err);
+ goto out_dump;
+ }
+ goto out;
+ }
+
+ if (type != UBIFS_DENT_KEY && type != UBIFS_XENT_KEY &&
+ type != UBIFS_DATA_KEY) {
+ ubifs_err(c, "unexpected node type %d at LEB %d:%d",
+ type, zbr->lnum, zbr->offs);
+ err = -EINVAL;
+ goto out_free;
+ }
+
+ ch = node;
+ if (le64_to_cpu(ch->sqnum) > c->max_sqnum) {
+ ubifs_err(c, "too high sequence number, max. is %llu",
+ c->max_sqnum);
+ err = -EINVAL;
+ goto out_dump;
+ }
+
+ if (type == UBIFS_DATA_KEY) {
+ long long blk_offs;
+ struct ubifs_data_node *dn = node;
+
+ ubifs_assert(c, zbr->len >= UBIFS_DATA_NODE_SZ);
+
+ /*
+ * Search the inode node this data node belongs to and insert
+ * it to the RB-tree of inodes.
+ */
+ inum = key_inum_flash(c, &dn->key);
+ fscki = read_add_inode(c, priv, inum);
+ if (IS_ERR(fscki)) {
+ err = PTR_ERR(fscki);
+ ubifs_err(c, "error %d while processing data node and trying to find inode node %lu",
+ err, (unsigned long)inum);
+ goto out_dump;
+ }
+
+ /* Make sure the data node is within inode size */
+ blk_offs = key_block_flash(c, &dn->key);
+ blk_offs <<= UBIFS_BLOCK_SHIFT;
+ blk_offs += le32_to_cpu(dn->size);
+ if (blk_offs > fscki->size) {
+ ubifs_err(c, "data node at LEB %d:%d is not within inode size %lld",
+ zbr->lnum, zbr->offs, fscki->size);
+ err = -EINVAL;
+ goto out_dump;
+ }
+ } else {
+ int nlen;
+ struct ubifs_dent_node *dent = node;
+ struct fsck_inode *fscki1;
+
+ ubifs_assert(c, zbr->len >= UBIFS_DENT_NODE_SZ);
+
+ err = ubifs_validate_entry(c, dent);
+ if (err)
+ goto out_dump;
+
+ /*
+ * Search the inode node this entry refers to and the parent
+ * inode node and insert them to the RB-tree of inodes.
+ */
+ inum = le64_to_cpu(dent->inum);
+ fscki = read_add_inode(c, priv, inum);
+ if (IS_ERR(fscki)) {
+ err = PTR_ERR(fscki);
+ ubifs_err(c, "error %d while processing entry node and trying to find inode node %lu",
+ err, (unsigned long)inum);
+ goto out_dump;
+ }
+
+ /* Count how many direntries or xentries refers this inode */
+ fscki->references += 1;
+
+ inum = key_inum_flash(c, &dent->key);
+ fscki1 = read_add_inode(c, priv, inum);
+ if (IS_ERR(fscki1)) {
+ err = PTR_ERR(fscki1);
+ ubifs_err(c, "error %d while processing entry node and trying to find parent inode node %lu",
+ err, (unsigned long)inum);
+ goto out_dump;
+ }
+
+ nlen = le16_to_cpu(dent->nlen);
+ if (type == UBIFS_XENT_KEY) {
+ fscki1->calc_xcnt += 1;
+ fscki1->calc_xsz += CALC_DENT_SIZE(nlen);
+ fscki1->calc_xsz += CALC_XATTR_BYTES(fscki->size);
+ fscki1->calc_xnms += nlen;
+ } else {
+ fscki1->calc_sz += CALC_DENT_SIZE(nlen);
+ if (dent->type == UBIFS_ITYPE_DIR)
+ fscki1->calc_cnt += 1;
+ }
+ }
+
+out:
+ kfree(node);
+ return 0;
+
+out_dump:
+ ubifs_msg(c, "dump of node at LEB %d:%d", zbr->lnum, zbr->offs);
+ ubifs_dump_node(c, node, zbr->len);
+out_free:
+ kfree(node);
+ return err;
+}
+
+/**
+ * free_inodes - free RB-tree of inodes.
+ * @fsckd: FS checking information
+ */
+static void free_inodes(struct fsck_data *fsckd)
+{
+ struct fsck_inode *fscki, *n;
+
+ rbtree_postorder_for_each_entry_safe(fscki, n, &fsckd->inodes, rb)
+ kfree(fscki);
+}
+
+/**
+ * check_inodes - checks all inodes.
+ * @c: UBIFS file-system description object
+ * @fsckd: FS checking information
+ *
+ * This is a helper function for 'dbg_check_filesystem()' which walks the
+ * RB-tree of inodes after the index scan has been finished, and checks that
+ * inode nlink, size, etc are correct. Returns zero if inodes are fine,
+ * %-EINVAL if not, and a negative error code in case of failure.
+ */
+static int check_inodes(struct ubifs_info *c, struct fsck_data *fsckd)
+{
+ int n, err;
+ union ubifs_key key;
+ struct ubifs_znode *znode;
+ struct ubifs_zbranch *zbr;
+ struct ubifs_ino_node *ino;
+ struct fsck_inode *fscki;
+ struct rb_node *this = rb_first(&fsckd->inodes);
+
+ while (this) {
+ fscki = rb_entry(this, struct fsck_inode, rb);
+ this = rb_next(this);
+
+ if (S_ISDIR(fscki->mode)) {
+ /*
+ * Directories have to have exactly one reference (they
+ * cannot have hardlinks), although root inode is an
+ * exception.
+ */
+ if (fscki->inum != UBIFS_ROOT_INO &&
+ fscki->references != 1) {
+ ubifs_err(c, "directory inode %lu has %d direntries which refer it, but should be 1",
+ (unsigned long)fscki->inum,
+ fscki->references);
+ goto out_dump;
+ }
+ if (fscki->inum == UBIFS_ROOT_INO &&
+ fscki->references != 0) {
+ ubifs_err(c, "root inode %lu has non-zero (%d) direntries which refer it",
+ (unsigned long)fscki->inum,
+ fscki->references);
+ goto out_dump;
+ }
+ if (fscki->calc_sz != fscki->size) {
+ ubifs_err(c, "directory inode %lu size is %lld, but calculated size is %lld",
+ (unsigned long)fscki->inum,
+ fscki->size, fscki->calc_sz);
+ goto out_dump;
+ }
+ if (fscki->calc_cnt != fscki->nlink) {
+ ubifs_err(c, "directory inode %lu nlink is %d, but calculated nlink is %d",
+ (unsigned long)fscki->inum,
+ fscki->nlink, fscki->calc_cnt);
+ goto out_dump;
+ }
+ } else {
+ if (fscki->references != fscki->nlink) {
+ ubifs_err(c, "inode %lu nlink is %d, but calculated nlink is %d",
+ (unsigned long)fscki->inum,
+ fscki->nlink, fscki->references);
+ goto out_dump;
+ }
+ }
+ if (fscki->xattr_sz != fscki->calc_xsz) {
+ ubifs_err(c, "inode %lu has xattr size %u, but calculated size is %lld",
+ (unsigned long)fscki->inum, fscki->xattr_sz,
+ fscki->calc_xsz);
+ goto out_dump;
+ }
+ if (fscki->xattr_cnt != fscki->calc_xcnt) {
+ ubifs_err(c, "inode %lu has %u xattrs, but calculated count is %lld",
+ (unsigned long)fscki->inum,
+ fscki->xattr_cnt, fscki->calc_xcnt);
+ goto out_dump;
+ }
+ if (fscki->xattr_nms != fscki->calc_xnms) {
+ ubifs_err(c, "inode %lu has xattr names' size %u, but calculated names' size is %lld",
+ (unsigned long)fscki->inum, fscki->xattr_nms,
+ fscki->calc_xnms);
+ goto out_dump;
+ }
+ }
+
+ return 0;
+
+out_dump:
+ /* Read the bad inode and dump it */
+ ino_key_init(c, &key, fscki->inum);
+ err = ubifs_lookup_level0(c, &key, &znode, &n);
+ if (!err) {
+ ubifs_err(c, "inode %lu not found in index",
+ (unsigned long)fscki->inum);
+ return -ENOENT;
+ } else if (err < 0) {
+ ubifs_err(c, "error %d while looking up inode %lu",
+ err, (unsigned long)fscki->inum);
+ return err;
+ }
+
+ zbr = &znode->zbranch[n];
+ ino = kmalloc(zbr->len, GFP_NOFS);
+ if (!ino)
+ return -ENOMEM;
+
+ err = ubifs_tnc_read_node(c, zbr, ino);
+ if (err) {
+ ubifs_err(c, "cannot read inode node at LEB %d:%d, error %d",
+ zbr->lnum, zbr->offs, err);
+ kfree(ino);
+ return err;
+ }
+
+ ubifs_msg(c, "dump of the inode %lu sitting in LEB %d:%d",
+ (unsigned long)fscki->inum, zbr->lnum, zbr->offs);
+ ubifs_dump_node(c, ino, zbr->len);
+ kfree(ino);
+ return -EINVAL;
+}
+
+/**
+ * dbg_check_filesystem - check the file-system.
+ * @c: UBIFS file-system description object
+ *
+ * This function checks the file system, namely:
+ * o makes sure that all leaf nodes exist and their CRCs are correct;
+ * o makes sure inode nlink, size, xattr size/count are correct (for all
+ * inodes).
+ *
+ * The function reads whole indexing tree and all nodes, so it is pretty
+ * heavy-weight. Returns zero if the file-system is consistent, %-EINVAL if
+ * not, and a negative error code in case of failure.
+ */
+int dbg_check_filesystem(struct ubifs_info *c)
+{
+ int err;
+ struct fsck_data fsckd;
+
+ if (!dbg_is_chk_fs(c))
+ return 0;
+
+ fsckd.inodes = RB_ROOT;
+ err = dbg_walk_index(c, check_leaf, NULL, &fsckd);
+ if (err)
+ goto out_free;
+
+ err = check_inodes(c, &fsckd);
+ if (err)
+ goto out_free;
+
+ free_inodes(&fsckd);
+ return 0;
+
+out_free:
+ ubifs_err(c, "file-system check failed with error %d", err);
+ dump_stack();
+ free_inodes(&fsckd);
+ return err;
+}
+
+/**
+ * dbg_check_data_nodes_order - check that list of data nodes is sorted.
+ * @c: UBIFS file-system description object
+ * @head: the list of nodes ('struct ubifs_scan_node' objects)
+ *
+ * This function returns zero if the list of data nodes is sorted correctly,
+ * and %-EINVAL if not.
+ */
+int dbg_check_data_nodes_order(struct ubifs_info *c, struct list_head *head)
+{
+ struct list_head *cur;
+ struct ubifs_scan_node *sa, *sb;
+
+ if (!dbg_is_chk_gen(c))
+ return 0;
+
+ for (cur = head->next; cur->next != head; cur = cur->next) {
+ ino_t inuma, inumb;
+ uint32_t blka, blkb;
+
+ cond_resched();
+ sa = container_of(cur, struct ubifs_scan_node, list);
+ sb = container_of(cur->next, struct ubifs_scan_node, list);
+
+ if (sa->type != UBIFS_DATA_NODE) {
+ ubifs_err(c, "bad node type %d", sa->type);
+ ubifs_dump_node(c, sa->node, c->leb_size - sa->offs);
+ return -EINVAL;
+ }
+ if (sb->type != UBIFS_DATA_NODE) {
+ ubifs_err(c, "bad node type %d", sb->type);
+ ubifs_dump_node(c, sb->node, c->leb_size - sb->offs);
+ return -EINVAL;
+ }
+
+ inuma = key_inum(c, &sa->key);
+ inumb = key_inum(c, &sb->key);
+
+ if (inuma < inumb)
+ continue;
+ if (inuma > inumb) {
+ ubifs_err(c, "larger inum %lu goes before inum %lu",
+ (unsigned long)inuma, (unsigned long)inumb);
+ goto error_dump;
+ }
+
+ blka = key_block(c, &sa->key);
+ blkb = key_block(c, &sb->key);
+
+ if (blka > blkb) {
+ ubifs_err(c, "larger block %u goes before %u", blka, blkb);
+ goto error_dump;
+ }
+ if (blka == blkb) {
+ ubifs_err(c, "two data nodes for the same block");
+ goto error_dump;
+ }
+ }
+
+ return 0;
+
+error_dump:
+ ubifs_dump_node(c, sa->node, c->leb_size - sa->offs);
+ ubifs_dump_node(c, sb->node, c->leb_size - sb->offs);
+ return -EINVAL;
+}
+
+/**
+ * dbg_check_nondata_nodes_order - check that list of data nodes is sorted.
+ * @c: UBIFS file-system description object
+ * @head: the list of nodes ('struct ubifs_scan_node' objects)
+ *
+ * This function returns zero if the list of non-data nodes is sorted correctly,
+ * and %-EINVAL if not.
+ */
+int dbg_check_nondata_nodes_order(struct ubifs_info *c, struct list_head *head)
+{
+ struct list_head *cur;
+ struct ubifs_scan_node *sa, *sb;
+
+ if (!dbg_is_chk_gen(c))
+ return 0;
+
+ for (cur = head->next; cur->next != head; cur = cur->next) {
+ ino_t inuma, inumb;
+ uint32_t hasha, hashb;
+
+ cond_resched();
+ sa = container_of(cur, struct ubifs_scan_node, list);
+ sb = container_of(cur->next, struct ubifs_scan_node, list);
+
+ if (sa->type != UBIFS_INO_NODE && sa->type != UBIFS_DENT_NODE &&
+ sa->type != UBIFS_XENT_NODE) {
+ ubifs_err(c, "bad node type %d", sa->type);
+ ubifs_dump_node(c, sa->node, c->leb_size - sa->offs);
+ return -EINVAL;
+ }
+ if (sb->type != UBIFS_INO_NODE && sb->type != UBIFS_DENT_NODE &&
+ sb->type != UBIFS_XENT_NODE) {
+ ubifs_err(c, "bad node type %d", sb->type);
+ ubifs_dump_node(c, sb->node, c->leb_size - sb->offs);
+ return -EINVAL;
+ }
+
+ if (sa->type != UBIFS_INO_NODE && sb->type == UBIFS_INO_NODE) {
+ ubifs_err(c, "non-inode node goes before inode node");
+ goto error_dump;
+ }
+
+ if (sa->type == UBIFS_INO_NODE && sb->type != UBIFS_INO_NODE)
+ continue;
+
+ if (sa->type == UBIFS_INO_NODE && sb->type == UBIFS_INO_NODE) {
+ /* Inode nodes are sorted in descending size order */
+ if (sa->len < sb->len) {
+ ubifs_err(c, "smaller inode node goes first");
+ goto error_dump;
+ }
+ continue;
+ }
+
+ /*
+ * This is either a dentry or xentry, which should be sorted in
+ * ascending (parent ino, hash) order.
+ */
+ inuma = key_inum(c, &sa->key);
+ inumb = key_inum(c, &sb->key);
+
+ if (inuma < inumb)
+ continue;
+ if (inuma > inumb) {
+ ubifs_err(c, "larger inum %lu goes before inum %lu",
+ (unsigned long)inuma, (unsigned long)inumb);
+ goto error_dump;
+ }
+
+ hasha = key_block(c, &sa->key);
+ hashb = key_block(c, &sb->key);
+
+ if (hasha > hashb) {
+ ubifs_err(c, "larger hash %u goes before %u",
+ hasha, hashb);
+ goto error_dump;
+ }
+ }
+
+ return 0;
+
+error_dump:
+ ubifs_msg(c, "dumping first node");
+ ubifs_dump_node(c, sa->node, c->leb_size - sa->offs);
+ ubifs_msg(c, "dumping second node");
+ ubifs_dump_node(c, sb->node, c->leb_size - sb->offs);
+ return -EINVAL;
+}
+
+static inline int chance(unsigned int n, unsigned int out_of)
+{
+ return !!(get_random_u32_below(out_of) + 1 <= n);
+
+}
+
+static int power_cut_emulated(struct ubifs_info *c, int lnum, int write)
+{
+ struct ubifs_debug_info *d = c->dbg;
+
+ ubifs_assert(c, dbg_is_tst_rcvry(c));
+
+ if (!d->pc_cnt) {
+ /* First call - decide delay to the power cut */
+ if (chance(1, 2)) {
+ unsigned long delay;
+
+ if (chance(1, 2)) {
+ d->pc_delay = 1;
+ /* Fail within 1 minute */
+ delay = get_random_u32_below(60000);
+ d->pc_timeout = jiffies;
+ d->pc_timeout += msecs_to_jiffies(delay);
+ ubifs_warn(c, "failing after %lums", delay);
+ } else {
+ d->pc_delay = 2;
+ delay = get_random_u32_below(10000);
+ /* Fail within 10000 operations */
+ d->pc_cnt_max = delay;
+ ubifs_warn(c, "failing after %lu calls", delay);
+ }
+ }
+
+ d->pc_cnt += 1;
+ }
+
+ /* Determine if failure delay has expired */
+ if (d->pc_delay == 1 && time_before(jiffies, d->pc_timeout))
+ return 0;
+ if (d->pc_delay == 2 && d->pc_cnt++ < d->pc_cnt_max)
+ return 0;
+
+ if (lnum == UBIFS_SB_LNUM) {
+ if (write && chance(1, 2))
+ return 0;
+ if (chance(19, 20))
+ return 0;
+ ubifs_warn(c, "failing in super block LEB %d", lnum);
+ } else if (lnum == UBIFS_MST_LNUM || lnum == UBIFS_MST_LNUM + 1) {
+ if (chance(19, 20))
+ return 0;
+ ubifs_warn(c, "failing in master LEB %d", lnum);
+ } else if (lnum >= UBIFS_LOG_LNUM && lnum <= c->log_last) {
+ if (write && chance(99, 100))
+ return 0;
+ if (chance(399, 400))
+ return 0;
+ ubifs_warn(c, "failing in log LEB %d", lnum);
+ } else if (lnum >= c->lpt_first && lnum <= c->lpt_last) {
+ if (write && chance(7, 8))
+ return 0;
+ if (chance(19, 20))
+ return 0;
+ ubifs_warn(c, "failing in LPT LEB %d", lnum);
+ } else if (lnum >= c->orph_first && lnum <= c->orph_last) {
+ if (write && chance(1, 2))
+ return 0;
+ if (chance(9, 10))
+ return 0;
+ ubifs_warn(c, "failing in orphan LEB %d", lnum);
+ } else if (lnum == c->ihead_lnum) {
+ if (chance(99, 100))
+ return 0;
+ ubifs_warn(c, "failing in index head LEB %d", lnum);
+ } else if (c->jheads && lnum == c->jheads[GCHD].wbuf.lnum) {
+ if (chance(9, 10))
+ return 0;
+ ubifs_warn(c, "failing in GC head LEB %d", lnum);
+ } else if (write && !RB_EMPTY_ROOT(&c->buds) &&
+ !ubifs_search_bud(c, lnum)) {
+ if (chance(19, 20))
+ return 0;
+ ubifs_warn(c, "failing in non-bud LEB %d", lnum);
+ } else if (c->cmt_state == COMMIT_RUNNING_BACKGROUND ||
+ c->cmt_state == COMMIT_RUNNING_REQUIRED) {
+ if (chance(999, 1000))
+ return 0;
+ ubifs_warn(c, "failing in bud LEB %d commit running", lnum);
+ } else {
+ if (chance(9999, 10000))
+ return 0;
+ ubifs_warn(c, "failing in bud LEB %d commit not running", lnum);
+ }
+
+ d->pc_happened = 1;
+ ubifs_warn(c, "========== Power cut emulated ==========");
+ dump_stack();
+ return 1;
+}
+
+static int corrupt_data(const struct ubifs_info *c, const void *buf,
+ unsigned int len)
+{
+ unsigned int from, to, ffs = chance(1, 2);
+ unsigned char *p = (void *)buf;
+
+ from = get_random_u32_below(len);
+ /* Corruption span max to end of write unit */
+ to = min(len, ALIGN(from + 1, c->max_write_size));
+
+ ubifs_warn(c, "filled bytes %u-%u with %s", from, to - 1,
+ ffs ? "0xFFs" : "random data");
+
+ if (ffs)
+ memset(p + from, 0xFF, to - from);
+ else
+ get_random_bytes(p + from, to - from);
+
+ return to;
+}
+
+int dbg_leb_write(struct ubifs_info *c, int lnum, const void *buf,
+ int offs, int len)
+{
+ int err, failing;
+
+ if (dbg_is_power_cut(c))
+ return -EROFS;
+
+ failing = power_cut_emulated(c, lnum, 1);
+ if (failing) {
+ len = corrupt_data(c, buf, len);
+ ubifs_warn(c, "actually write %d bytes to LEB %d:%d (the buffer was corrupted)",
+ len, lnum, offs);
+ }
+ err = ubi_leb_write(c->ubi, lnum, buf, offs, len);
+ if (err)
+ return err;
+ if (failing)
+ return -EROFS;
+ return 0;
+}
+
+int dbg_leb_change(struct ubifs_info *c, int lnum, const void *buf,
+ int len)
+{
+ int err;
+
+ if (dbg_is_power_cut(c))
+ return -EROFS;
+ if (power_cut_emulated(c, lnum, 1))
+ return -EROFS;
+ err = ubi_leb_change(c->ubi, lnum, buf, len);
+ if (err)
+ return err;
+ if (power_cut_emulated(c, lnum, 1))
+ return -EROFS;
+ return 0;
+}
+
+int dbg_leb_unmap(struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ if (dbg_is_power_cut(c))
+ return -EROFS;
+ if (power_cut_emulated(c, lnum, 0))
+ return -EROFS;
+ err = ubi_leb_unmap(c->ubi, lnum);
+ if (err)
+ return err;
+ if (power_cut_emulated(c, lnum, 0))
+ return -EROFS;
+ return 0;
+}
+
+int dbg_leb_map(struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ if (dbg_is_power_cut(c))
+ return -EROFS;
+ if (power_cut_emulated(c, lnum, 0))
+ return -EROFS;
+ err = ubi_leb_map(c->ubi, lnum);
+ if (err)
+ return err;
+ if (power_cut_emulated(c, lnum, 0))
+ return -EROFS;
+ return 0;
+}
+
+/*
+ * Root directory for UBIFS stuff in debugfs. Contains sub-directories which
+ * contain the stuff specific to particular file-system mounts.
+ */
+static struct dentry *dfs_rootdir;
+
+static int dfs_file_open(struct inode *inode, struct file *file)
+{
+ file->private_data = inode->i_private;
+ return nonseekable_open(inode, file);
+}
+
+/**
+ * provide_user_output - provide output to the user reading a debugfs file.
+ * @val: boolean value for the answer
+ * @u: the buffer to store the answer at
+ * @count: size of the buffer
+ * @ppos: position in the @u output buffer
+ *
+ * This is a simple helper function which stores @val boolean value in the user
+ * buffer when the user reads one of UBIFS debugfs files. Returns amount of
+ * bytes written to @u in case of success and a negative error code in case of
+ * failure.
+ */
+static int provide_user_output(int val, char __user *u, size_t count,
+ loff_t *ppos)
+{
+ char buf[3];
+
+ if (val)
+ buf[0] = '1';
+ else
+ buf[0] = '0';
+ buf[1] = '\n';
+ buf[2] = 0x00;
+
+ return simple_read_from_buffer(u, count, ppos, buf, 2);
+}
+
+static ssize_t dfs_file_read(struct file *file, char __user *u, size_t count,
+ loff_t *ppos)
+{
+ struct dentry *dent = file->f_path.dentry;
+ struct ubifs_info *c = file->private_data;
+ struct ubifs_debug_info *d = c->dbg;
+ int val;
+
+ if (dent == d->dfs_chk_gen)
+ val = d->chk_gen;
+ else if (dent == d->dfs_chk_index)
+ val = d->chk_index;
+ else if (dent == d->dfs_chk_orph)
+ val = d->chk_orph;
+ else if (dent == d->dfs_chk_lprops)
+ val = d->chk_lprops;
+ else if (dent == d->dfs_chk_fs)
+ val = d->chk_fs;
+ else if (dent == d->dfs_tst_rcvry)
+ val = d->tst_rcvry;
+ else if (dent == d->dfs_ro_error)
+ val = c->ro_error;
+ else
+ return -EINVAL;
+
+ return provide_user_output(val, u, count, ppos);
+}
+
+/**
+ * interpret_user_input - interpret user debugfs file input.
+ * @u: user-provided buffer with the input
+ * @count: buffer size
+ *
+ * This is a helper function which interpret user input to a boolean UBIFS
+ * debugfs file. Returns %0 or %1 in case of success and a negative error code
+ * in case of failure.
+ */
+static int interpret_user_input(const char __user *u, size_t count)
+{
+ size_t buf_size;
+ char buf[8];
+
+ buf_size = min_t(size_t, count, (sizeof(buf) - 1));
+ if (copy_from_user(buf, u, buf_size))
+ return -EFAULT;
+
+ if (buf[0] == '1')
+ return 1;
+ else if (buf[0] == '0')
+ return 0;
+
+ return -EINVAL;
+}
+
+static ssize_t dfs_file_write(struct file *file, const char __user *u,
+ size_t count, loff_t *ppos)
+{
+ struct ubifs_info *c = file->private_data;
+ struct ubifs_debug_info *d = c->dbg;
+ struct dentry *dent = file->f_path.dentry;
+ int val;
+
+ if (file->f_path.dentry == d->dfs_dump_lprops) {
+ ubifs_dump_lprops(c);
+ return count;
+ }
+ if (file->f_path.dentry == d->dfs_dump_budg) {
+ ubifs_dump_budg(c, &c->bi);
+ return count;
+ }
+ if (file->f_path.dentry == d->dfs_dump_tnc) {
+ mutex_lock(&c->tnc_mutex);
+ ubifs_dump_tnc(c);
+ mutex_unlock(&c->tnc_mutex);
+ return count;
+ }
+
+ val = interpret_user_input(u, count);
+ if (val < 0)
+ return val;
+
+ if (dent == d->dfs_chk_gen)
+ d->chk_gen = val;
+ else if (dent == d->dfs_chk_index)
+ d->chk_index = val;
+ else if (dent == d->dfs_chk_orph)
+ d->chk_orph = val;
+ else if (dent == d->dfs_chk_lprops)
+ d->chk_lprops = val;
+ else if (dent == d->dfs_chk_fs)
+ d->chk_fs = val;
+ else if (dent == d->dfs_tst_rcvry)
+ d->tst_rcvry = val;
+ else if (dent == d->dfs_ro_error)
+ c->ro_error = !!val;
+ else
+ return -EINVAL;
+
+ return count;
+}
+
+static const struct file_operations dfs_fops = {
+ .open = dfs_file_open,
+ .read = dfs_file_read,
+ .write = dfs_file_write,
+ .owner = THIS_MODULE,
+ .llseek = no_llseek,
+};
+
+/**
+ * dbg_debugfs_init_fs - initialize debugfs for UBIFS instance.
+ * @c: UBIFS file-system description object
+ *
+ * This function creates all debugfs files for this instance of UBIFS.
+ *
+ * Note, the only reason we have not merged this function with the
+ * 'ubifs_debugging_init()' function is because it is better to initialize
+ * debugfs interfaces at the very end of the mount process, and remove them at
+ * the very beginning of the mount process.
+ */
+void dbg_debugfs_init_fs(struct ubifs_info *c)
+{
+ int n;
+ const char *fname;
+ struct ubifs_debug_info *d = c->dbg;
+
+ n = snprintf(d->dfs_dir_name, UBIFS_DFS_DIR_LEN + 1, UBIFS_DFS_DIR_NAME,
+ c->vi.ubi_num, c->vi.vol_id);
+ if (n > UBIFS_DFS_DIR_LEN) {
+ /* The array size is too small */
+ return;
+ }
+
+ fname = d->dfs_dir_name;
+ d->dfs_dir = debugfs_create_dir(fname, dfs_rootdir);
+
+ fname = "dump_lprops";
+ d->dfs_dump_lprops = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c,
+ &dfs_fops);
+
+ fname = "dump_budg";
+ d->dfs_dump_budg = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c,
+ &dfs_fops);
+
+ fname = "dump_tnc";
+ d->dfs_dump_tnc = debugfs_create_file(fname, S_IWUSR, d->dfs_dir, c,
+ &dfs_fops);
+
+ fname = "chk_general";
+ d->dfs_chk_gen = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "chk_index";
+ d->dfs_chk_index = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "chk_orphans";
+ d->dfs_chk_orph = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "chk_lprops";
+ d->dfs_chk_lprops = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "chk_fs";
+ d->dfs_chk_fs = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "tst_recovery";
+ d->dfs_tst_rcvry = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+
+ fname = "ro_error";
+ d->dfs_ro_error = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ d->dfs_dir, c, &dfs_fops);
+}
+
+/**
+ * dbg_debugfs_exit_fs - remove all debugfs files.
+ * @c: UBIFS file-system description object
+ */
+void dbg_debugfs_exit_fs(struct ubifs_info *c)
+{
+ debugfs_remove_recursive(c->dbg->dfs_dir);
+}
+
+struct ubifs_global_debug_info ubifs_dbg;
+
+static struct dentry *dfs_chk_gen;
+static struct dentry *dfs_chk_index;
+static struct dentry *dfs_chk_orph;
+static struct dentry *dfs_chk_lprops;
+static struct dentry *dfs_chk_fs;
+static struct dentry *dfs_tst_rcvry;
+
+static ssize_t dfs_global_file_read(struct file *file, char __user *u,
+ size_t count, loff_t *ppos)
+{
+ struct dentry *dent = file->f_path.dentry;
+ int val;
+
+ if (dent == dfs_chk_gen)
+ val = ubifs_dbg.chk_gen;
+ else if (dent == dfs_chk_index)
+ val = ubifs_dbg.chk_index;
+ else if (dent == dfs_chk_orph)
+ val = ubifs_dbg.chk_orph;
+ else if (dent == dfs_chk_lprops)
+ val = ubifs_dbg.chk_lprops;
+ else if (dent == dfs_chk_fs)
+ val = ubifs_dbg.chk_fs;
+ else if (dent == dfs_tst_rcvry)
+ val = ubifs_dbg.tst_rcvry;
+ else
+ return -EINVAL;
+
+ return provide_user_output(val, u, count, ppos);
+}
+
+static ssize_t dfs_global_file_write(struct file *file, const char __user *u,
+ size_t count, loff_t *ppos)
+{
+ struct dentry *dent = file->f_path.dentry;
+ int val;
+
+ val = interpret_user_input(u, count);
+ if (val < 0)
+ return val;
+
+ if (dent == dfs_chk_gen)
+ ubifs_dbg.chk_gen = val;
+ else if (dent == dfs_chk_index)
+ ubifs_dbg.chk_index = val;
+ else if (dent == dfs_chk_orph)
+ ubifs_dbg.chk_orph = val;
+ else if (dent == dfs_chk_lprops)
+ ubifs_dbg.chk_lprops = val;
+ else if (dent == dfs_chk_fs)
+ ubifs_dbg.chk_fs = val;
+ else if (dent == dfs_tst_rcvry)
+ ubifs_dbg.tst_rcvry = val;
+ else
+ return -EINVAL;
+
+ return count;
+}
+
+static const struct file_operations dfs_global_fops = {
+ .read = dfs_global_file_read,
+ .write = dfs_global_file_write,
+ .owner = THIS_MODULE,
+ .llseek = no_llseek,
+};
+
+/**
+ * dbg_debugfs_init - initialize debugfs file-system.
+ *
+ * UBIFS uses debugfs file-system to expose various debugging knobs to
+ * user-space. This function creates "ubifs" directory in the debugfs
+ * file-system.
+ */
+void dbg_debugfs_init(void)
+{
+ const char *fname;
+
+ fname = "ubifs";
+ dfs_rootdir = debugfs_create_dir(fname, NULL);
+
+ fname = "chk_general";
+ dfs_chk_gen = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir,
+ NULL, &dfs_global_fops);
+
+ fname = "chk_index";
+ dfs_chk_index = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ dfs_rootdir, NULL, &dfs_global_fops);
+
+ fname = "chk_orphans";
+ dfs_chk_orph = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ dfs_rootdir, NULL, &dfs_global_fops);
+
+ fname = "chk_lprops";
+ dfs_chk_lprops = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ dfs_rootdir, NULL, &dfs_global_fops);
+
+ fname = "chk_fs";
+ dfs_chk_fs = debugfs_create_file(fname, S_IRUSR | S_IWUSR, dfs_rootdir,
+ NULL, &dfs_global_fops);
+
+ fname = "tst_recovery";
+ dfs_tst_rcvry = debugfs_create_file(fname, S_IRUSR | S_IWUSR,
+ dfs_rootdir, NULL, &dfs_global_fops);
+}
+
+/**
+ * dbg_debugfs_exit - remove the "ubifs" directory from debugfs file-system.
+ */
+void dbg_debugfs_exit(void)
+{
+ debugfs_remove_recursive(dfs_rootdir);
+}
+
+void ubifs_assert_failed(struct ubifs_info *c, const char *expr,
+ const char *file, int line)
+{
+ ubifs_err(c, "UBIFS assert failed: %s, in %s:%u", expr, file, line);
+
+ switch (c->assert_action) {
+ case ASSACT_PANIC:
+ BUG();
+ break;
+
+ case ASSACT_RO:
+ ubifs_ro_mode(c, -EINVAL);
+ break;
+
+ case ASSACT_REPORT:
+ default:
+ dump_stack();
+ break;
+
+ }
+}
+
+/**
+ * ubifs_debugging_init - initialize UBIFS debugging.
+ * @c: UBIFS file-system description object
+ *
+ * This function initializes debugging-related data for the file system.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_debugging_init(struct ubifs_info *c)
+{
+ c->dbg = kzalloc(sizeof(struct ubifs_debug_info), GFP_KERNEL);
+ if (!c->dbg)
+ return -ENOMEM;
+
+ return 0;
+}
+
+/**
+ * ubifs_debugging_exit - free debugging data.
+ * @c: UBIFS file-system description object
+ */
+void ubifs_debugging_exit(struct ubifs_info *c)
+{
+ kfree(c->dbg);
+}
new file mode 100644
@@ -0,0 +1,304 @@
+/* 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 plus 2x2 for 2 UBI numbers and 1 for the trailing zero byte.
+ */
+#define UBIFS_DFS_DIR_NAME "ubi%d_%d"
+#define UBIFS_DFS_DIR_LEN (3 + 1 + 2*2 + 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 + 1];
+ 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__ */
new file mode 100644
@@ -0,0 +1,1744 @@
+// 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: Artem Bityutskiy (Битюцкий Артём)
+ * Adrian Hunter
+ * Zoltan Sogor
+ */
+
+/*
+ * This file implements directory operations.
+ *
+ * All FS operations in this file allocate budget before writing anything to the
+ * media. If they fail to allocate it, the error is returned. The only
+ * exceptions are 'ubifs_unlink()' and 'ubifs_rmdir()' which keep working even
+ * if they unable to allocate the budget, because deletion %-ENOSPC failure is
+ * not what users are usually ready to get. UBIFS budgeting subsystem has some
+ * space reserved for these purposes.
+ *
+ * All operations in this file write all inodes which they change straight
+ * away, instead of marking them dirty. For example, 'ubifs_link()' changes
+ * @i_size of the parent inode and writes the parent inode together with the
+ * target inode. This was done to simplify file-system recovery which would
+ * otherwise be very difficult to do. The only exception is rename which marks
+ * the re-named inode dirty (because its @i_ctime is updated) but does not
+ * write it, but just marks it as dirty.
+ */
+
+#include "ubifs.h"
+
+/**
+ * inherit_flags - inherit flags of the parent inode.
+ * @dir: parent inode
+ * @mode: new inode mode flags
+ *
+ * This is a helper function for 'ubifs_new_inode()' which inherits flag of the
+ * parent directory inode @dir. UBIFS inodes inherit the following flags:
+ * o %UBIFS_COMPR_FL, which is useful to switch compression on/of on
+ * sub-directory basis;
+ * o %UBIFS_SYNC_FL - useful for the same reasons;
+ * o %UBIFS_DIRSYNC_FL - similar, but relevant only to directories.
+ *
+ * This function returns the inherited flags.
+ */
+static int inherit_flags(const struct inode *dir, umode_t mode)
+{
+ int flags;
+ const struct ubifs_inode *ui = ubifs_inode(dir);
+
+ if (!S_ISDIR(dir->i_mode))
+ /*
+ * The parent is not a directory, which means that an extended
+ * attribute inode is being created. No flags.
+ */
+ return 0;
+
+ flags = ui->flags & (UBIFS_COMPR_FL | UBIFS_SYNC_FL | UBIFS_DIRSYNC_FL);
+ if (!S_ISDIR(mode))
+ /* The "DIRSYNC" flag only applies to directories */
+ flags &= ~UBIFS_DIRSYNC_FL;
+ return flags;
+}
+
+/**
+ * ubifs_new_inode - allocate new UBIFS inode object.
+ * @c: UBIFS file-system description object
+ * @dir: parent directory inode
+ * @mode: inode mode flags
+ * @is_xattr: whether the inode is xattr inode
+ *
+ * This function finds an unused inode number, allocates new inode and
+ * initializes it. Non-xattr new inode may be written with xattrs(selinux/
+ * encryption) before writing dentry, which could cause inconsistent problem
+ * when powercut happens between two operations. To deal with it, non-xattr
+ * new inode is initialized with zero-nlink and added into orphan list, caller
+ * should make sure that inode is relinked later, and make sure that orphan
+ * removing and journal writing into an committing atomic operation. Returns
+ * new inode in case of success and an error code in case of failure.
+ */
+struct inode *ubifs_new_inode(struct ubifs_info *c, struct inode *dir,
+ umode_t mode, bool is_xattr)
+{
+ int err;
+ struct inode *inode;
+ struct ubifs_inode *ui;
+ bool encrypted = false;
+
+ inode = new_inode(c->vfs_sb);
+ ui = ubifs_inode(inode);
+ if (!inode)
+ return ERR_PTR(-ENOMEM);
+
+ /*
+ * Set 'S_NOCMTIME' to prevent VFS form updating [mc]time of inodes and
+ * marking them dirty in file write path (see 'file_update_time()').
+ * UBIFS has to fully control "clean <-> dirty" transitions of inodes
+ * to make budgeting work.
+ */
+ inode->i_flags |= S_NOCMTIME;
+
+ inode_init_owner(&nop_mnt_idmap, inode, dir, mode);
+ simple_inode_init_ts(inode);
+ inode->i_mapping->nrpages = 0;
+
+ if (!is_xattr) {
+ err = fscrypt_prepare_new_inode(dir, inode, &encrypted);
+ if (err) {
+ ubifs_err(c, "fscrypt_prepare_new_inode failed: %i", err);
+ goto out_iput;
+ }
+ }
+
+ switch (mode & S_IFMT) {
+ case S_IFREG:
+ inode->i_mapping->a_ops = &ubifs_file_address_operations;
+ inode->i_op = &ubifs_file_inode_operations;
+ inode->i_fop = &ubifs_file_operations;
+ break;
+ case S_IFDIR:
+ inode->i_op = &ubifs_dir_inode_operations;
+ inode->i_fop = &ubifs_dir_operations;
+ inode->i_size = ui->ui_size = UBIFS_INO_NODE_SZ;
+ break;
+ case S_IFLNK:
+ inode->i_op = &ubifs_symlink_inode_operations;
+ break;
+ case S_IFSOCK:
+ case S_IFIFO:
+ case S_IFBLK:
+ case S_IFCHR:
+ inode->i_op = &ubifs_file_inode_operations;
+ break;
+ default:
+ BUG();
+ }
+
+ ui->flags = inherit_flags(dir, mode);
+ ubifs_set_inode_flags(inode);
+ if (S_ISREG(mode))
+ ui->compr_type = c->default_compr;
+ else
+ ui->compr_type = UBIFS_COMPR_NONE;
+ ui->synced_i_size = 0;
+
+ spin_lock(&c->cnt_lock);
+ /* Inode number overflow is currently not supported */
+ if (c->highest_inum >= INUM_WARN_WATERMARK) {
+ if (c->highest_inum >= INUM_WATERMARK) {
+ spin_unlock(&c->cnt_lock);
+ ubifs_err(c, "out of inode numbers");
+ err = -EINVAL;
+ goto out_iput;
+ }
+ ubifs_warn(c, "running out of inode numbers (current %lu, max %u)",
+ (unsigned long)c->highest_inum, INUM_WATERMARK);
+ }
+
+ inode->i_ino = ++c->highest_inum;
+ /*
+ * The creation sequence number remains with this inode for its
+ * lifetime. All nodes for this inode have a greater sequence number,
+ * and so it is possible to distinguish obsolete nodes belonging to a
+ * previous incarnation of the same inode number - for example, for the
+ * purpose of rebuilding the index.
+ */
+ ui->creat_sqnum = ++c->max_sqnum;
+ spin_unlock(&c->cnt_lock);
+
+ if (!is_xattr) {
+ set_nlink(inode, 0);
+ err = ubifs_add_orphan(c, inode->i_ino);
+ if (err) {
+ ubifs_err(c, "ubifs_add_orphan failed: %i", err);
+ goto out_iput;
+ }
+ down_read(&c->commit_sem);
+ ui->del_cmtno = c->cmt_no;
+ up_read(&c->commit_sem);
+ }
+
+ if (encrypted) {
+ err = fscrypt_set_context(inode, NULL);
+ if (err) {
+ if (!is_xattr) {
+ set_nlink(inode, 1);
+ ubifs_delete_orphan(c, inode->i_ino);
+ }
+ ubifs_err(c, "fscrypt_set_context failed: %i", err);
+ goto out_iput;
+ }
+ }
+
+ return inode;
+
+out_iput:
+ make_bad_inode(inode);
+ iput(inode);
+ return ERR_PTR(err);
+}
+
+static int dbg_check_name(const struct ubifs_info *c,
+ const struct ubifs_dent_node *dent,
+ const struct fscrypt_name *nm)
+{
+ if (!dbg_is_chk_gen(c))
+ return 0;
+ if (le16_to_cpu(dent->nlen) != fname_len(nm))
+ return -EINVAL;
+ if (memcmp(dent->name, fname_name(nm), fname_len(nm)))
+ return -EINVAL;
+ return 0;
+}
+
+static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry,
+ unsigned int flags)
+{
+ int err;
+ union ubifs_key key;
+ struct inode *inode = NULL;
+ struct ubifs_dent_node *dent = NULL;
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct fscrypt_name nm;
+
+ dbg_gen("'%pd' in dir ino %lu", dentry, dir->i_ino);
+
+ err = fscrypt_prepare_lookup(dir, dentry, &nm);
+ if (err == -ENOENT)
+ return d_splice_alias(NULL, dentry);
+ if (err)
+ return ERR_PTR(err);
+
+ if (fname_len(&nm) > UBIFS_MAX_NLEN) {
+ inode = ERR_PTR(-ENAMETOOLONG);
+ goto done;
+ }
+
+ dent = kmalloc(UBIFS_MAX_DENT_NODE_SZ, GFP_NOFS);
+ if (!dent) {
+ inode = ERR_PTR(-ENOMEM);
+ goto done;
+ }
+
+ if (fname_name(&nm) == NULL) {
+ if (nm.hash & ~UBIFS_S_KEY_HASH_MASK)
+ goto done; /* ENOENT */
+ dent_key_init_hash(c, &key, dir->i_ino, nm.hash);
+ err = ubifs_tnc_lookup_dh(c, &key, dent, nm.minor_hash);
+ } else {
+ dent_key_init(c, &key, dir->i_ino, &nm);
+ err = ubifs_tnc_lookup_nm(c, &key, dent, &nm);
+ }
+
+ if (err) {
+ if (err == -ENOENT)
+ dbg_gen("not found");
+ else
+ inode = ERR_PTR(err);
+ goto done;
+ }
+
+ if (dbg_check_name(c, dent, &nm)) {
+ inode = ERR_PTR(-EINVAL);
+ goto done;
+ }
+
+ inode = ubifs_iget(dir->i_sb, le64_to_cpu(dent->inum));
+ if (IS_ERR(inode)) {
+ /*
+ * This should not happen. Probably the file-system needs
+ * checking.
+ */
+ err = PTR_ERR(inode);
+ ubifs_err(c, "dead directory entry '%pd', error %d",
+ dentry, err);
+ ubifs_ro_mode(c, err);
+ goto done;
+ }
+
+ if (IS_ENCRYPTED(dir) &&
+ (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode)) &&
+ !fscrypt_has_permitted_context(dir, inode)) {
+ ubifs_warn(c, "Inconsistent encryption contexts: %lu/%lu",
+ dir->i_ino, inode->i_ino);
+ iput(inode);
+ inode = ERR_PTR(-EPERM);
+ }
+
+done:
+ kfree(dent);
+ fscrypt_free_filename(&nm);
+ return d_splice_alias(inode, dentry);
+}
+
+static int ubifs_prepare_create(struct inode *dir, struct dentry *dentry,
+ struct fscrypt_name *nm)
+{
+ if (fscrypt_is_nokey_name(dentry))
+ return -ENOKEY;
+
+ return fscrypt_setup_filename(dir, &dentry->d_name, 0, nm);
+}
+
+static int ubifs_create(struct mnt_idmap *idmap, struct inode *dir,
+ struct dentry *dentry, umode_t mode, bool excl)
+{
+ struct inode *inode;
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+ .dirtied_ino = 1 };
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ struct fscrypt_name nm;
+ int err, sz_change;
+
+ /*
+ * Budget request settings: new inode, new direntry, changing the
+ * parent directory inode.
+ */
+
+ dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
+ dentry, mode, dir->i_ino);
+
+ err = ubifs_budget_space(c, &req);
+ if (err)
+ return err;
+
+ err = ubifs_prepare_create(dir, dentry, &nm);
+ if (err)
+ goto out_budg;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ inode = ubifs_new_inode(c, dir, mode, false);
+ if (IS_ERR(inode)) {
+ err = PTR_ERR(inode);
+ goto out_fname;
+ }
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
+ set_nlink(inode, 1);
+ mutex_lock(&dir_ui->ui_mutex);
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
+ if (err)
+ goto out_cancel;
+ mutex_unlock(&dir_ui->ui_mutex);
+
+ ubifs_release_budget(c, &req);
+ fscrypt_free_filename(&nm);
+ insert_inode_hash(inode);
+ d_instantiate(dentry, inode);
+ return 0;
+
+out_cancel:
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
+out_inode:
+ iput(inode);
+out_fname:
+ fscrypt_free_filename(&nm);
+out_budg:
+ ubifs_release_budget(c, &req);
+ ubifs_err(c, "cannot create regular file, error %d", err);
+ return err;
+}
+
+static struct inode *create_whiteout(struct inode *dir, struct dentry *dentry)
+{
+ int err;
+ umode_t mode = S_IFCHR | WHITEOUT_MODE;
+ struct inode *inode;
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+
+ /*
+ * Create an inode('nlink = 1') for whiteout without updating journal,
+ * let ubifs_jnl_rename() store it on flash to complete rename whiteout
+ * atomically.
+ */
+
+ dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
+ dentry, mode, dir->i_ino);
+
+ inode = ubifs_new_inode(c, dir, mode, false);
+ if (IS_ERR(inode)) {
+ err = PTR_ERR(inode);
+ goto out_free;
+ }
+
+ init_special_inode(inode, inode->i_mode, WHITEOUT_DEV);
+ ubifs_assert(c, inode->i_op == &ubifs_file_inode_operations);
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
+ /* The dir size is updated by do_rename. */
+ insert_inode_hash(inode);
+
+ return inode;
+
+out_inode:
+ iput(inode);
+out_free:
+ ubifs_err(c, "cannot create whiteout file, error %d", err);
+ return ERR_PTR(err);
+}
+
+/**
+ * lock_2_inodes - a wrapper for locking two UBIFS inodes.
+ * @inode1: first inode
+ * @inode2: second inode
+ *
+ * We do not implement any tricks to guarantee strict lock ordering, because
+ * VFS has already done it for us on the @i_mutex. So this is just a simple
+ * wrapper function.
+ */
+static void lock_2_inodes(struct inode *inode1, struct inode *inode2)
+{
+ mutex_lock_nested(&ubifs_inode(inode1)->ui_mutex, WB_MUTEX_1);
+ mutex_lock_nested(&ubifs_inode(inode2)->ui_mutex, WB_MUTEX_2);
+}
+
+/**
+ * unlock_2_inodes - a wrapper for unlocking two UBIFS inodes.
+ * @inode1: first inode
+ * @inode2: second inode
+ */
+static void unlock_2_inodes(struct inode *inode1, struct inode *inode2)
+{
+ mutex_unlock(&ubifs_inode(inode2)->ui_mutex);
+ mutex_unlock(&ubifs_inode(inode1)->ui_mutex);
+}
+
+static int ubifs_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
+ struct file *file, umode_t mode)
+{
+ struct dentry *dentry = file->f_path.dentry;
+ struct inode *inode;
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+ .dirtied_ino = 1};
+ struct ubifs_budget_req ino_req = { .dirtied_ino = 1 };
+ struct ubifs_inode *ui;
+ int err, instantiated = 0;
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: new inode, new direntry, changing the
+ * parent directory inode.
+ * Allocate budget separately for new dirtied inode, the budget will
+ * be released via writeback.
+ */
+
+ dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
+ dentry, mode, dir->i_ino);
+
+ err = fscrypt_setup_filename(dir, &dentry->d_name, 0, &nm);
+ if (err)
+ return err;
+
+ err = ubifs_budget_space(c, &req);
+ if (err) {
+ fscrypt_free_filename(&nm);
+ return err;
+ }
+
+ err = ubifs_budget_space(c, &ino_req);
+ if (err) {
+ ubifs_release_budget(c, &req);
+ fscrypt_free_filename(&nm);
+ return err;
+ }
+
+ inode = ubifs_new_inode(c, dir, mode, false);
+ if (IS_ERR(inode)) {
+ err = PTR_ERR(inode);
+ goto out_budg;
+ }
+ ui = ubifs_inode(inode);
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
+ set_nlink(inode, 1);
+ mutex_lock(&ui->ui_mutex);
+ insert_inode_hash(inode);
+ d_tmpfile(file, inode);
+ ubifs_assert(c, ui->dirty);
+
+ instantiated = 1;
+ mutex_unlock(&ui->ui_mutex);
+
+ lock_2_inodes(dir, inode);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 1);
+ if (err)
+ goto out_cancel;
+ unlock_2_inodes(dir, inode);
+
+ ubifs_release_budget(c, &req);
+ fscrypt_free_filename(&nm);
+
+ return finish_open_simple(file, 0);
+
+out_cancel:
+ unlock_2_inodes(dir, inode);
+out_inode:
+ if (!instantiated)
+ iput(inode);
+out_budg:
+ ubifs_release_budget(c, &req);
+ if (!instantiated)
+ ubifs_release_budget(c, &ino_req);
+ fscrypt_free_filename(&nm);
+ ubifs_err(c, "cannot create temporary file, error %d", err);
+ return err;
+}
+
+/**
+ * vfs_dent_type - get VFS directory entry type.
+ * @type: UBIFS directory entry type
+ *
+ * This function converts UBIFS directory entry type into VFS directory entry
+ * type.
+ */
+static unsigned int vfs_dent_type(uint8_t type)
+{
+ switch (type) {
+ case UBIFS_ITYPE_REG:
+ return DT_REG;
+ case UBIFS_ITYPE_DIR:
+ return DT_DIR;
+ case UBIFS_ITYPE_LNK:
+ return DT_LNK;
+ case UBIFS_ITYPE_BLK:
+ return DT_BLK;
+ case UBIFS_ITYPE_CHR:
+ return DT_CHR;
+ case UBIFS_ITYPE_FIFO:
+ return DT_FIFO;
+ case UBIFS_ITYPE_SOCK:
+ return DT_SOCK;
+ default:
+ BUG();
+ }
+ return 0;
+}
+
+/*
+ * The classical Unix view for directory is that it is a linear array of
+ * (name, inode number) entries. Linux/VFS assumes this model as well.
+ * Particularly, 'readdir()' call wants us to return a directory entry offset
+ * which later may be used to continue 'readdir()'ing the directory or to
+ * 'seek()' to that specific direntry. Obviously UBIFS does not really fit this
+ * model because directory entries are identified by keys, which may collide.
+ *
+ * UBIFS uses directory entry hash value for directory offsets, so
+ * 'seekdir()'/'telldir()' may not always work because of possible key
+ * collisions. But UBIFS guarantees that consecutive 'readdir()' calls work
+ * properly by means of saving full directory entry name in the private field
+ * of the file description object.
+ *
+ * This means that UBIFS cannot support NFS which requires full
+ * 'seekdir()'/'telldir()' support.
+ */
+static int ubifs_readdir(struct file *file, struct dir_context *ctx)
+{
+ int fstr_real_len = 0, err = 0;
+ struct fscrypt_name nm;
+ struct fscrypt_str fstr = {0};
+ union ubifs_key key;
+ struct ubifs_dent_node *dent;
+ struct inode *dir = file_inode(file);
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ bool encrypted = IS_ENCRYPTED(dir);
+
+ dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, ctx->pos);
+
+ if (ctx->pos > UBIFS_S_KEY_HASH_MASK || ctx->pos == 2)
+ /*
+ * The directory was seek'ed to a senseless position or there
+ * are no more entries.
+ */
+ return 0;
+
+ if (encrypted) {
+ err = fscrypt_prepare_readdir(dir);
+ if (err)
+ return err;
+
+ err = fscrypt_fname_alloc_buffer(UBIFS_MAX_NLEN, &fstr);
+ if (err)
+ return err;
+
+ fstr_real_len = fstr.len;
+ }
+
+ if (file->f_version == 0) {
+ /*
+ * The file was seek'ed, which means that @file->private_data
+ * is now invalid. This may also be just the first
+ * 'ubifs_readdir()' invocation, in which case
+ * @file->private_data is NULL, and the below code is
+ * basically a no-op.
+ */
+ kfree(file->private_data);
+ file->private_data = NULL;
+ }
+
+ /*
+ * 'generic_file_llseek()' unconditionally sets @file->f_version to
+ * zero, and we use this for detecting whether the file was seek'ed.
+ */
+ file->f_version = 1;
+
+ /* File positions 0 and 1 correspond to "." and ".." */
+ if (ctx->pos < 2) {
+ ubifs_assert(c, !file->private_data);
+ if (!dir_emit_dots(file, ctx)) {
+ if (encrypted)
+ fscrypt_fname_free_buffer(&fstr);
+ return 0;
+ }
+
+ /* Find the first entry in TNC and save it */
+ lowest_dent_key(c, &key, dir->i_ino);
+ fname_len(&nm) = 0;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ ctx->pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ }
+
+ dent = file->private_data;
+ if (!dent) {
+ /*
+ * The directory was seek'ed to and is now readdir'ed.
+ * Find the entry corresponding to @ctx->pos or the closest one.
+ */
+ dent_key_init_hash(c, &key, dir->i_ino, ctx->pos);
+ fname_len(&nm) = 0;
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+ ctx->pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ }
+
+ while (1) {
+ dbg_gen("ino %llu, new f_pos %#x",
+ (unsigned long long)le64_to_cpu(dent->inum),
+ key_hash_flash(c, &dent->key));
+ ubifs_assert(c, le64_to_cpu(dent->ch.sqnum) >
+ ubifs_inode(dir)->creat_sqnum);
+
+ fname_len(&nm) = le16_to_cpu(dent->nlen);
+ fname_name(&nm) = dent->name;
+
+ if (encrypted) {
+ fstr.len = fstr_real_len;
+
+ err = fscrypt_fname_disk_to_usr(dir, key_hash_flash(c,
+ &dent->key),
+ le32_to_cpu(dent->cookie),
+ &nm.disk_name, &fstr);
+ if (err)
+ goto out;
+ } else {
+ fstr.len = fname_len(&nm);
+ fstr.name = fname_name(&nm);
+ }
+
+ if (!dir_emit(ctx, fstr.name, fstr.len,
+ le64_to_cpu(dent->inum),
+ vfs_dent_type(dent->type))) {
+ if (encrypted)
+ fscrypt_fname_free_buffer(&fstr);
+ return 0;
+ }
+
+ /* Switch to the next entry */
+ key_read(c, &dent->key, &key);
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ goto out;
+ }
+
+ kfree(file->private_data);
+ ctx->pos = key_hash_flash(c, &dent->key);
+ file->private_data = dent;
+ cond_resched();
+ }
+
+out:
+ kfree(file->private_data);
+ file->private_data = NULL;
+
+ if (encrypted)
+ fscrypt_fname_free_buffer(&fstr);
+
+ if (err != -ENOENT)
+ ubifs_err(c, "cannot find next direntry, error %d", err);
+ else
+ /*
+ * -ENOENT is a non-fatal error in this context, the TNC uses
+ * it to indicate that the cursor moved past the current directory
+ * and readdir() has to stop.
+ */
+ err = 0;
+
+
+ /* 2 is a special value indicating that there are no more direntries */
+ ctx->pos = 2;
+ return err;
+}
+
+/* Free saved readdir() state when the directory is closed */
+static int ubifs_dir_release(struct inode *dir, struct file *file)
+{
+ kfree(file->private_data);
+ file->private_data = NULL;
+ return 0;
+}
+
+static int ubifs_link(struct dentry *old_dentry, struct inode *dir,
+ struct dentry *dentry)
+{
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct inode *inode = d_inode(old_dentry);
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ int err, sz_change;
+ struct ubifs_budget_req req = { .new_dent = 1, .dirtied_ino = 2,
+ .dirtied_ino_d = ALIGN(ui->data_len, 8) };
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: new direntry, changing the target inode,
+ * changing the parent inode.
+ */
+
+ dbg_gen("dent '%pd' to ino %lu (nlink %d) in dir ino %lu",
+ dentry, inode->i_ino,
+ inode->i_nlink, dir->i_ino);
+ ubifs_assert(c, inode_is_locked(dir));
+ ubifs_assert(c, inode_is_locked(inode));
+
+ err = fscrypt_prepare_link(old_dentry, dir, dentry);
+ if (err)
+ return err;
+
+ err = fscrypt_setup_filename(dir, &dentry->d_name, 0, &nm);
+ if (err)
+ return err;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ err = dbg_check_synced_i_size(c, inode);
+ if (err)
+ goto out_fname;
+
+ err = ubifs_budget_space(c, &req);
+ if (err)
+ goto out_fname;
+
+ lock_2_inodes(dir, inode);
+
+ inc_nlink(inode);
+ ihold(inode);
+ inode_set_ctime_current(inode);
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, inode->i_nlink == 1);
+ if (err)
+ goto out_cancel;
+ unlock_2_inodes(dir, inode);
+
+ ubifs_release_budget(c, &req);
+ d_instantiate(dentry, inode);
+ fscrypt_free_filename(&nm);
+ return 0;
+
+out_cancel:
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ drop_nlink(inode);
+ unlock_2_inodes(dir, inode);
+ ubifs_release_budget(c, &req);
+ iput(inode);
+out_fname:
+ fscrypt_free_filename(&nm);
+ return err;
+}
+
+static int ubifs_unlink(struct inode *dir, struct dentry *dentry)
+{
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct inode *inode = d_inode(dentry);
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ int err, sz_change, budgeted = 1;
+ struct ubifs_budget_req req = { .mod_dent = 1, .dirtied_ino = 2 };
+ unsigned int saved_nlink = inode->i_nlink;
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: deletion direntry, deletion inode (+1 for
+ * @dirtied_ino), changing the parent directory inode. If budgeting
+ * fails, go ahead anyway because we have extra space reserved for
+ * deletions.
+ */
+
+ dbg_gen("dent '%pd' from ino %lu (nlink %d) in dir ino %lu",
+ dentry, inode->i_ino,
+ inode->i_nlink, dir->i_ino);
+
+ err = fscrypt_setup_filename(dir, &dentry->d_name, 1, &nm);
+ if (err)
+ return err;
+
+ err = ubifs_purge_xattrs(inode);
+ if (err)
+ return err;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ ubifs_assert(c, inode_is_locked(dir));
+ ubifs_assert(c, inode_is_locked(inode));
+ err = dbg_check_synced_i_size(c, inode);
+ if (err)
+ goto out_fname;
+
+ err = ubifs_budget_space(c, &req);
+ if (err) {
+ if (err != -ENOSPC)
+ goto out_fname;
+ budgeted = 0;
+ }
+
+ lock_2_inodes(dir, inode);
+ inode_set_ctime_current(inode);
+ drop_nlink(inode);
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 0);
+ if (err)
+ goto out_cancel;
+ unlock_2_inodes(dir, inode);
+
+ if (budgeted)
+ ubifs_release_budget(c, &req);
+ else {
+ /* We've deleted something - clean the "no space" flags */
+ c->bi.nospace = c->bi.nospace_rp = 0;
+ smp_wmb();
+ }
+ fscrypt_free_filename(&nm);
+ return 0;
+
+out_cancel:
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ set_nlink(inode, saved_nlink);
+ unlock_2_inodes(dir, inode);
+ if (budgeted)
+ ubifs_release_budget(c, &req);
+out_fname:
+ fscrypt_free_filename(&nm);
+ return err;
+}
+
+/**
+ * ubifs_check_dir_empty - check if a directory is empty or not.
+ * @dir: VFS inode object of the directory to check
+ *
+ * This function checks if directory @dir is empty. Returns zero if the
+ * directory is empty, %-ENOTEMPTY if it is not, and other negative error codes
+ * in case of errors.
+ */
+int ubifs_check_dir_empty(struct inode *dir)
+{
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct fscrypt_name nm = { 0 };
+ struct ubifs_dent_node *dent;
+ union ubifs_key key;
+ int err;
+
+ lowest_dent_key(c, &key, dir->i_ino);
+ dent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(dent)) {
+ err = PTR_ERR(dent);
+ if (err == -ENOENT)
+ err = 0;
+ } else {
+ kfree(dent);
+ err = -ENOTEMPTY;
+ }
+ return err;
+}
+
+static int ubifs_rmdir(struct inode *dir, struct dentry *dentry)
+{
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ struct inode *inode = d_inode(dentry);
+ int err, sz_change, budgeted = 1;
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ struct ubifs_budget_req req = { .mod_dent = 1, .dirtied_ino = 2 };
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: deletion direntry, deletion inode and
+ * changing the parent inode. If budgeting fails, go ahead anyway
+ * because we have extra space reserved for deletions.
+ */
+
+ dbg_gen("directory '%pd', ino %lu in dir ino %lu", dentry,
+ inode->i_ino, dir->i_ino);
+ ubifs_assert(c, inode_is_locked(dir));
+ ubifs_assert(c, inode_is_locked(inode));
+ err = ubifs_check_dir_empty(d_inode(dentry));
+ if (err)
+ return err;
+
+ err = fscrypt_setup_filename(dir, &dentry->d_name, 1, &nm);
+ if (err)
+ return err;
+
+ err = ubifs_purge_xattrs(inode);
+ if (err)
+ return err;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ err = ubifs_budget_space(c, &req);
+ if (err) {
+ if (err != -ENOSPC)
+ goto out_fname;
+ budgeted = 0;
+ }
+
+ lock_2_inodes(dir, inode);
+ inode_set_ctime_current(inode);
+ clear_nlink(inode);
+ drop_nlink(dir);
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 0);
+ if (err)
+ goto out_cancel;
+ unlock_2_inodes(dir, inode);
+
+ if (budgeted)
+ ubifs_release_budget(c, &req);
+ else {
+ /* We've deleted something - clean the "no space" flags */
+ c->bi.nospace = c->bi.nospace_rp = 0;
+ smp_wmb();
+ }
+ fscrypt_free_filename(&nm);
+ return 0;
+
+out_cancel:
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inc_nlink(dir);
+ set_nlink(inode, 2);
+ unlock_2_inodes(dir, inode);
+ if (budgeted)
+ ubifs_release_budget(c, &req);
+out_fname:
+ fscrypt_free_filename(&nm);
+ return err;
+}
+
+static int ubifs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
+ struct dentry *dentry, umode_t mode)
+{
+ struct inode *inode;
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ int err, sz_change;
+ struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+ .dirtied_ino = 1};
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: new inode, new direntry and changing parent
+ * directory inode.
+ */
+
+ dbg_gen("dent '%pd', mode %#hx in dir ino %lu",
+ dentry, mode, dir->i_ino);
+
+ err = ubifs_budget_space(c, &req);
+ if (err)
+ return err;
+
+ err = ubifs_prepare_create(dir, dentry, &nm);
+ if (err)
+ goto out_budg;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ inode = ubifs_new_inode(c, dir, S_IFDIR | mode, false);
+ if (IS_ERR(inode)) {
+ err = PTR_ERR(inode);
+ goto out_fname;
+ }
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
+ set_nlink(inode, 1);
+ mutex_lock(&dir_ui->ui_mutex);
+ insert_inode_hash(inode);
+ inc_nlink(inode);
+ inc_nlink(dir);
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
+ if (err) {
+ ubifs_err(c, "cannot create directory, error %d", err);
+ goto out_cancel;
+ }
+ mutex_unlock(&dir_ui->ui_mutex);
+
+ ubifs_release_budget(c, &req);
+ d_instantiate(dentry, inode);
+ fscrypt_free_filename(&nm);
+ return 0;
+
+out_cancel:
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ drop_nlink(dir);
+ mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
+out_inode:
+ iput(inode);
+out_fname:
+ fscrypt_free_filename(&nm);
+out_budg:
+ ubifs_release_budget(c, &req);
+ return err;
+}
+
+static int ubifs_mknod(struct mnt_idmap *idmap, struct inode *dir,
+ struct dentry *dentry, umode_t mode, dev_t rdev)
+{
+ struct inode *inode;
+ struct ubifs_inode *ui;
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ union ubifs_dev_desc *dev = NULL;
+ int sz_change;
+ int err, devlen = 0;
+ struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+ .dirtied_ino = 1 };
+ struct fscrypt_name nm;
+
+ /*
+ * Budget request settings: new inode, new direntry and changing parent
+ * directory inode.
+ */
+
+ dbg_gen("dent '%pd' in dir ino %lu", dentry, dir->i_ino);
+
+ if (S_ISBLK(mode) || S_ISCHR(mode)) {
+ dev = kmalloc(sizeof(union ubifs_dev_desc), GFP_NOFS);
+ if (!dev)
+ return -ENOMEM;
+ devlen = ubifs_encode_dev(dev, rdev);
+ }
+
+ req.new_ino_d = ALIGN(devlen, 8);
+ err = ubifs_budget_space(c, &req);
+ if (err) {
+ kfree(dev);
+ return err;
+ }
+
+ err = ubifs_prepare_create(dir, dentry, &nm);
+ if (err) {
+ kfree(dev);
+ goto out_budg;
+ }
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ inode = ubifs_new_inode(c, dir, mode, false);
+ if (IS_ERR(inode)) {
+ kfree(dev);
+ err = PTR_ERR(inode);
+ goto out_fname;
+ }
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err) {
+ kfree(dev);
+ goto out_inode;
+ }
+
+ init_special_inode(inode, inode->i_mode, rdev);
+ inode->i_size = ubifs_inode(inode)->ui_size = devlen;
+ ui = ubifs_inode(inode);
+ ui->data = dev;
+ ui->data_len = devlen;
+ set_nlink(inode, 1);
+
+ mutex_lock(&dir_ui->ui_mutex);
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
+ if (err)
+ goto out_cancel;
+ mutex_unlock(&dir_ui->ui_mutex);
+
+ ubifs_release_budget(c, &req);
+ insert_inode_hash(inode);
+ d_instantiate(dentry, inode);
+ fscrypt_free_filename(&nm);
+ return 0;
+
+out_cancel:
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
+out_inode:
+ iput(inode);
+out_fname:
+ fscrypt_free_filename(&nm);
+out_budg:
+ ubifs_release_budget(c, &req);
+ return err;
+}
+
+static int ubifs_symlink(struct mnt_idmap *idmap, struct inode *dir,
+ struct dentry *dentry, const char *symname)
+{
+ struct inode *inode;
+ struct ubifs_inode *ui;
+ struct ubifs_inode *dir_ui = ubifs_inode(dir);
+ struct ubifs_info *c = dir->i_sb->s_fs_info;
+ int err, sz_change, len = strlen(symname);
+ struct fscrypt_str disk_link;
+ struct ubifs_budget_req req = { .new_ino = 1, .new_dent = 1,
+ .dirtied_ino = 1 };
+ struct fscrypt_name nm;
+
+ dbg_gen("dent '%pd', target '%s' in dir ino %lu", dentry,
+ symname, dir->i_ino);
+
+ err = fscrypt_prepare_symlink(dir, symname, len, UBIFS_MAX_INO_DATA,
+ &disk_link);
+ if (err)
+ return err;
+
+ /*
+ * Budget request settings: new inode, new direntry and changing parent
+ * directory inode.
+ */
+ req.new_ino_d = ALIGN(disk_link.len - 1, 8);
+ err = ubifs_budget_space(c, &req);
+ if (err)
+ return err;
+
+ err = ubifs_prepare_create(dir, dentry, &nm);
+ if (err)
+ goto out_budg;
+
+ sz_change = CALC_DENT_SIZE(fname_len(&nm));
+
+ inode = ubifs_new_inode(c, dir, S_IFLNK | S_IRWXUGO, false);
+ if (IS_ERR(inode)) {
+ err = PTR_ERR(inode);
+ goto out_fname;
+ }
+
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
+ ui = ubifs_inode(inode);
+ ui->data = kmalloc(disk_link.len, GFP_NOFS);
+ if (!ui->data) {
+ err = -ENOMEM;
+ goto out_inode;
+ }
+
+ if (IS_ENCRYPTED(inode)) {
+ disk_link.name = ui->data; /* encrypt directly into ui->data */
+ err = fscrypt_encrypt_symlink(inode, symname, len, &disk_link);
+ if (err)
+ goto out_inode;
+ } else {
+ memcpy(ui->data, disk_link.name, disk_link.len);
+ inode->i_link = ui->data;
+ }
+
+ /*
+ * The terminating zero byte is not written to the flash media and it
+ * is put just to make later in-memory string processing simpler. Thus,
+ * data length is @disk_link.len - 1, not @disk_link.len.
+ */
+ ui->data_len = disk_link.len - 1;
+ inode->i_size = ubifs_inode(inode)->ui_size = disk_link.len - 1;
+ set_nlink(inode, 1);
+
+ mutex_lock(&dir_ui->ui_mutex);
+ dir->i_size += sz_change;
+ dir_ui->ui_size = dir->i_size;
+ inode_set_mtime_to_ts(dir,
+ inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
+ if (err)
+ goto out_cancel;
+ mutex_unlock(&dir_ui->ui_mutex);
+
+ insert_inode_hash(inode);
+ d_instantiate(dentry, inode);
+ err = 0;
+ goto out_fname;
+
+out_cancel:
+ dir->i_size -= sz_change;
+ dir_ui->ui_size = dir->i_size;
+ mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
+out_inode:
+ /* Free inode->i_link before inode is marked as bad. */
+ fscrypt_free_inode(inode);
+ iput(inode);
+out_fname:
+ fscrypt_free_filename(&nm);
+out_budg:
+ ubifs_release_budget(c, &req);
+ return err;
+}
+
+/**
+ * lock_4_inodes - a wrapper for locking three UBIFS inodes.
+ * @inode1: first inode
+ * @inode2: second inode
+ * @inode3: third inode
+ * @inode4: fourth inode
+ *
+ * This function is used for 'ubifs_rename()' and @inode1 may be the same as
+ * @inode2 whereas @inode3 and @inode4 may be %NULL.
+ *
+ * We do not implement any tricks to guarantee strict lock ordering, because
+ * VFS has already done it for us on the @i_mutex. So this is just a simple
+ * wrapper function.
+ */
+static void lock_4_inodes(struct inode *inode1, struct inode *inode2,
+ struct inode *inode3, struct inode *inode4)
+{
+ mutex_lock_nested(&ubifs_inode(inode1)->ui_mutex, WB_MUTEX_1);
+ if (inode2 != inode1)
+ mutex_lock_nested(&ubifs_inode(inode2)->ui_mutex, WB_MUTEX_2);
+ if (inode3)
+ mutex_lock_nested(&ubifs_inode(inode3)->ui_mutex, WB_MUTEX_3);
+ if (inode4)
+ mutex_lock_nested(&ubifs_inode(inode4)->ui_mutex, WB_MUTEX_4);
+}
+
+/**
+ * unlock_4_inodes - a wrapper for unlocking three UBIFS inodes for rename.
+ * @inode1: first inode
+ * @inode2: second inode
+ * @inode3: third inode
+ * @inode4: fourth inode
+ */
+static void unlock_4_inodes(struct inode *inode1, struct inode *inode2,
+ struct inode *inode3, struct inode *inode4)
+{
+ if (inode4)
+ mutex_unlock(&ubifs_inode(inode4)->ui_mutex);
+ if (inode3)
+ mutex_unlock(&ubifs_inode(inode3)->ui_mutex);
+ if (inode1 != inode2)
+ mutex_unlock(&ubifs_inode(inode2)->ui_mutex);
+ mutex_unlock(&ubifs_inode(inode1)->ui_mutex);
+}
+
+static int do_rename(struct inode *old_dir, struct dentry *old_dentry,
+ struct inode *new_dir, struct dentry *new_dentry,
+ unsigned int flags)
+{
+ struct ubifs_info *c = old_dir->i_sb->s_fs_info;
+ struct inode *old_inode = d_inode(old_dentry);
+ struct inode *new_inode = d_inode(new_dentry);
+ struct inode *whiteout = NULL;
+ struct ubifs_inode *old_inode_ui = ubifs_inode(old_inode);
+ struct ubifs_inode *whiteout_ui = NULL;
+ int err, release, sync = 0, move = (new_dir != old_dir);
+ int is_dir = S_ISDIR(old_inode->i_mode);
+ int unlink = !!new_inode, new_sz, old_sz;
+ struct ubifs_budget_req req = { .new_dent = 1, .mod_dent = 1,
+ .dirtied_ino = 3 };
+ struct ubifs_budget_req ino_req = { .dirtied_ino = 1,
+ .dirtied_ino_d = ALIGN(old_inode_ui->data_len, 8) };
+ struct ubifs_budget_req wht_req;
+ unsigned int saved_nlink;
+ struct fscrypt_name old_nm, new_nm;
+
+ /*
+ * Budget request settings:
+ * req: deletion direntry, new direntry, removing the old inode,
+ * and changing old and new parent directory inodes.
+ *
+ * wht_req: new whiteout inode for RENAME_WHITEOUT.
+ *
+ * ino_req: marks the target inode as dirty and does not write it.
+ */
+
+ dbg_gen("dent '%pd' ino %lu in dir ino %lu to dent '%pd' in dir ino %lu flags 0x%x",
+ old_dentry, old_inode->i_ino, old_dir->i_ino,
+ new_dentry, new_dir->i_ino, flags);
+
+ if (unlink) {
+ ubifs_assert(c, inode_is_locked(new_inode));
+
+ /* Budget for old inode's data when its nlink > 1. */
+ req.dirtied_ino_d = ALIGN(ubifs_inode(new_inode)->data_len, 8);
+ err = ubifs_purge_xattrs(new_inode);
+ if (err)
+ return err;
+ }
+
+ if (unlink && is_dir) {
+ err = ubifs_check_dir_empty(new_inode);
+ if (err)
+ return err;
+ }
+
+ err = fscrypt_setup_filename(old_dir, &old_dentry->d_name, 0, &old_nm);
+ if (err)
+ return err;
+
+ err = fscrypt_setup_filename(new_dir, &new_dentry->d_name, 0, &new_nm);
+ if (err) {
+ fscrypt_free_filename(&old_nm);
+ return err;
+ }
+
+ new_sz = CALC_DENT_SIZE(fname_len(&new_nm));
+ old_sz = CALC_DENT_SIZE(fname_len(&old_nm));
+
+ err = ubifs_budget_space(c, &req);
+ if (err) {
+ fscrypt_free_filename(&old_nm);
+ fscrypt_free_filename(&new_nm);
+ return err;
+ }
+ err = ubifs_budget_space(c, &ino_req);
+ if (err) {
+ fscrypt_free_filename(&old_nm);
+ fscrypt_free_filename(&new_nm);
+ ubifs_release_budget(c, &req);
+ return err;
+ }
+
+ if (flags & RENAME_WHITEOUT) {
+ union ubifs_dev_desc *dev = NULL;
+
+ dev = kmalloc(sizeof(union ubifs_dev_desc), GFP_NOFS);
+ if (!dev) {
+ err = -ENOMEM;
+ goto out_release;
+ }
+
+ /*
+ * The whiteout inode without dentry is pinned in memory,
+ * umount won't happen during rename process because we
+ * got parent dentry.
+ */
+ whiteout = create_whiteout(old_dir, old_dentry);
+ if (IS_ERR(whiteout)) {
+ err = PTR_ERR(whiteout);
+ kfree(dev);
+ goto out_release;
+ }
+
+ whiteout_ui = ubifs_inode(whiteout);
+ whiteout_ui->data = dev;
+ whiteout_ui->data_len = ubifs_encode_dev(dev, MKDEV(0, 0));
+ ubifs_assert(c, !whiteout_ui->dirty);
+
+ memset(&wht_req, 0, sizeof(struct ubifs_budget_req));
+ wht_req.new_ino = 1;
+ wht_req.new_ino_d = ALIGN(whiteout_ui->data_len, 8);
+ /*
+ * To avoid deadlock between space budget (holds ui_mutex and
+ * waits wb work) and writeback work(waits ui_mutex), do space
+ * budget before ubifs inodes locked.
+ */
+ err = ubifs_budget_space(c, &wht_req);
+ if (err) {
+ iput(whiteout);
+ goto out_release;
+ }
+ set_nlink(whiteout, 1);
+
+ /* Add the old_dentry size to the old_dir size. */
+ old_sz -= CALC_DENT_SIZE(fname_len(&old_nm));
+ }
+
+ lock_4_inodes(old_dir, new_dir, new_inode, whiteout);
+
+ /*
+ * Like most other Unix systems, set the @i_ctime for inodes on a
+ * rename.
+ */
+ simple_rename_timestamp(old_dir, old_dentry, new_dir, new_dentry);
+
+ /* We must adjust parent link count when renaming directories */
+ if (is_dir) {
+ if (move) {
+ /*
+ * @old_dir loses a link because we are moving
+ * @old_inode to a different directory.
+ */
+ drop_nlink(old_dir);
+ /*
+ * @new_dir only gains a link if we are not also
+ * overwriting an existing directory.
+ */
+ if (!unlink)
+ inc_nlink(new_dir);
+ } else {
+ /*
+ * @old_inode is not moving to a different directory,
+ * but @old_dir still loses a link if we are
+ * overwriting an existing directory.
+ */
+ if (unlink)
+ drop_nlink(old_dir);
+ }
+ }
+
+ old_dir->i_size -= old_sz;
+ ubifs_inode(old_dir)->ui_size = old_dir->i_size;
+
+ /*
+ * And finally, if we unlinked a direntry which happened to have the
+ * same name as the moved direntry, we have to decrement @i_nlink of
+ * the unlinked inode.
+ */
+ if (unlink) {
+ /*
+ * Directories cannot have hard-links, so if this is a
+ * directory, just clear @i_nlink.
+ */
+ saved_nlink = new_inode->i_nlink;
+ if (is_dir)
+ clear_nlink(new_inode);
+ else
+ drop_nlink(new_inode);
+ } else {
+ new_dir->i_size += new_sz;
+ ubifs_inode(new_dir)->ui_size = new_dir->i_size;
+ }
+
+ /*
+ * Do not ask 'ubifs_jnl_rename()' to flush write-buffer if @old_inode
+ * is dirty, because this will be done later on at the end of
+ * 'ubifs_rename()'.
+ */
+ if (IS_SYNC(old_inode)) {
+ sync = IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir);
+ if (unlink && IS_SYNC(new_inode))
+ sync = 1;
+ /*
+ * S_SYNC flag of whiteout inherits from the old_dir, and we
+ * have already checked the old dir inode. So there is no need
+ * to check whiteout.
+ */
+ }
+
+ err = ubifs_jnl_rename(c, old_dir, old_inode, &old_nm, new_dir,
+ new_inode, &new_nm, whiteout, sync, !!whiteout);
+ if (err)
+ goto out_cancel;
+
+ unlock_4_inodes(old_dir, new_dir, new_inode, whiteout);
+ ubifs_release_budget(c, &req);
+
+ if (whiteout) {
+ ubifs_release_budget(c, &wht_req);
+ iput(whiteout);
+ }
+
+ mutex_lock(&old_inode_ui->ui_mutex);
+ release = old_inode_ui->dirty;
+ mark_inode_dirty_sync(old_inode);
+ mutex_unlock(&old_inode_ui->ui_mutex);
+
+ if (release)
+ ubifs_release_budget(c, &ino_req);
+ if (IS_SYNC(old_inode))
+ /*
+ * Rename finished here. Although old inode cannot be updated
+ * on flash, old ctime is not a big problem, don't return err
+ * code to userspace.
+ */
+ old_inode->i_sb->s_op->write_inode(old_inode, NULL);
+
+ fscrypt_free_filename(&old_nm);
+ fscrypt_free_filename(&new_nm);
+ return 0;
+
+out_cancel:
+ if (unlink) {
+ set_nlink(new_inode, saved_nlink);
+ } else {
+ new_dir->i_size -= new_sz;
+ ubifs_inode(new_dir)->ui_size = new_dir->i_size;
+ }
+ old_dir->i_size += old_sz;
+ ubifs_inode(old_dir)->ui_size = old_dir->i_size;
+ if (is_dir) {
+ if (move) {
+ inc_nlink(old_dir);
+ if (!unlink)
+ drop_nlink(new_dir);
+ } else {
+ if (unlink)
+ inc_nlink(old_dir);
+ }
+ }
+ unlock_4_inodes(old_dir, new_dir, new_inode, whiteout);
+ if (whiteout) {
+ ubifs_release_budget(c, &wht_req);
+ set_nlink(whiteout, 0);
+ iput(whiteout);
+ }
+out_release:
+ ubifs_release_budget(c, &ino_req);
+ ubifs_release_budget(c, &req);
+ fscrypt_free_filename(&old_nm);
+ fscrypt_free_filename(&new_nm);
+ return err;
+}
+
+static int ubifs_xrename(struct inode *old_dir, struct dentry *old_dentry,
+ struct inode *new_dir, struct dentry *new_dentry)
+{
+ struct ubifs_info *c = old_dir->i_sb->s_fs_info;
+ struct ubifs_budget_req req = { .new_dent = 1, .mod_dent = 1,
+ .dirtied_ino = 2 };
+ int sync = IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir);
+ struct inode *fst_inode = d_inode(old_dentry);
+ struct inode *snd_inode = d_inode(new_dentry);
+ int err;
+ struct fscrypt_name fst_nm, snd_nm;
+
+ ubifs_assert(c, fst_inode && snd_inode);
+
+ /*
+ * Budget request settings: changing two direntries, changing the two
+ * parent directory inodes.
+ */
+
+ dbg_gen("dent '%pd' ino %lu in dir ino %lu exchange dent '%pd' ino %lu in dir ino %lu",
+ old_dentry, fst_inode->i_ino, old_dir->i_ino,
+ new_dentry, snd_inode->i_ino, new_dir->i_ino);
+
+ err = fscrypt_setup_filename(old_dir, &old_dentry->d_name, 0, &fst_nm);
+ if (err)
+ return err;
+
+ err = fscrypt_setup_filename(new_dir, &new_dentry->d_name, 0, &snd_nm);
+ if (err) {
+ fscrypt_free_filename(&fst_nm);
+ return err;
+ }
+
+ err = ubifs_budget_space(c, &req);
+ if (err)
+ goto out;
+
+ lock_4_inodes(old_dir, new_dir, NULL, NULL);
+
+ simple_rename_timestamp(old_dir, old_dentry, new_dir, new_dentry);
+
+ if (old_dir != new_dir) {
+ if (S_ISDIR(fst_inode->i_mode) && !S_ISDIR(snd_inode->i_mode)) {
+ inc_nlink(new_dir);
+ drop_nlink(old_dir);
+ }
+ else if (!S_ISDIR(fst_inode->i_mode) && S_ISDIR(snd_inode->i_mode)) {
+ drop_nlink(new_dir);
+ inc_nlink(old_dir);
+ }
+ }
+
+ err = ubifs_jnl_xrename(c, old_dir, fst_inode, &fst_nm, new_dir,
+ snd_inode, &snd_nm, sync);
+
+ unlock_4_inodes(old_dir, new_dir, NULL, NULL);
+ ubifs_release_budget(c, &req);
+
+out:
+ fscrypt_free_filename(&fst_nm);
+ fscrypt_free_filename(&snd_nm);
+ return err;
+}
+
+static int ubifs_rename(struct mnt_idmap *idmap,
+ struct inode *old_dir, struct dentry *old_dentry,
+ struct inode *new_dir, struct dentry *new_dentry,
+ unsigned int flags)
+{
+ int err;
+ struct ubifs_info *c = old_dir->i_sb->s_fs_info;
+
+ if (flags & ~(RENAME_NOREPLACE | RENAME_WHITEOUT | RENAME_EXCHANGE))
+ return -EINVAL;
+
+ ubifs_assert(c, inode_is_locked(old_dir));
+ ubifs_assert(c, inode_is_locked(new_dir));
+
+ err = fscrypt_prepare_rename(old_dir, old_dentry, new_dir, new_dentry,
+ flags);
+ if (err)
+ return err;
+
+ if (flags & RENAME_EXCHANGE)
+ return ubifs_xrename(old_dir, old_dentry, new_dir, new_dentry);
+
+ return do_rename(old_dir, old_dentry, new_dir, new_dentry, flags);
+}
+
+int ubifs_getattr(struct mnt_idmap *idmap, const struct path *path,
+ struct kstat *stat, u32 request_mask, unsigned int flags)
+{
+ loff_t size;
+ struct inode *inode = d_inode(path->dentry);
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ mutex_lock(&ui->ui_mutex);
+
+ if (ui->flags & UBIFS_APPEND_FL)
+ stat->attributes |= STATX_ATTR_APPEND;
+ if (ui->flags & UBIFS_COMPR_FL)
+ stat->attributes |= STATX_ATTR_COMPRESSED;
+ if (ui->flags & UBIFS_CRYPT_FL)
+ stat->attributes |= STATX_ATTR_ENCRYPTED;
+ if (ui->flags & UBIFS_IMMUTABLE_FL)
+ stat->attributes |= STATX_ATTR_IMMUTABLE;
+
+ stat->attributes_mask |= (STATX_ATTR_APPEND |
+ STATX_ATTR_COMPRESSED |
+ STATX_ATTR_ENCRYPTED |
+ STATX_ATTR_IMMUTABLE);
+
+ generic_fillattr(&nop_mnt_idmap, request_mask, inode, stat);
+ stat->blksize = UBIFS_BLOCK_SIZE;
+ stat->size = ui->ui_size;
+
+ /*
+ * Unfortunately, the 'stat()' system call was designed for block
+ * device based file systems, and it is not appropriate for UBIFS,
+ * because UBIFS does not have notion of "block". For example, it is
+ * difficult to tell how many block a directory takes - it actually
+ * takes less than 300 bytes, but we have to round it to block size,
+ * which introduces large mistake. This makes utilities like 'du' to
+ * report completely senseless numbers. This is the reason why UBIFS
+ * goes the same way as JFFS2 - it reports zero blocks for everything
+ * but regular files, which makes more sense than reporting completely
+ * wrong sizes.
+ */
+ if (S_ISREG(inode->i_mode)) {
+ size = ui->xattr_size;
+ size += stat->size;
+ size = ALIGN(size, UBIFS_BLOCK_SIZE);
+ /*
+ * Note, user-space expects 512-byte blocks count irrespectively
+ * of what was reported in @stat->size.
+ */
+ stat->blocks = size >> 9;
+ } else
+ stat->blocks = 0;
+ mutex_unlock(&ui->ui_mutex);
+ return 0;
+}
+
+const struct inode_operations ubifs_dir_inode_operations = {
+ .lookup = ubifs_lookup,
+ .create = ubifs_create,
+ .link = ubifs_link,
+ .symlink = ubifs_symlink,
+ .unlink = ubifs_unlink,
+ .mkdir = ubifs_mkdir,
+ .rmdir = ubifs_rmdir,
+ .mknod = ubifs_mknod,
+ .rename = ubifs_rename,
+ .setattr = ubifs_setattr,
+ .getattr = ubifs_getattr,
+ .listxattr = ubifs_listxattr,
+ .update_time = ubifs_update_time,
+ .tmpfile = ubifs_tmpfile,
+ .fileattr_get = ubifs_fileattr_get,
+ .fileattr_set = ubifs_fileattr_set,
+};
+
+const struct file_operations ubifs_dir_operations = {
+ .llseek = generic_file_llseek,
+ .release = ubifs_dir_release,
+ .read = generic_read_dir,
+ .iterate_shared = ubifs_readdir,
+ .fsync = ubifs_fsync,
+ .unlocked_ioctl = ubifs_ioctl,
+#ifdef CONFIG_COMPAT
+ .compat_ioctl = ubifs_compat_ioctl,
+#endif
+};
new file mode 100644
@@ -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
+ * @data: 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,
+ struct scan_data *data)
+{
+ 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,
+ (ubifs_lpt_scan_callback)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
+ * @data: 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,
+ struct scan_data *data)
+{
+ 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,
+ (ubifs_lpt_scan_callback)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
+ * @data: 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,
+ struct scan_data *data)
+{
+ 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,
+ (ubifs_lpt_scan_callback)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 struct ubifs_lprops **a,
+ const struct ubifs_lprops **b)
+{
+ const struct ubifs_lprops *lpa = *a;
+ const struct ubifs_lprops *lpb = *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 *),
+ (int (*)(const void *, const 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
+ * @data: 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,
+ struct scan_data *data)
+{
+ 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,
+ (ubifs_lpt_scan_callback)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;
+}
new file mode 100644
@@ -0,0 +1,1017 @@
+// 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 garbage collection. The procedure for garbage collection
+ * is different depending on whether a LEB as an index LEB (contains index
+ * nodes) or not. For non-index LEBs, garbage collection finds a LEB which
+ * contains a lot of dirty space (obsolete nodes), and copies the non-obsolete
+ * nodes to the journal, at which point the garbage-collected LEB is free to be
+ * reused. For index LEBs, garbage collection marks the non-obsolete index nodes
+ * dirty in the TNC, and after the next commit, the garbage-collected LEB is
+ * to be reused. Garbage collection will cause the number of dirty index nodes
+ * to grow, however sufficient space is reserved for the index to ensure the
+ * commit will never run out of space.
+ *
+ * Notes about dead watermark. At current UBIFS implementation we assume that
+ * LEBs which have less than @c->dead_wm bytes of free + dirty space are full
+ * and not worth garbage-collecting. The dead watermark is one min. I/O unit
+ * size, or min. UBIFS node size, depending on what is greater. Indeed, UBIFS
+ * Garbage Collector has to synchronize the GC head's write buffer before
+ * returning, so this is about wasting one min. I/O unit. However, UBIFS GC can
+ * actually reclaim even very small pieces of dirty space by garbage collecting
+ * enough dirty LEBs, but we do not bother doing this at this implementation.
+ *
+ * Notes about dark watermark. The results of GC work depends on how big are
+ * the UBIFS nodes GC deals with. Large nodes make GC waste more space. Indeed,
+ * if GC move data from LEB A to LEB B and nodes in LEB A are large, GC would
+ * have to waste large pieces of free space at the end of LEB B, because nodes
+ * from LEB A would not fit. And the worst situation is when all nodes are of
+ * maximum size. So dark watermark is the amount of free + dirty space in LEB
+ * which are guaranteed to be reclaimable. If LEB has less space, the GC might
+ * be unable to reclaim it. So, LEBs with free + dirty greater than dark
+ * watermark are "good" LEBs from GC's point of view. The other LEBs are not so
+ * good, and GC takes extra care when moving them.
+ */
+
+#include <linux/slab.h>
+#include <linux/pagemap.h>
+#include <linux/list_sort.h>
+#include "ubifs.h"
+
+/*
+ * GC may need to move more than one LEB to make progress. The below constants
+ * define "soft" and "hard" limits on the number of LEBs the garbage collector
+ * may move.
+ */
+#define SOFT_LEBS_LIMIT 4
+#define HARD_LEBS_LIMIT 32
+
+/**
+ * switch_gc_head - switch the garbage collection journal head.
+ * @c: UBIFS file-system description object
+ *
+ * This function switch the GC head to the next LEB which is reserved in
+ * @c->gc_lnum. Returns %0 in case of success, %-EAGAIN if commit is required,
+ * and other negative error code in case of failures.
+ */
+static int switch_gc_head(struct ubifs_info *c)
+{
+ int err, gc_lnum = c->gc_lnum;
+ struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf;
+
+ ubifs_assert(c, gc_lnum != -1);
+ dbg_gc("switch GC head from LEB %d:%d to LEB %d (waste %d bytes)",
+ wbuf->lnum, wbuf->offs + wbuf->used, gc_lnum,
+ c->leb_size - wbuf->offs - wbuf->used);
+
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ if (err)
+ return err;
+
+ /*
+ * The GC write-buffer was synchronized, we may safely unmap
+ * 'c->gc_lnum'.
+ */
+ err = ubifs_leb_unmap(c, gc_lnum);
+ if (err)
+ return err;
+
+ err = ubifs_add_bud_to_log(c, GCHD, gc_lnum, 0);
+ if (err)
+ return err;
+
+ c->gc_lnum = -1;
+ err = ubifs_wbuf_seek_nolock(wbuf, gc_lnum, 0);
+ return err;
+}
+
+/**
+ * data_nodes_cmp - compare 2 data nodes.
+ * @priv: UBIFS file-system description object
+ * @a: first data node
+ * @b: second data node
+ *
+ * This function compares data nodes @a and @b. Returns %1 if @a has greater
+ * inode or block number, and %-1 otherwise.
+ */
+static int data_nodes_cmp(void *priv, const struct list_head *a,
+ const struct list_head *b)
+{
+ ino_t inuma, inumb;
+ struct ubifs_info *c = priv;
+ struct ubifs_scan_node *sa, *sb;
+
+ cond_resched();
+ if (a == b)
+ return 0;
+
+ sa = list_entry(a, struct ubifs_scan_node, list);
+ sb = list_entry(b, struct ubifs_scan_node, list);
+
+ ubifs_assert(c, key_type(c, &sa->key) == UBIFS_DATA_KEY);
+ ubifs_assert(c, key_type(c, &sb->key) == UBIFS_DATA_KEY);
+ ubifs_assert(c, sa->type == UBIFS_DATA_NODE);
+ ubifs_assert(c, sb->type == UBIFS_DATA_NODE);
+
+ inuma = key_inum(c, &sa->key);
+ inumb = key_inum(c, &sb->key);
+
+ if (inuma == inumb) {
+ unsigned int blka = key_block(c, &sa->key);
+ unsigned int blkb = key_block(c, &sb->key);
+
+ if (blka <= blkb)
+ return -1;
+ } else if (inuma <= inumb)
+ return -1;
+
+ return 1;
+}
+
+/*
+ * nondata_nodes_cmp - compare 2 non-data nodes.
+ * @priv: UBIFS file-system description object
+ * @a: first node
+ * @a: second node
+ *
+ * This function compares nodes @a and @b. It makes sure that inode nodes go
+ * first and sorted by length in descending order. Directory entry nodes go
+ * after inode nodes and are sorted in ascending hash valuer order.
+ */
+static int nondata_nodes_cmp(void *priv, const struct list_head *a,
+ const struct list_head *b)
+{
+ ino_t inuma, inumb;
+ struct ubifs_info *c = priv;
+ struct ubifs_scan_node *sa, *sb;
+
+ cond_resched();
+ if (a == b)
+ return 0;
+
+ sa = list_entry(a, struct ubifs_scan_node, list);
+ sb = list_entry(b, struct ubifs_scan_node, list);
+
+ ubifs_assert(c, key_type(c, &sa->key) != UBIFS_DATA_KEY &&
+ key_type(c, &sb->key) != UBIFS_DATA_KEY);
+ ubifs_assert(c, sa->type != UBIFS_DATA_NODE &&
+ sb->type != UBIFS_DATA_NODE);
+
+ /* Inodes go before directory entries */
+ if (sa->type == UBIFS_INO_NODE) {
+ if (sb->type == UBIFS_INO_NODE)
+ return sb->len - sa->len;
+ return -1;
+ }
+ if (sb->type == UBIFS_INO_NODE)
+ return 1;
+
+ ubifs_assert(c, key_type(c, &sa->key) == UBIFS_DENT_KEY ||
+ key_type(c, &sa->key) == UBIFS_XENT_KEY);
+ ubifs_assert(c, key_type(c, &sb->key) == UBIFS_DENT_KEY ||
+ key_type(c, &sb->key) == UBIFS_XENT_KEY);
+ ubifs_assert(c, sa->type == UBIFS_DENT_NODE ||
+ sa->type == UBIFS_XENT_NODE);
+ ubifs_assert(c, sb->type == UBIFS_DENT_NODE ||
+ sb->type == UBIFS_XENT_NODE);
+
+ inuma = key_inum(c, &sa->key);
+ inumb = key_inum(c, &sb->key);
+
+ if (inuma == inumb) {
+ uint32_t hasha = key_hash(c, &sa->key);
+ uint32_t hashb = key_hash(c, &sb->key);
+
+ if (hasha <= hashb)
+ return -1;
+ } else if (inuma <= inumb)
+ return -1;
+
+ return 1;
+}
+
+/**
+ * sort_nodes - sort nodes for GC.
+ * @c: UBIFS file-system description object
+ * @sleb: describes nodes to sort and contains the result on exit
+ * @nondata: contains non-data nodes on exit
+ * @min: minimum node size is returned here
+ *
+ * This function sorts the list of inodes to garbage collect. First of all, it
+ * kills obsolete nodes and separates data and non-data nodes to the
+ * @sleb->nodes and @nondata lists correspondingly.
+ *
+ * Data nodes are then sorted in block number order - this is important for
+ * bulk-read; data nodes with lower inode number go before data nodes with
+ * higher inode number, and data nodes with lower block number go before data
+ * nodes with higher block number;
+ *
+ * Non-data nodes are sorted as follows.
+ * o First go inode nodes - they are sorted in descending length order.
+ * o Then go directory entry nodes - they are sorted in hash order, which
+ * should supposedly optimize 'readdir()'. Direntry nodes with lower parent
+ * inode number go before direntry nodes with higher parent inode number,
+ * and direntry nodes with lower name hash values go before direntry nodes
+ * with higher name hash values.
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+static int sort_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ struct list_head *nondata, int *min)
+{
+ int err;
+ struct ubifs_scan_node *snod, *tmp;
+
+ *min = INT_MAX;
+
+ /* Separate data nodes and non-data nodes */
+ list_for_each_entry_safe(snod, tmp, &sleb->nodes, list) {
+ ubifs_assert(c, snod->type == UBIFS_INO_NODE ||
+ snod->type == UBIFS_DATA_NODE ||
+ snod->type == UBIFS_DENT_NODE ||
+ snod->type == UBIFS_XENT_NODE ||
+ snod->type == UBIFS_TRUN_NODE ||
+ snod->type == UBIFS_AUTH_NODE);
+
+ if (snod->type != UBIFS_INO_NODE &&
+ snod->type != UBIFS_DATA_NODE &&
+ snod->type != UBIFS_DENT_NODE &&
+ snod->type != UBIFS_XENT_NODE) {
+ /* Probably truncation node, zap it */
+ list_del(&snod->list);
+ kfree(snod);
+ continue;
+ }
+
+ ubifs_assert(c, key_type(c, &snod->key) == UBIFS_DATA_KEY ||
+ key_type(c, &snod->key) == UBIFS_INO_KEY ||
+ key_type(c, &snod->key) == UBIFS_DENT_KEY ||
+ key_type(c, &snod->key) == UBIFS_XENT_KEY);
+
+ err = ubifs_tnc_has_node(c, &snod->key, 0, sleb->lnum,
+ snod->offs, 0);
+ if (err < 0)
+ return err;
+
+ if (!err) {
+ /* The node is obsolete, remove it from the list */
+ list_del(&snod->list);
+ kfree(snod);
+ continue;
+ }
+
+ if (snod->len < *min)
+ *min = snod->len;
+
+ if (key_type(c, &snod->key) != UBIFS_DATA_KEY)
+ list_move_tail(&snod->list, nondata);
+ }
+
+ /* Sort data and non-data nodes */
+ list_sort(c, &sleb->nodes, &data_nodes_cmp);
+ list_sort(c, nondata, &nondata_nodes_cmp);
+
+ err = dbg_check_data_nodes_order(c, &sleb->nodes);
+ if (err)
+ return err;
+ err = dbg_check_nondata_nodes_order(c, nondata);
+ if (err)
+ return err;
+ return 0;
+}
+
+/**
+ * move_node - move a node.
+ * @c: UBIFS file-system description object
+ * @sleb: describes the LEB to move nodes from
+ * @snod: the mode to move
+ * @wbuf: write-buffer to move node to
+ *
+ * This function moves node @snod to @wbuf, changes TNC correspondingly, and
+ * destroys @snod. Returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+static int move_node(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
+ struct ubifs_scan_node *snod, struct ubifs_wbuf *wbuf)
+{
+ int err, new_lnum = wbuf->lnum, new_offs = wbuf->offs + wbuf->used;
+
+ cond_resched();
+ err = ubifs_wbuf_write_nolock(wbuf, snod->node, snod->len);
+ if (err)
+ return err;
+
+ err = ubifs_tnc_replace(c, &snod->key, sleb->lnum,
+ snod->offs, new_lnum, new_offs,
+ snod->len);
+ list_del(&snod->list);
+ kfree(snod);
+ return err;
+}
+
+/**
+ * move_nodes - move nodes.
+ * @c: UBIFS file-system description object
+ * @sleb: describes the LEB to move nodes from
+ *
+ * This function moves valid nodes from data LEB described by @sleb to the GC
+ * journal head. This function returns zero in case of success, %-EAGAIN if
+ * commit is required, and other negative error codes in case of other
+ * failures.
+ */
+static int move_nodes(struct ubifs_info *c, struct ubifs_scan_leb *sleb)
+{
+ int err, min;
+ LIST_HEAD(nondata);
+ struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf;
+
+ if (wbuf->lnum == -1) {
+ /*
+ * The GC journal head is not set, because it is the first GC
+ * invocation since mount.
+ */
+ err = switch_gc_head(c);
+ if (err)
+ return err;
+ }
+
+ err = sort_nodes(c, sleb, &nondata, &min);
+ if (err)
+ goto out;
+
+ /* Write nodes to their new location. Use the first-fit strategy */
+ while (1) {
+ int avail, moved = 0;
+ struct ubifs_scan_node *snod, *tmp;
+
+ /* Move data nodes */
+ list_for_each_entry_safe(snod, tmp, &sleb->nodes, list) {
+ avail = c->leb_size - wbuf->offs - wbuf->used -
+ ubifs_auth_node_sz(c);
+ if (snod->len > avail)
+ /*
+ * Do not skip data nodes in order to optimize
+ * bulk-read.
+ */
+ break;
+
+ err = ubifs_shash_update(c, c->jheads[GCHD].log_hash,
+ snod->node, snod->len);
+ if (err)
+ goto out;
+
+ err = move_node(c, sleb, snod, wbuf);
+ if (err)
+ goto out;
+ moved = 1;
+ }
+
+ /* Move non-data nodes */
+ list_for_each_entry_safe(snod, tmp, &nondata, list) {
+ avail = c->leb_size - wbuf->offs - wbuf->used -
+ ubifs_auth_node_sz(c);
+ if (avail < min)
+ break;
+
+ if (snod->len > avail) {
+ /*
+ * Keep going only if this is an inode with
+ * some data. Otherwise stop and switch the GC
+ * head. IOW, we assume that data-less inode
+ * nodes and direntry nodes are roughly of the
+ * same size.
+ */
+ if (key_type(c, &snod->key) == UBIFS_DENT_KEY ||
+ snod->len == UBIFS_INO_NODE_SZ)
+ break;
+ continue;
+ }
+
+ err = ubifs_shash_update(c, c->jheads[GCHD].log_hash,
+ snod->node, snod->len);
+ if (err)
+ goto out;
+
+ err = move_node(c, sleb, snod, wbuf);
+ if (err)
+ goto out;
+ moved = 1;
+ }
+
+ if (ubifs_authenticated(c) && moved) {
+ struct ubifs_auth_node *auth;
+
+ auth = kmalloc(ubifs_auth_node_sz(c), GFP_NOFS);
+ if (!auth) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ err = ubifs_prepare_auth_node(c, auth,
+ c->jheads[GCHD].log_hash);
+ if (err) {
+ kfree(auth);
+ goto out;
+ }
+
+ err = ubifs_wbuf_write_nolock(wbuf, auth,
+ ubifs_auth_node_sz(c));
+ if (err) {
+ kfree(auth);
+ goto out;
+ }
+
+ ubifs_add_dirt(c, wbuf->lnum, ubifs_auth_node_sz(c));
+ }
+
+ if (list_empty(&sleb->nodes) && list_empty(&nondata))
+ break;
+
+ /*
+ * Waste the rest of the space in the LEB and switch to the
+ * next LEB.
+ */
+ err = switch_gc_head(c);
+ if (err)
+ goto out;
+ }
+
+ return 0;
+
+out:
+ list_splice_tail(&nondata, &sleb->nodes);
+ return err;
+}
+
+/**
+ * gc_sync_wbufs - sync write-buffers for GC.
+ * @c: UBIFS file-system description object
+ *
+ * We must guarantee that obsoleting nodes are on flash. Unfortunately they may
+ * be in a write-buffer instead. That is, a node could be written to a
+ * write-buffer, obsoleting another node in a LEB that is GC'd. If that LEB is
+ * erased before the write-buffer is sync'd and then there is an unclean
+ * unmount, then an existing node is lost. To avoid this, we sync all
+ * write-buffers.
+ *
+ * This function returns %0 on success or a negative error code on failure.
+ */
+static int gc_sync_wbufs(struct ubifs_info *c)
+{
+ int err, i;
+
+ for (i = 0; i < c->jhead_cnt; i++) {
+ if (i == GCHD)
+ continue;
+ err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
+ if (err)
+ return err;
+ }
+ return 0;
+}
+
+/**
+ * ubifs_garbage_collect_leb - garbage-collect a logical eraseblock.
+ * @c: UBIFS file-system description object
+ * @lp: describes the LEB to garbage collect
+ *
+ * This function garbage-collects an LEB and returns one of the @LEB_FREED,
+ * @LEB_RETAINED, etc positive codes in case of success, %-EAGAIN if commit is
+ * required, and other negative error codes in case of failures.
+ */
+int ubifs_garbage_collect_leb(struct ubifs_info *c, struct ubifs_lprops *lp)
+{
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf;
+ int err = 0, lnum = lp->lnum;
+
+ ubifs_assert(c, c->gc_lnum != -1 || wbuf->offs + wbuf->used == 0 ||
+ c->need_recovery);
+ ubifs_assert(c, c->gc_lnum != lnum);
+ ubifs_assert(c, wbuf->lnum != lnum);
+
+ if (lp->free + lp->dirty == c->leb_size) {
+ /* Special case - a free LEB */
+ dbg_gc("LEB %d is free, return it", lp->lnum);
+ ubifs_assert(c, !(lp->flags & LPROPS_INDEX));
+
+ if (lp->free != c->leb_size) {
+ /*
+ * Write buffers must be sync'd before unmapping
+ * freeable LEBs, because one of them may contain data
+ * which obsoletes something in 'lp->lnum'.
+ */
+ err = gc_sync_wbufs(c);
+ if (err)
+ return err;
+ err = ubifs_change_one_lp(c, lp->lnum, c->leb_size,
+ 0, 0, 0, 0);
+ if (err)
+ return err;
+ }
+ err = ubifs_leb_unmap(c, lp->lnum);
+ if (err)
+ return err;
+
+ if (c->gc_lnum == -1) {
+ c->gc_lnum = lnum;
+ return LEB_RETAINED;
+ }
+
+ return LEB_FREED;
+ }
+
+ /*
+ * We scan the entire LEB even though we only really need to scan up to
+ * (c->leb_size - lp->free).
+ */
+ sleb = ubifs_scan(c, lnum, 0, c->sbuf, 0);
+ if (IS_ERR(sleb))
+ return PTR_ERR(sleb);
+
+ ubifs_assert(c, !list_empty(&sleb->nodes));
+ snod = list_entry(sleb->nodes.next, struct ubifs_scan_node, list);
+
+ if (snod->type == UBIFS_IDX_NODE) {
+ struct ubifs_gced_idx_leb *idx_gc;
+
+ dbg_gc("indexing LEB %d (free %d, dirty %d)",
+ lnum, lp->free, lp->dirty);
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ struct ubifs_idx_node *idx = snod->node;
+ int level = le16_to_cpu(idx->level);
+
+ ubifs_assert(c, snod->type == UBIFS_IDX_NODE);
+ key_read(c, ubifs_idx_key(c, idx), &snod->key);
+ err = ubifs_dirty_idx_node(c, &snod->key, level, lnum,
+ snod->offs);
+ if (err)
+ goto out;
+ }
+
+ idx_gc = kmalloc(sizeof(struct ubifs_gced_idx_leb), GFP_NOFS);
+ if (!idx_gc) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ idx_gc->lnum = lnum;
+ idx_gc->unmap = 0;
+ list_add(&idx_gc->list, &c->idx_gc);
+
+ /*
+ * Don't release the LEB until after the next commit, because
+ * it may contain data which is needed for recovery. So
+ * although we freed this LEB, it will become usable only after
+ * the commit.
+ */
+ err = ubifs_change_one_lp(c, lnum, c->leb_size, 0, 0,
+ LPROPS_INDEX, 1);
+ if (err)
+ goto out;
+ err = LEB_FREED_IDX;
+ } else {
+ dbg_gc("data LEB %d (free %d, dirty %d)",
+ lnum, lp->free, lp->dirty);
+
+ err = move_nodes(c, sleb);
+ if (err)
+ goto out_inc_seq;
+
+ err = gc_sync_wbufs(c);
+ if (err)
+ goto out_inc_seq;
+
+ err = ubifs_change_one_lp(c, lnum, c->leb_size, 0, 0, 0, 0);
+ if (err)
+ goto out_inc_seq;
+
+ /* Allow for races with TNC */
+ c->gced_lnum = lnum;
+ smp_wmb();
+ c->gc_seq += 1;
+ smp_wmb();
+
+ if (c->gc_lnum == -1) {
+ c->gc_lnum = lnum;
+ err = LEB_RETAINED;
+ } else {
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ if (err)
+ goto out;
+
+ err = ubifs_leb_unmap(c, lnum);
+ if (err)
+ goto out;
+
+ err = LEB_FREED;
+ }
+ }
+
+out:
+ ubifs_scan_destroy(sleb);
+ return err;
+
+out_inc_seq:
+ /* We may have moved at least some nodes so allow for races with TNC */
+ c->gced_lnum = lnum;
+ smp_wmb();
+ c->gc_seq += 1;
+ smp_wmb();
+ goto out;
+}
+
+/**
+ * ubifs_garbage_collect - UBIFS garbage collector.
+ * @c: UBIFS file-system description object
+ * @anyway: do GC even if there are free LEBs
+ *
+ * This function does out-of-place garbage collection. The return codes are:
+ * o positive LEB number if the LEB has been freed and may be used;
+ * o %-EAGAIN if the caller has to run commit;
+ * o %-ENOSPC if GC failed to make any progress;
+ * o other negative error codes in case of other errors.
+ *
+ * Garbage collector writes data to the journal when GC'ing data LEBs, and just
+ * marking indexing nodes dirty when GC'ing indexing LEBs. Thus, at some point
+ * commit may be required. But commit cannot be run from inside GC, because the
+ * caller might be holding the commit lock, so %-EAGAIN is returned instead;
+ * And this error code means that the caller has to run commit, and re-run GC
+ * if there is still no free space.
+ *
+ * There are many reasons why this function may return %-EAGAIN:
+ * o the log is full and there is no space to write an LEB reference for
+ * @c->gc_lnum;
+ * o the journal is too large and exceeds size limitations;
+ * o GC moved indexing LEBs, but they can be used only after the commit;
+ * o the shrinker fails to find clean znodes to free and requests the commit;
+ * o etc.
+ *
+ * Note, if the file-system is close to be full, this function may return
+ * %-EAGAIN infinitely, so the caller has to limit amount of re-invocations of
+ * the function. E.g., this happens if the limits on the journal size are too
+ * tough and GC writes too much to the journal before an LEB is freed. This
+ * might also mean that the journal is too large, and the TNC becomes to big,
+ * so that the shrinker is constantly called, finds not clean znodes to free,
+ * and requests commit. Well, this may also happen if the journal is all right,
+ * but another kernel process consumes too much memory. Anyway, infinite
+ * %-EAGAIN may happen, but in some extreme/misconfiguration cases.
+ */
+int ubifs_garbage_collect(struct ubifs_info *c, int anyway)
+{
+ int i, err, ret, min_space = c->dead_wm;
+ struct ubifs_lprops lp;
+ struct ubifs_wbuf *wbuf = &c->jheads[GCHD].wbuf;
+
+ ubifs_assert_cmt_locked(c);
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+
+ if (ubifs_gc_should_commit(c))
+ return -EAGAIN;
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+
+ if (c->ro_error) {
+ ret = -EROFS;
+ goto out_unlock;
+ }
+
+ /* We expect the write-buffer to be empty on entry */
+ ubifs_assert(c, !wbuf->used);
+
+ for (i = 0; ; i++) {
+ int space_before, space_after;
+
+ /* Maybe continue after find and break before find */
+ lp.lnum = -1;
+
+ cond_resched();
+
+ /* Give the commit an opportunity to run */
+ if (ubifs_gc_should_commit(c)) {
+ ret = -EAGAIN;
+ break;
+ }
+
+ if (i > SOFT_LEBS_LIMIT && !list_empty(&c->idx_gc)) {
+ /*
+ * We've done enough iterations. Indexing LEBs were
+ * moved and will be available after the commit.
+ */
+ dbg_gc("soft limit, some index LEBs GC'ed, -EAGAIN");
+ ubifs_commit_required(c);
+ ret = -EAGAIN;
+ break;
+ }
+
+ if (i > HARD_LEBS_LIMIT) {
+ /*
+ * We've moved too many LEBs and have not made
+ * progress, give up.
+ */
+ dbg_gc("hard limit, -ENOSPC");
+ ret = -ENOSPC;
+ break;
+ }
+
+ /*
+ * Empty and freeable LEBs can turn up while we waited for
+ * the wbuf lock, or while we have been running GC. In that
+ * case, we should just return one of those instead of
+ * continuing to GC dirty LEBs. Hence we request
+ * 'ubifs_find_dirty_leb()' to return an empty LEB if it can.
+ */
+ ret = ubifs_find_dirty_leb(c, &lp, min_space, anyway ? 0 : 1);
+ if (ret) {
+ if (ret == -ENOSPC)
+ dbg_gc("no more dirty LEBs");
+ break;
+ }
+
+ dbg_gc("found LEB %d: free %d, dirty %d, sum %d (min. space %d)",
+ lp.lnum, lp.free, lp.dirty, lp.free + lp.dirty,
+ min_space);
+
+ space_before = c->leb_size - wbuf->offs - wbuf->used;
+ if (wbuf->lnum == -1)
+ space_before = 0;
+
+ ret = ubifs_garbage_collect_leb(c, &lp);
+ if (ret < 0) {
+ if (ret == -EAGAIN) {
+ /*
+ * This is not error, so we have to return the
+ * LEB to lprops. But if 'ubifs_return_leb()'
+ * fails, its failure code is propagated to the
+ * caller instead of the original '-EAGAIN'.
+ */
+ err = ubifs_return_leb(c, lp.lnum);
+ if (err) {
+ ret = err;
+ /*
+ * An LEB may always be "taken",
+ * so setting ubifs to read-only,
+ * and then executing sync wbuf will
+ * return -EROFS and enter the "out"
+ * error branch.
+ */
+ ubifs_ro_mode(c, ret);
+ }
+ /* Maybe double return LEB if goto out */
+ lp.lnum = -1;
+ break;
+ }
+ goto out;
+ }
+
+ if (ret == LEB_FREED) {
+ /* An LEB has been freed and is ready for use */
+ dbg_gc("LEB %d freed, return", lp.lnum);
+ ret = lp.lnum;
+ break;
+ }
+
+ if (ret == LEB_FREED_IDX) {
+ /*
+ * This was an indexing LEB and it cannot be
+ * immediately used. And instead of requesting the
+ * commit straight away, we try to garbage collect some
+ * more.
+ */
+ dbg_gc("indexing LEB %d freed, continue", lp.lnum);
+ continue;
+ }
+
+ ubifs_assert(c, ret == LEB_RETAINED);
+ space_after = c->leb_size - wbuf->offs - wbuf->used;
+ dbg_gc("LEB %d retained, freed %d bytes", lp.lnum,
+ space_after - space_before);
+
+ if (space_after > space_before) {
+ /* GC makes progress, keep working */
+ min_space >>= 1;
+ if (min_space < c->dead_wm)
+ min_space = c->dead_wm;
+ continue;
+ }
+
+ dbg_gc("did not make progress");
+
+ /*
+ * GC moved an LEB bud have not done any progress. This means
+ * that the previous GC head LEB contained too few free space
+ * and the LEB which was GC'ed contained only large nodes which
+ * did not fit that space.
+ *
+ * We can do 2 things:
+ * 1. pick another LEB in a hope it'll contain a small node
+ * which will fit the space we have at the end of current GC
+ * head LEB, but there is no guarantee, so we try this out
+ * unless we have already been working for too long;
+ * 2. request an LEB with more dirty space, which will force
+ * 'ubifs_find_dirty_leb()' to start scanning the lprops
+ * table, instead of just picking one from the heap
+ * (previously it already picked the dirtiest LEB).
+ */
+ if (i < SOFT_LEBS_LIMIT) {
+ dbg_gc("try again");
+ continue;
+ }
+
+ min_space <<= 1;
+ if (min_space > c->dark_wm)
+ min_space = c->dark_wm;
+ dbg_gc("set min. space to %d", min_space);
+ }
+
+ if (ret == -ENOSPC && !list_empty(&c->idx_gc)) {
+ dbg_gc("no space, some index LEBs GC'ed, -EAGAIN");
+ ubifs_commit_required(c);
+ ret = -EAGAIN;
+ }
+
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ if (!err)
+ err = ubifs_leb_unmap(c, c->gc_lnum);
+ if (err) {
+ ret = err;
+ goto out;
+ }
+out_unlock:
+ mutex_unlock(&wbuf->io_mutex);
+ return ret;
+
+out:
+ ubifs_assert(c, ret < 0);
+ ubifs_assert(c, ret != -ENOSPC && ret != -EAGAIN);
+ ubifs_wbuf_sync_nolock(wbuf);
+ ubifs_ro_mode(c, ret);
+ mutex_unlock(&wbuf->io_mutex);
+ if (lp.lnum != -1)
+ ubifs_return_leb(c, lp.lnum);
+ return ret;
+}
+
+/**
+ * ubifs_gc_start_commit - garbage collection at start of commit.
+ * @c: UBIFS file-system description object
+ *
+ * If a LEB has only dirty and free space, then we may safely unmap it and make
+ * it free. Note, we cannot do this with indexing LEBs because dirty space may
+ * correspond index nodes that are required for recovery. In that case, the
+ * LEB cannot be unmapped until after the next commit.
+ *
+ * This function returns %0 upon success and a negative error code upon failure.
+ */
+int ubifs_gc_start_commit(struct ubifs_info *c)
+{
+ struct ubifs_gced_idx_leb *idx_gc;
+ const struct ubifs_lprops *lp;
+ int err = 0, flags;
+
+ ubifs_get_lprops(c);
+
+ /*
+ * Unmap (non-index) freeable LEBs. Note that recovery requires that all
+ * wbufs are sync'd before this, which is done in 'do_commit()'.
+ */
+ while (1) {
+ lp = ubifs_fast_find_freeable(c);
+ if (!lp)
+ break;
+ ubifs_assert(c, !(lp->flags & LPROPS_TAKEN));
+ ubifs_assert(c, !(lp->flags & LPROPS_INDEX));
+ err = ubifs_leb_unmap(c, lp->lnum);
+ if (err)
+ goto out;
+ lp = ubifs_change_lp(c, lp, c->leb_size, 0, lp->flags, 0);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+ ubifs_assert(c, !(lp->flags & LPROPS_TAKEN));
+ ubifs_assert(c, !(lp->flags & LPROPS_INDEX));
+ }
+
+ /* Mark GC'd index LEBs OK to unmap after this commit finishes */
+ list_for_each_entry(idx_gc, &c->idx_gc, list)
+ idx_gc->unmap = 1;
+
+ /* Record index freeable LEBs for unmapping after commit */
+ while (1) {
+ lp = ubifs_fast_find_frdi_idx(c);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+ if (!lp)
+ break;
+ idx_gc = kmalloc(sizeof(struct ubifs_gced_idx_leb), GFP_NOFS);
+ if (!idx_gc) {
+ err = -ENOMEM;
+ goto out;
+ }
+ ubifs_assert(c, !(lp->flags & LPROPS_TAKEN));
+ ubifs_assert(c, lp->flags & LPROPS_INDEX);
+ /* Don't release the LEB until after the next commit */
+ flags = (lp->flags | LPROPS_TAKEN) ^ LPROPS_INDEX;
+ lp = ubifs_change_lp(c, lp, c->leb_size, 0, flags, 1);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ kfree(idx_gc);
+ goto out;
+ }
+ ubifs_assert(c, lp->flags & LPROPS_TAKEN);
+ ubifs_assert(c, !(lp->flags & LPROPS_INDEX));
+ idx_gc->lnum = lp->lnum;
+ idx_gc->unmap = 1;
+ list_add(&idx_gc->list, &c->idx_gc);
+ }
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_gc_end_commit - garbage collection at end of commit.
+ * @c: UBIFS file-system description object
+ *
+ * This function completes out-of-place garbage collection of index LEBs.
+ */
+int ubifs_gc_end_commit(struct ubifs_info *c)
+{
+ struct ubifs_gced_idx_leb *idx_gc, *tmp;
+ struct ubifs_wbuf *wbuf;
+ int err = 0;
+
+ wbuf = &c->jheads[GCHD].wbuf;
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ list_for_each_entry_safe(idx_gc, tmp, &c->idx_gc, list)
+ if (idx_gc->unmap) {
+ dbg_gc("LEB %d", idx_gc->lnum);
+ err = ubifs_leb_unmap(c, idx_gc->lnum);
+ if (err)
+ goto out;
+ err = ubifs_change_one_lp(c, idx_gc->lnum, LPROPS_NC,
+ LPROPS_NC, 0, LPROPS_TAKEN, -1);
+ if (err)
+ goto out;
+ list_del(&idx_gc->list);
+ kfree(idx_gc);
+ }
+out:
+ mutex_unlock(&wbuf->io_mutex);
+ return err;
+}
+
+/**
+ * ubifs_destroy_idx_gc - destroy idx_gc list.
+ * @c: UBIFS file-system description object
+ *
+ * This function destroys the @c->idx_gc list. It is called when unmounting
+ * so locks are not needed. Returns zero in case of success and a negative
+ * error code in case of failure.
+ */
+void ubifs_destroy_idx_gc(struct ubifs_info *c)
+{
+ while (!list_empty(&c->idx_gc)) {
+ struct ubifs_gced_idx_leb *idx_gc;
+
+ idx_gc = list_entry(c->idx_gc.next, struct ubifs_gced_idx_leb,
+ list);
+ c->idx_gc_cnt -= 1;
+ list_del(&idx_gc->list);
+ kfree(idx_gc);
+ }
+}
+
+/**
+ * ubifs_get_idx_gc_leb - get a LEB from GC'd index LEB list.
+ * @c: UBIFS file-system description object
+ *
+ * Called during start commit so locks are not needed.
+ */
+int ubifs_get_idx_gc_leb(struct ubifs_info *c)
+{
+ struct ubifs_gced_idx_leb *idx_gc;
+ int lnum;
+
+ if (list_empty(&c->idx_gc))
+ return -ENOSPC;
+ idx_gc = list_entry(c->idx_gc.next, struct ubifs_gced_idx_leb, list);
+ lnum = idx_gc->lnum;
+ /* c->idx_gc_cnt is updated by the caller when lprops are updated */
+ list_del(&idx_gc->list);
+ kfree(idx_gc);
+ return lnum;
+}
new file mode 100644
@@ -0,0 +1,1268 @@
+// 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: Artem Bityutskiy (Битюцкий Артём)
+ * Adrian Hunter
+ * Zoltan Sogor
+ */
+
+/*
+ * This file implements UBIFS I/O subsystem which provides various I/O-related
+ * helper functions (reading/writing/checking/validating nodes) and implements
+ * write-buffering support. Write buffers help to save space which otherwise
+ * would have been wasted for padding to the nearest minimal I/O unit boundary.
+ * Instead, data first goes to the write-buffer and is flushed when the
+ * buffer is full or when it is not used for some time (by timer). This is
+ * similar to the mechanism is used by JFFS2.
+ *
+ * UBIFS distinguishes between minimum write size (@c->min_io_size) and maximum
+ * write size (@c->max_write_size). The latter is the maximum amount of bytes
+ * the underlying flash is able to program at a time, and writing in
+ * @c->max_write_size units should presumably be faster. Obviously,
+ * @c->min_io_size <= @c->max_write_size. Write-buffers are of
+ * @c->max_write_size bytes in size for maximum performance. However, when a
+ * write-buffer is flushed, only the portion of it (aligned to @c->min_io_size
+ * boundary) which contains data is written, not the whole write-buffer,
+ * because this is more space-efficient.
+ *
+ * This optimization adds few complications to the code. Indeed, on the one
+ * hand, we want to write in optimal @c->max_write_size bytes chunks, which
+ * also means aligning writes at the @c->max_write_size bytes offsets. On the
+ * other hand, we do not want to waste space when synchronizing the write
+ * buffer, so during synchronization we writes in smaller chunks. And this makes
+ * the next write offset to be not aligned to @c->max_write_size bytes. So the
+ * have to make sure that the write-buffer offset (@wbuf->offs) becomes aligned
+ * to @c->max_write_size bytes again. We do this by temporarily shrinking
+ * write-buffer size (@wbuf->size).
+ *
+ * Write-buffers are defined by 'struct ubifs_wbuf' objects and protected by
+ * mutexes defined inside these objects. Since sometimes upper-level code
+ * has to lock the write-buffer (e.g. journal space reservation code), many
+ * functions related to write-buffers have "nolock" suffix which means that the
+ * caller has to lock the write-buffer before calling this function.
+ *
+ * UBIFS stores nodes at 64 bit-aligned addresses. If the node length is not
+ * aligned, UBIFS starts the next node from the aligned address, and the padded
+ * bytes may contain any rubbish. In other words, UBIFS does not put padding
+ * bytes in those small gaps. Common headers of nodes store real node lengths,
+ * not aligned lengths. Indexing nodes also store real lengths in branches.
+ *
+ * UBIFS uses padding when it pads to the next min. I/O unit. In this case it
+ * uses padding nodes or padding bytes, if the padding node does not fit.
+ *
+ * All UBIFS nodes are protected by CRC checksums and UBIFS checks CRC when
+ * they are read from the flash media.
+ */
+
+#include <linux/crc32.h>
+#include <linux/slab.h>
+#include "ubifs.h"
+
+/**
+ * ubifs_ro_mode - switch UBIFS to read read-only mode.
+ * @c: UBIFS file-system description object
+ * @err: error code which is the reason of switching to R/O mode
+ */
+void ubifs_ro_mode(struct ubifs_info *c, int err)
+{
+ if (!c->ro_error) {
+ c->ro_error = 1;
+ c->no_chk_data_crc = 0;
+ c->vfs_sb->s_flags |= SB_RDONLY;
+ ubifs_warn(c, "switched to read-only mode, error %d", err);
+ dump_stack();
+ }
+}
+
+/*
+ * Below are simple wrappers over UBI I/O functions which include some
+ * additional checks and UBIFS debugging stuff. See corresponding UBI function
+ * for more information.
+ */
+
+int ubifs_leb_read(const struct ubifs_info *c, int lnum, void *buf, int offs,
+ int len, int even_ebadmsg)
+{
+ int err;
+
+ err = ubi_read(c->ubi, lnum, buf, offs, len);
+ /*
+ * In case of %-EBADMSG print the error message only if the
+ * @even_ebadmsg is true.
+ */
+ if (err && (err != -EBADMSG || even_ebadmsg)) {
+ ubifs_err(c, "reading %d bytes from LEB %d:%d failed, error %d",
+ len, lnum, offs, err);
+ dump_stack();
+ }
+ return err;
+}
+
+int ubifs_leb_write(struct ubifs_info *c, int lnum, const void *buf, int offs,
+ int len)
+{
+ int err;
+
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (c->ro_error)
+ return -EROFS;
+ if (!dbg_is_tst_rcvry(c))
+ err = ubi_leb_write(c->ubi, lnum, buf, offs, len);
+ else
+ err = dbg_leb_write(c, lnum, buf, offs, len);
+ if (err) {
+ ubifs_err(c, "writing %d bytes to LEB %d:%d failed, error %d",
+ len, lnum, offs, err);
+ ubifs_ro_mode(c, err);
+ dump_stack();
+ }
+ return err;
+}
+
+int ubifs_leb_change(struct ubifs_info *c, int lnum, const void *buf, int len)
+{
+ int err;
+
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (c->ro_error)
+ return -EROFS;
+ if (!dbg_is_tst_rcvry(c))
+ err = ubi_leb_change(c->ubi, lnum, buf, len);
+ else
+ err = dbg_leb_change(c, lnum, buf, len);
+ if (err) {
+ ubifs_err(c, "changing %d bytes in LEB %d failed, error %d",
+ len, lnum, err);
+ ubifs_ro_mode(c, err);
+ dump_stack();
+ }
+ return err;
+}
+
+int ubifs_leb_unmap(struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (c->ro_error)
+ return -EROFS;
+ if (!dbg_is_tst_rcvry(c))
+ err = ubi_leb_unmap(c->ubi, lnum);
+ else
+ err = dbg_leb_unmap(c, lnum);
+ if (err) {
+ ubifs_err(c, "unmap LEB %d failed, error %d", lnum, err);
+ ubifs_ro_mode(c, err);
+ dump_stack();
+ }
+ return err;
+}
+
+int ubifs_leb_map(struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (c->ro_error)
+ return -EROFS;
+ if (!dbg_is_tst_rcvry(c))
+ err = ubi_leb_map(c->ubi, lnum);
+ else
+ err = dbg_leb_map(c, lnum);
+ if (err) {
+ ubifs_err(c, "mapping LEB %d failed, error %d", lnum, err);
+ ubifs_ro_mode(c, err);
+ dump_stack();
+ }
+ return err;
+}
+
+int ubifs_is_mapped(const struct ubifs_info *c, int lnum)
+{
+ int err;
+
+ err = ubi_is_mapped(c->ubi, lnum);
+ if (err < 0) {
+ ubifs_err(c, "ubi_is_mapped failed for LEB %d, error %d",
+ lnum, err);
+ dump_stack();
+ }
+ return err;
+}
+
+static void record_magic_error(struct ubifs_stats_info *stats)
+{
+ if (stats)
+ stats->magic_errors++;
+}
+
+static void record_node_error(struct ubifs_stats_info *stats)
+{
+ if (stats)
+ stats->node_errors++;
+}
+
+static void record_crc_error(struct ubifs_stats_info *stats)
+{
+ if (stats)
+ stats->crc_errors++;
+}
+
+/**
+ * ubifs_check_node - check node.
+ * @c: UBIFS file-system description object
+ * @buf: node to check
+ * @len: node length
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ * @quiet: print no messages
+ * @must_chk_crc: indicates whether to always check the CRC
+ *
+ * This function checks node magic number and CRC checksum. This function also
+ * validates node length to prevent UBIFS from becoming crazy when an attacker
+ * feeds it a file-system image with incorrect nodes. For example, too large
+ * node length in the common header could cause UBIFS to read memory outside of
+ * allocated buffer when checking the CRC checksum.
+ *
+ * This function may skip data nodes CRC checking if @c->no_chk_data_crc is
+ * true, which is controlled by corresponding UBIFS mount option. However, if
+ * @must_chk_crc is true, then @c->no_chk_data_crc is ignored and CRC is
+ * checked. Similarly, if @c->mounting or @c->remounting_rw is true (we are
+ * mounting or re-mounting to R/W mode), @c->no_chk_data_crc is ignored and CRC
+ * is checked. This is because during mounting or re-mounting from R/O mode to
+ * R/W mode we may read journal nodes (when replying the journal or doing the
+ * recovery) and the journal nodes may potentially be corrupted, so checking is
+ * required.
+ *
+ * This function returns zero in case of success and %-EUCLEAN in case of bad
+ * CRC or magic.
+ */
+int ubifs_check_node(const struct ubifs_info *c, const void *buf, int len,
+ int lnum, int offs, int quiet, int must_chk_crc)
+{
+ int err = -EINVAL, type, node_len;
+ uint32_t crc, node_crc, magic;
+ const struct ubifs_ch *ch = buf;
+
+ ubifs_assert(c, lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(c, !(offs & 7) && offs < c->leb_size);
+
+ magic = le32_to_cpu(ch->magic);
+ if (magic != UBIFS_NODE_MAGIC) {
+ if (!quiet)
+ ubifs_err(c, "bad magic %#08x, expected %#08x",
+ magic, UBIFS_NODE_MAGIC);
+ record_magic_error(c->stats);
+ err = -EUCLEAN;
+ goto out;
+ }
+
+ type = ch->node_type;
+ if (type < 0 || type >= UBIFS_NODE_TYPES_CNT) {
+ if (!quiet)
+ ubifs_err(c, "bad node type %d", type);
+ record_node_error(c->stats);
+ goto out;
+ }
+
+ node_len = le32_to_cpu(ch->len);
+ if (node_len + offs > c->leb_size)
+ goto out_len;
+
+ if (c->ranges[type].max_len == 0) {
+ if (node_len != c->ranges[type].len)
+ goto out_len;
+ } else if (node_len < c->ranges[type].min_len ||
+ node_len > c->ranges[type].max_len)
+ goto out_len;
+
+ if (!must_chk_crc && type == UBIFS_DATA_NODE && !c->mounting &&
+ !c->remounting_rw && c->no_chk_data_crc)
+ return 0;
+
+ crc = crc32(UBIFS_CRC32_INIT, buf + 8, node_len - 8);
+ node_crc = le32_to_cpu(ch->crc);
+ if (crc != node_crc) {
+ if (!quiet)
+ ubifs_err(c, "bad CRC: calculated %#08x, read %#08x",
+ crc, node_crc);
+ record_crc_error(c->stats);
+ err = -EUCLEAN;
+ goto out;
+ }
+
+ return 0;
+
+out_len:
+ if (!quiet)
+ ubifs_err(c, "bad node length %d", node_len);
+out:
+ if (!quiet) {
+ ubifs_err(c, "bad node at LEB %d:%d", lnum, offs);
+ ubifs_dump_node(c, buf, len);
+ dump_stack();
+ }
+ return err;
+}
+
+/**
+ * ubifs_pad - pad flash space.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to put padding to
+ * @pad: how many bytes to pad
+ *
+ * The flash media obliges us to write only in chunks of %c->min_io_size and
+ * when we have to write less data we add padding node to the write-buffer and
+ * pad it to the next minimal I/O unit's boundary. Padding nodes help when the
+ * media is being scanned. If the amount of wasted space is not enough to fit a
+ * padding node which takes %UBIFS_PAD_NODE_SZ bytes, we write padding bytes
+ * pattern (%UBIFS_PADDING_BYTE).
+ *
+ * Padding nodes are also used to fill gaps when the "commit-in-gaps" method is
+ * used.
+ */
+void ubifs_pad(const struct ubifs_info *c, void *buf, int pad)
+{
+ uint32_t crc;
+
+ ubifs_assert(c, pad >= 0);
+
+ if (pad >= UBIFS_PAD_NODE_SZ) {
+ struct ubifs_ch *ch = buf;
+ struct ubifs_pad_node *pad_node = buf;
+
+ ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
+ ch->node_type = UBIFS_PAD_NODE;
+ ch->group_type = UBIFS_NO_NODE_GROUP;
+ ch->padding[0] = ch->padding[1] = 0;
+ ch->sqnum = 0;
+ ch->len = cpu_to_le32(UBIFS_PAD_NODE_SZ);
+ pad -= UBIFS_PAD_NODE_SZ;
+ pad_node->pad_len = cpu_to_le32(pad);
+ crc = crc32(UBIFS_CRC32_INIT, buf + 8, UBIFS_PAD_NODE_SZ - 8);
+ ch->crc = cpu_to_le32(crc);
+ memset(buf + UBIFS_PAD_NODE_SZ, 0, pad);
+ } else if (pad > 0)
+ /* Too little space, padding node won't fit */
+ memset(buf, UBIFS_PADDING_BYTE, pad);
+}
+
+/**
+ * next_sqnum - get next sequence number.
+ * @c: UBIFS file-system description object
+ */
+static unsigned long long next_sqnum(struct ubifs_info *c)
+{
+ unsigned long long sqnum;
+
+ spin_lock(&c->cnt_lock);
+ sqnum = ++c->max_sqnum;
+ spin_unlock(&c->cnt_lock);
+
+ if (unlikely(sqnum >= SQNUM_WARN_WATERMARK)) {
+ if (sqnum >= SQNUM_WATERMARK) {
+ ubifs_err(c, "sequence number overflow %llu, end of life",
+ sqnum);
+ ubifs_ro_mode(c, -EINVAL);
+ }
+ ubifs_warn(c, "running out of sequence numbers, end of life soon");
+ }
+
+ return sqnum;
+}
+
+void ubifs_init_node(struct ubifs_info *c, void *node, int len, int pad)
+{
+ struct ubifs_ch *ch = node;
+ unsigned long long sqnum = next_sqnum(c);
+
+ ubifs_assert(c, len >= UBIFS_CH_SZ);
+
+ ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
+ ch->len = cpu_to_le32(len);
+ ch->group_type = UBIFS_NO_NODE_GROUP;
+ ch->sqnum = cpu_to_le64(sqnum);
+ ch->padding[0] = ch->padding[1] = 0;
+
+ if (pad) {
+ len = ALIGN(len, 8);
+ pad = ALIGN(len, c->min_io_size) - len;
+ ubifs_pad(c, node + len, pad);
+ }
+}
+
+void ubifs_crc_node(struct ubifs_info *c, void *node, int len)
+{
+ struct ubifs_ch *ch = node;
+ uint32_t crc;
+
+ crc = crc32(UBIFS_CRC32_INIT, node + 8, len - 8);
+ ch->crc = cpu_to_le32(crc);
+}
+
+/**
+ * ubifs_prepare_node_hmac - prepare node to be written to flash.
+ * @c: UBIFS file-system description object
+ * @node: the node to pad
+ * @len: node length
+ * @hmac_offs: offset of the HMAC in the node
+ * @pad: if the buffer has to be padded
+ *
+ * This function prepares node at @node to be written to the media - it
+ * calculates node CRC, fills the common header, and adds proper padding up to
+ * the next minimum I/O unit if @pad is not zero. if @hmac_offs is positive then
+ * a HMAC is inserted into the node at the given offset.
+ *
+ * This function returns 0 for success or a negative error code otherwise.
+ */
+int ubifs_prepare_node_hmac(struct ubifs_info *c, void *node, int len,
+ int hmac_offs, int pad)
+{
+ int err;
+
+ ubifs_init_node(c, node, len, pad);
+
+ if (hmac_offs > 0) {
+ err = ubifs_node_insert_hmac(c, node, len, hmac_offs);
+ if (err)
+ return err;
+ }
+
+ ubifs_crc_node(c, node, len);
+
+ return 0;
+}
+
+/**
+ * ubifs_prepare_node - prepare node to be written to flash.
+ * @c: UBIFS file-system description object
+ * @node: the node to pad
+ * @len: node length
+ * @pad: if the buffer has to be padded
+ *
+ * This function prepares node at @node to be written to the media - it
+ * calculates node CRC, fills the common header, and adds proper padding up to
+ * the next minimum I/O unit if @pad is not zero.
+ */
+void ubifs_prepare_node(struct ubifs_info *c, void *node, int len, int pad)
+{
+ /*
+ * Deliberately ignore return value since this function can only fail
+ * when a hmac offset is given.
+ */
+ ubifs_prepare_node_hmac(c, node, len, 0, pad);
+}
+
+/**
+ * ubifs_prep_grp_node - prepare node of a group to be written to flash.
+ * @c: UBIFS file-system description object
+ * @node: the node to pad
+ * @len: node length
+ * @last: indicates the last node of the group
+ *
+ * This function prepares node at @node to be written to the media - it
+ * calculates node CRC and fills the common header.
+ */
+void ubifs_prep_grp_node(struct ubifs_info *c, void *node, int len, int last)
+{
+ uint32_t crc;
+ struct ubifs_ch *ch = node;
+ unsigned long long sqnum = next_sqnum(c);
+
+ ubifs_assert(c, len >= UBIFS_CH_SZ);
+
+ ch->magic = cpu_to_le32(UBIFS_NODE_MAGIC);
+ ch->len = cpu_to_le32(len);
+ if (last)
+ ch->group_type = UBIFS_LAST_OF_NODE_GROUP;
+ else
+ ch->group_type = UBIFS_IN_NODE_GROUP;
+ ch->sqnum = cpu_to_le64(sqnum);
+ ch->padding[0] = ch->padding[1] = 0;
+ crc = crc32(UBIFS_CRC32_INIT, node + 8, len - 8);
+ ch->crc = cpu_to_le32(crc);
+}
+
+/**
+ * wbuf_timer_callback_nolock - write-buffer timer callback function.
+ * @timer: timer data (write-buffer descriptor)
+ *
+ * This function is called when the write-buffer timer expires.
+ */
+static enum hrtimer_restart wbuf_timer_callback_nolock(struct hrtimer *timer)
+{
+ struct ubifs_wbuf *wbuf = container_of(timer, struct ubifs_wbuf, timer);
+
+ dbg_io("jhead %s", dbg_jhead(wbuf->jhead));
+ wbuf->need_sync = 1;
+ wbuf->c->need_wbuf_sync = 1;
+ ubifs_wake_up_bgt(wbuf->c);
+ return HRTIMER_NORESTART;
+}
+
+/**
+ * new_wbuf_timer_nolock - start new write-buffer timer.
+ * @c: UBIFS file-system description object
+ * @wbuf: write-buffer descriptor
+ */
+static void new_wbuf_timer_nolock(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
+{
+ ktime_t softlimit = ms_to_ktime(dirty_writeback_interval * 10);
+ unsigned long long delta = dirty_writeback_interval;
+
+ /* centi to milli, milli to nano, then 10% */
+ delta *= 10ULL * NSEC_PER_MSEC / 10ULL;
+
+ ubifs_assert(c, !hrtimer_active(&wbuf->timer));
+ ubifs_assert(c, delta <= ULONG_MAX);
+
+ if (wbuf->no_timer)
+ return;
+ dbg_io("set timer for jhead %s, %llu-%llu millisecs",
+ dbg_jhead(wbuf->jhead),
+ div_u64(ktime_to_ns(softlimit), USEC_PER_SEC),
+ div_u64(ktime_to_ns(softlimit) + delta, USEC_PER_SEC));
+ hrtimer_start_range_ns(&wbuf->timer, softlimit, delta,
+ HRTIMER_MODE_REL);
+}
+
+/**
+ * cancel_wbuf_timer_nolock - cancel write-buffer timer.
+ * @wbuf: write-buffer descriptor
+ */
+static void cancel_wbuf_timer_nolock(struct ubifs_wbuf *wbuf)
+{
+ if (wbuf->no_timer)
+ return;
+ wbuf->need_sync = 0;
+ hrtimer_cancel(&wbuf->timer);
+}
+
+/**
+ * ubifs_wbuf_sync_nolock - synchronize write-buffer.
+ * @wbuf: write-buffer to synchronize
+ *
+ * This function synchronizes write-buffer @buf and returns zero in case of
+ * success or a negative error code in case of failure.
+ *
+ * Note, although write-buffers are of @c->max_write_size, this function does
+ * not necessarily writes all @c->max_write_size bytes to the flash. Instead,
+ * if the write-buffer is only partially filled with data, only the used part
+ * of the write-buffer (aligned on @c->min_io_size boundary) is synchronized.
+ * This way we waste less space.
+ */
+int ubifs_wbuf_sync_nolock(struct ubifs_wbuf *wbuf)
+{
+ struct ubifs_info *c = wbuf->c;
+ int err, dirt, sync_len;
+
+ cancel_wbuf_timer_nolock(wbuf);
+ if (!wbuf->used || wbuf->lnum == -1)
+ /* Write-buffer is empty or not seeked */
+ return 0;
+
+ dbg_io("LEB %d:%d, %d bytes, jhead %s",
+ wbuf->lnum, wbuf->offs, wbuf->used, dbg_jhead(wbuf->jhead));
+ ubifs_assert(c, !(wbuf->avail & 7));
+ ubifs_assert(c, wbuf->offs + wbuf->size <= c->leb_size);
+ ubifs_assert(c, wbuf->size >= c->min_io_size);
+ ubifs_assert(c, wbuf->size <= c->max_write_size);
+ ubifs_assert(c, wbuf->size % c->min_io_size == 0);
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (c->leb_size - wbuf->offs >= c->max_write_size)
+ ubifs_assert(c, !((wbuf->offs + wbuf->size) % c->max_write_size));
+
+ if (c->ro_error)
+ return -EROFS;
+
+ /*
+ * Do not write whole write buffer but write only the minimum necessary
+ * amount of min. I/O units.
+ */
+ sync_len = ALIGN(wbuf->used, c->min_io_size);
+ dirt = sync_len - wbuf->used;
+ if (dirt)
+ ubifs_pad(c, wbuf->buf + wbuf->used, dirt);
+ err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs, sync_len);
+ if (err)
+ return err;
+
+ spin_lock(&wbuf->lock);
+ wbuf->offs += sync_len;
+ /*
+ * Now @wbuf->offs is not necessarily aligned to @c->max_write_size.
+ * But our goal is to optimize writes and make sure we write in
+ * @c->max_write_size chunks and to @c->max_write_size-aligned offset.
+ * Thus, if @wbuf->offs is not aligned to @c->max_write_size now, make
+ * sure that @wbuf->offs + @wbuf->size is aligned to
+ * @c->max_write_size. This way we make sure that after next
+ * write-buffer flush we are again at the optimal offset (aligned to
+ * @c->max_write_size).
+ */
+ if (c->leb_size - wbuf->offs < c->max_write_size)
+ wbuf->size = c->leb_size - wbuf->offs;
+ else if (wbuf->offs & (c->max_write_size - 1))
+ wbuf->size = ALIGN(wbuf->offs, c->max_write_size) - wbuf->offs;
+ else
+ wbuf->size = c->max_write_size;
+ wbuf->avail = wbuf->size;
+ wbuf->used = 0;
+ wbuf->next_ino = 0;
+ spin_unlock(&wbuf->lock);
+
+ if (wbuf->sync_callback)
+ err = wbuf->sync_callback(c, wbuf->lnum,
+ c->leb_size - wbuf->offs, dirt);
+ return err;
+}
+
+/**
+ * ubifs_wbuf_seek_nolock - seek write-buffer.
+ * @wbuf: write-buffer
+ * @lnum: logical eraseblock number to seek to
+ * @offs: logical eraseblock offset to seek to
+ *
+ * This function targets the write-buffer to logical eraseblock @lnum:@offs.
+ * The write-buffer has to be empty. Returns zero in case of success and a
+ * negative error code in case of failure.
+ */
+int ubifs_wbuf_seek_nolock(struct ubifs_wbuf *wbuf, int lnum, int offs)
+{
+ const struct ubifs_info *c = wbuf->c;
+
+ dbg_io("LEB %d:%d, jhead %s", lnum, offs, dbg_jhead(wbuf->jhead));
+ ubifs_assert(c, lnum >= 0 && lnum < c->leb_cnt);
+ ubifs_assert(c, offs >= 0 && offs <= c->leb_size);
+ ubifs_assert(c, offs % c->min_io_size == 0 && !(offs & 7));
+ ubifs_assert(c, lnum != wbuf->lnum);
+ ubifs_assert(c, wbuf->used == 0);
+
+ spin_lock(&wbuf->lock);
+ wbuf->lnum = lnum;
+ wbuf->offs = offs;
+ if (c->leb_size - wbuf->offs < c->max_write_size)
+ wbuf->size = c->leb_size - wbuf->offs;
+ else if (wbuf->offs & (c->max_write_size - 1))
+ wbuf->size = ALIGN(wbuf->offs, c->max_write_size) - wbuf->offs;
+ else
+ wbuf->size = c->max_write_size;
+ wbuf->avail = wbuf->size;
+ wbuf->used = 0;
+ spin_unlock(&wbuf->lock);
+
+ return 0;
+}
+
+/**
+ * ubifs_bg_wbufs_sync - synchronize write-buffers.
+ * @c: UBIFS file-system description object
+ *
+ * This function is called by background thread to synchronize write-buffers.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_bg_wbufs_sync(struct ubifs_info *c)
+{
+ int err, i;
+
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ if (!c->need_wbuf_sync)
+ return 0;
+ c->need_wbuf_sync = 0;
+
+ if (c->ro_error) {
+ err = -EROFS;
+ goto out_timers;
+ }
+
+ dbg_io("synchronize");
+ for (i = 0; i < c->jhead_cnt; i++) {
+ struct ubifs_wbuf *wbuf = &c->jheads[i].wbuf;
+
+ cond_resched();
+
+ /*
+ * If the mutex is locked then wbuf is being changed, so
+ * synchronization is not necessary.
+ */
+ if (mutex_is_locked(&wbuf->io_mutex))
+ continue;
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ if (!wbuf->need_sync) {
+ mutex_unlock(&wbuf->io_mutex);
+ continue;
+ }
+
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ mutex_unlock(&wbuf->io_mutex);
+ if (err) {
+ ubifs_err(c, "cannot sync write-buffer, error %d", err);
+ ubifs_ro_mode(c, err);
+ goto out_timers;
+ }
+ }
+
+ return 0;
+
+out_timers:
+ /* Cancel all timers to prevent repeated errors */
+ for (i = 0; i < c->jhead_cnt; i++) {
+ struct ubifs_wbuf *wbuf = &c->jheads[i].wbuf;
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ cancel_wbuf_timer_nolock(wbuf);
+ mutex_unlock(&wbuf->io_mutex);
+ }
+ return err;
+}
+
+/**
+ * ubifs_wbuf_write_nolock - write data to flash via write-buffer.
+ * @wbuf: write-buffer
+ * @buf: node to write
+ * @len: node length
+ *
+ * This function writes data to flash via write-buffer @wbuf. This means that
+ * the last piece of the node won't reach the flash media immediately if it
+ * does not take whole max. write unit (@c->max_write_size). Instead, the node
+ * will sit in RAM until the write-buffer is synchronized (e.g., by timer, or
+ * because more data are appended to the write-buffer).
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure. If the node cannot be written because there is no more
+ * space in this logical eraseblock, %-ENOSPC is returned.
+ */
+int ubifs_wbuf_write_nolock(struct ubifs_wbuf *wbuf, void *buf, int len)
+{
+ struct ubifs_info *c = wbuf->c;
+ int err, n, written = 0, aligned_len = ALIGN(len, 8);
+
+ dbg_io("%d bytes (%s) to jhead %s wbuf at LEB %d:%d", len,
+ dbg_ntype(((struct ubifs_ch *)buf)->node_type),
+ dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs + wbuf->used);
+ ubifs_assert(c, len > 0 && wbuf->lnum >= 0 && wbuf->lnum < c->leb_cnt);
+ ubifs_assert(c, wbuf->offs >= 0 && wbuf->offs % c->min_io_size == 0);
+ ubifs_assert(c, !(wbuf->offs & 7) && wbuf->offs <= c->leb_size);
+ ubifs_assert(c, wbuf->avail > 0 && wbuf->avail <= wbuf->size);
+ ubifs_assert(c, wbuf->size >= c->min_io_size);
+ ubifs_assert(c, wbuf->size <= c->max_write_size);
+ ubifs_assert(c, wbuf->size % c->min_io_size == 0);
+ ubifs_assert(c, mutex_is_locked(&wbuf->io_mutex));
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ ubifs_assert(c, !c->space_fixup);
+ if (c->leb_size - wbuf->offs >= c->max_write_size)
+ ubifs_assert(c, !((wbuf->offs + wbuf->size) % c->max_write_size));
+
+ if (c->leb_size - wbuf->offs - wbuf->used < aligned_len) {
+ err = -ENOSPC;
+ goto out;
+ }
+
+ cancel_wbuf_timer_nolock(wbuf);
+
+ if (c->ro_error)
+ return -EROFS;
+
+ if (aligned_len <= wbuf->avail) {
+ /*
+ * The node is not very large and fits entirely within
+ * write-buffer.
+ */
+ memcpy(wbuf->buf + wbuf->used, buf, len);
+ if (aligned_len > len) {
+ ubifs_assert(c, aligned_len - len < 8);
+ ubifs_pad(c, wbuf->buf + wbuf->used + len, aligned_len - len);
+ }
+
+ if (aligned_len == wbuf->avail) {
+ dbg_io("flush jhead %s wbuf to LEB %d:%d",
+ dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
+ err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf,
+ wbuf->offs, wbuf->size);
+ if (err)
+ goto out;
+
+ spin_lock(&wbuf->lock);
+ wbuf->offs += wbuf->size;
+ if (c->leb_size - wbuf->offs >= c->max_write_size)
+ wbuf->size = c->max_write_size;
+ else
+ wbuf->size = c->leb_size - wbuf->offs;
+ wbuf->avail = wbuf->size;
+ wbuf->used = 0;
+ wbuf->next_ino = 0;
+ spin_unlock(&wbuf->lock);
+ } else {
+ spin_lock(&wbuf->lock);
+ wbuf->avail -= aligned_len;
+ wbuf->used += aligned_len;
+ spin_unlock(&wbuf->lock);
+ }
+
+ goto exit;
+ }
+
+ if (wbuf->used) {
+ /*
+ * The node is large enough and does not fit entirely within
+ * current available space. We have to fill and flush
+ * write-buffer and switch to the next max. write unit.
+ */
+ dbg_io("flush jhead %s wbuf to LEB %d:%d",
+ dbg_jhead(wbuf->jhead), wbuf->lnum, wbuf->offs);
+ memcpy(wbuf->buf + wbuf->used, buf, wbuf->avail);
+ err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs,
+ wbuf->size);
+ if (err)
+ goto out;
+
+ wbuf->offs += wbuf->size;
+ len -= wbuf->avail;
+ aligned_len -= wbuf->avail;
+ written += wbuf->avail;
+ } else if (wbuf->offs & (c->max_write_size - 1)) {
+ /*
+ * The write-buffer offset is not aligned to
+ * @c->max_write_size and @wbuf->size is less than
+ * @c->max_write_size. Write @wbuf->size bytes to make sure the
+ * following writes are done in optimal @c->max_write_size
+ * chunks.
+ */
+ dbg_io("write %d bytes to LEB %d:%d",
+ wbuf->size, wbuf->lnum, wbuf->offs);
+ err = ubifs_leb_write(c, wbuf->lnum, buf, wbuf->offs,
+ wbuf->size);
+ if (err)
+ goto out;
+
+ wbuf->offs += wbuf->size;
+ len -= wbuf->size;
+ aligned_len -= wbuf->size;
+ written += wbuf->size;
+ }
+
+ /*
+ * The remaining data may take more whole max. write units, so write the
+ * remains multiple to max. write unit size directly to the flash media.
+ * We align node length to 8-byte boundary because we anyway flash wbuf
+ * if the remaining space is less than 8 bytes.
+ */
+ n = aligned_len >> c->max_write_shift;
+ if (n) {
+ int m = n - 1;
+
+ dbg_io("write %d bytes to LEB %d:%d", n, wbuf->lnum,
+ wbuf->offs);
+
+ if (m) {
+ /* '(n-1)<<c->max_write_shift < len' is always true. */
+ m <<= c->max_write_shift;
+ err = ubifs_leb_write(c, wbuf->lnum, buf + written,
+ wbuf->offs, m);
+ if (err)
+ goto out;
+ wbuf->offs += m;
+ aligned_len -= m;
+ len -= m;
+ written += m;
+ }
+
+ /*
+ * The non-written len of buf may be less than 'n' because
+ * parameter 'len' is not 8 bytes aligned, so here we read
+ * min(len, n) bytes from buf.
+ */
+ n = 1 << c->max_write_shift;
+ memcpy(wbuf->buf, buf + written, min(len, n));
+ if (n > len) {
+ ubifs_assert(c, n - len < 8);
+ ubifs_pad(c, wbuf->buf + len, n - len);
+ }
+
+ err = ubifs_leb_write(c, wbuf->lnum, wbuf->buf, wbuf->offs, n);
+ if (err)
+ goto out;
+ wbuf->offs += n;
+ aligned_len -= n;
+ len -= min(len, n);
+ written += n;
+ }
+
+ spin_lock(&wbuf->lock);
+ if (aligned_len) {
+ /*
+ * And now we have what's left and what does not take whole
+ * max. write unit, so write it to the write-buffer and we are
+ * done.
+ */
+ memcpy(wbuf->buf, buf + written, len);
+ if (aligned_len > len) {
+ ubifs_assert(c, aligned_len - len < 8);
+ ubifs_pad(c, wbuf->buf + len, aligned_len - len);
+ }
+ }
+
+ if (c->leb_size - wbuf->offs >= c->max_write_size)
+ wbuf->size = c->max_write_size;
+ else
+ wbuf->size = c->leb_size - wbuf->offs;
+ wbuf->avail = wbuf->size - aligned_len;
+ wbuf->used = aligned_len;
+ wbuf->next_ino = 0;
+ spin_unlock(&wbuf->lock);
+
+exit:
+ if (wbuf->sync_callback) {
+ int free = c->leb_size - wbuf->offs - wbuf->used;
+
+ err = wbuf->sync_callback(c, wbuf->lnum, free, 0);
+ if (err)
+ goto out;
+ }
+
+ if (wbuf->used)
+ new_wbuf_timer_nolock(c, wbuf);
+
+ return 0;
+
+out:
+ ubifs_err(c, "cannot write %d bytes to LEB %d:%d, error %d",
+ len, wbuf->lnum, wbuf->offs, err);
+ ubifs_dump_node(c, buf, written + len);
+ dump_stack();
+ ubifs_dump_leb(c, wbuf->lnum);
+ return err;
+}
+
+/**
+ * ubifs_write_node_hmac - write node to the media.
+ * @c: UBIFS file-system description object
+ * @buf: the node to write
+ * @len: node length
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ * @hmac_offs: offset of the HMAC within the node
+ *
+ * This function automatically fills node magic number, assigns sequence
+ * number, and calculates node CRC checksum. The length of the @buf buffer has
+ * to be aligned to the minimal I/O unit size. This function automatically
+ * appends padding node and padding bytes if needed. Returns zero in case of
+ * success and a negative error code in case of failure.
+ */
+int ubifs_write_node_hmac(struct ubifs_info *c, void *buf, int len, int lnum,
+ int offs, int hmac_offs)
+{
+ int err, buf_len = ALIGN(len, c->min_io_size);
+
+ dbg_io("LEB %d:%d, %s, length %d (aligned %d)",
+ lnum, offs, dbg_ntype(((struct ubifs_ch *)buf)->node_type), len,
+ buf_len);
+ ubifs_assert(c, lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(c, offs % c->min_io_size == 0 && offs < c->leb_size);
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ ubifs_assert(c, !c->space_fixup);
+
+ if (c->ro_error)
+ return -EROFS;
+
+ err = ubifs_prepare_node_hmac(c, buf, len, hmac_offs, 1);
+ if (err)
+ return err;
+
+ err = ubifs_leb_write(c, lnum, buf, offs, buf_len);
+ if (err)
+ ubifs_dump_node(c, buf, len);
+
+ return err;
+}
+
+/**
+ * ubifs_write_node - write node to the media.
+ * @c: UBIFS file-system description object
+ * @buf: the node to write
+ * @len: node length
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ *
+ * This function automatically fills node magic number, assigns sequence
+ * number, and calculates node CRC checksum. The length of the @buf buffer has
+ * to be aligned to the minimal I/O unit size. This function automatically
+ * appends padding node and padding bytes if needed. Returns zero in case of
+ * success and a negative error code in case of failure.
+ */
+int ubifs_write_node(struct ubifs_info *c, void *buf, int len, int lnum,
+ int offs)
+{
+ return ubifs_write_node_hmac(c, buf, len, lnum, offs, -1);
+}
+
+/**
+ * ubifs_read_node_wbuf - read node from the media or write-buffer.
+ * @wbuf: wbuf to check for un-written data
+ * @buf: buffer to read to
+ * @type: node type
+ * @len: node length
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ *
+ * This function reads a node of known type and length, checks it and stores
+ * in @buf. If the node partially or fully sits in the write-buffer, this
+ * function takes data from the buffer, otherwise it reads the flash media.
+ * Returns zero in case of success, %-EUCLEAN if CRC mismatched and a negative
+ * error code in case of failure.
+ */
+int ubifs_read_node_wbuf(struct ubifs_wbuf *wbuf, void *buf, int type, int len,
+ int lnum, int offs)
+{
+ const struct ubifs_info *c = wbuf->c;
+ int err, rlen, overlap;
+ struct ubifs_ch *ch = buf;
+
+ dbg_io("LEB %d:%d, %s, length %d, jhead %s", lnum, offs,
+ dbg_ntype(type), len, dbg_jhead(wbuf->jhead));
+ ubifs_assert(c, wbuf && lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(c, !(offs & 7) && offs < c->leb_size);
+ ubifs_assert(c, type >= 0 && type < UBIFS_NODE_TYPES_CNT);
+
+ spin_lock(&wbuf->lock);
+ overlap = (lnum == wbuf->lnum && offs + len > wbuf->offs);
+ if (!overlap) {
+ /* We may safely unlock the write-buffer and read the data */
+ spin_unlock(&wbuf->lock);
+ return ubifs_read_node(c, buf, type, len, lnum, offs);
+ }
+
+ /* Don't read under wbuf */
+ rlen = wbuf->offs - offs;
+ if (rlen < 0)
+ rlen = 0;
+
+ /* Copy the rest from the write-buffer */
+ memcpy(buf + rlen, wbuf->buf + offs + rlen - wbuf->offs, len - rlen);
+ spin_unlock(&wbuf->lock);
+
+ if (rlen > 0) {
+ /* Read everything that goes before write-buffer */
+ err = ubifs_leb_read(c, lnum, buf, offs, rlen, 0);
+ if (err && err != -EBADMSG)
+ return err;
+ }
+
+ if (type != ch->node_type) {
+ ubifs_err(c, "bad node type (%d but expected %d)",
+ ch->node_type, type);
+ goto out;
+ }
+
+ err = ubifs_check_node(c, buf, len, lnum, offs, 0, 0);
+ if (err) {
+ ubifs_err(c, "expected node type %d", type);
+ return err;
+ }
+
+ rlen = le32_to_cpu(ch->len);
+ if (rlen != len) {
+ ubifs_err(c, "bad node length %d, expected %d", rlen, len);
+ goto out;
+ }
+
+ return 0;
+
+out:
+ ubifs_err(c, "bad node at LEB %d:%d", lnum, offs);
+ ubifs_dump_node(c, buf, len);
+ dump_stack();
+ return -EINVAL;
+}
+
+/**
+ * ubifs_read_node - read node.
+ * @c: UBIFS file-system description object
+ * @buf: buffer to read to
+ * @type: node type
+ * @len: node length (not aligned)
+ * @lnum: logical eraseblock number
+ * @offs: offset within the logical eraseblock
+ *
+ * This function reads a node of known type and length, checks it and
+ * stores in @buf. Returns zero in case of success, %-EUCLEAN if CRC mismatched
+ * and a negative error code in case of failure.
+ */
+int ubifs_read_node(const struct ubifs_info *c, void *buf, int type, int len,
+ int lnum, int offs)
+{
+ int err, l;
+ struct ubifs_ch *ch = buf;
+
+ dbg_io("LEB %d:%d, %s, length %d", lnum, offs, dbg_ntype(type), len);
+ ubifs_assert(c, lnum >= 0 && lnum < c->leb_cnt && offs >= 0);
+ ubifs_assert(c, len >= UBIFS_CH_SZ && offs + len <= c->leb_size);
+ ubifs_assert(c, !(offs & 7) && offs < c->leb_size);
+ ubifs_assert(c, type >= 0 && type < UBIFS_NODE_TYPES_CNT);
+
+ err = ubifs_leb_read(c, lnum, buf, offs, len, 0);
+ if (err && err != -EBADMSG)
+ return err;
+
+ if (type != ch->node_type) {
+ ubifs_errc(c, "bad node type (%d but expected %d)",
+ ch->node_type, type);
+ goto out;
+ }
+
+ err = ubifs_check_node(c, buf, len, lnum, offs, 0, 0);
+ if (err) {
+ ubifs_errc(c, "expected node type %d", type);
+ return err;
+ }
+
+ l = le32_to_cpu(ch->len);
+ if (l != len) {
+ ubifs_errc(c, "bad node length %d, expected %d", l, len);
+ goto out;
+ }
+
+ return 0;
+
+out:
+ ubifs_errc(c, "bad node at LEB %d:%d, LEB mapping status %d", lnum,
+ offs, ubi_is_mapped(c->ubi, lnum));
+ if (!c->probing) {
+ ubifs_dump_node(c, buf, len);
+ dump_stack();
+ }
+ return -EINVAL;
+}
+
+/**
+ * ubifs_wbuf_init - initialize write-buffer.
+ * @c: UBIFS file-system description object
+ * @wbuf: write-buffer to initialize
+ *
+ * This function initializes write-buffer. Returns zero in case of success
+ * %-ENOMEM in case of failure.
+ */
+int ubifs_wbuf_init(struct ubifs_info *c, struct ubifs_wbuf *wbuf)
+{
+ size_t size;
+
+ wbuf->buf = kmalloc(c->max_write_size, GFP_KERNEL);
+ if (!wbuf->buf)
+ return -ENOMEM;
+
+ size = (c->max_write_size / UBIFS_CH_SZ + 1) * sizeof(ino_t);
+ wbuf->inodes = kmalloc(size, GFP_KERNEL);
+ if (!wbuf->inodes) {
+ kfree(wbuf->buf);
+ wbuf->buf = NULL;
+ return -ENOMEM;
+ }
+
+ wbuf->used = 0;
+ wbuf->lnum = wbuf->offs = -1;
+ /*
+ * If the LEB starts at the max. write size aligned address, then
+ * write-buffer size has to be set to @c->max_write_size. Otherwise,
+ * set it to something smaller so that it ends at the closest max.
+ * write size boundary.
+ */
+ size = c->max_write_size - (c->leb_start % c->max_write_size);
+ wbuf->avail = wbuf->size = size;
+ wbuf->sync_callback = NULL;
+ mutex_init(&wbuf->io_mutex);
+ spin_lock_init(&wbuf->lock);
+ wbuf->c = c;
+ wbuf->next_ino = 0;
+
+ hrtimer_init(&wbuf->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+ wbuf->timer.function = wbuf_timer_callback_nolock;
+ return 0;
+}
+
+/**
+ * ubifs_wbuf_add_ino_nolock - add an inode number into the wbuf inode array.
+ * @wbuf: the write-buffer where to add
+ * @inum: the inode number
+ *
+ * This function adds an inode number to the inode array of the write-buffer.
+ */
+void ubifs_wbuf_add_ino_nolock(struct ubifs_wbuf *wbuf, ino_t inum)
+{
+ if (!wbuf->buf)
+ /* NOR flash or something similar */
+ return;
+
+ spin_lock(&wbuf->lock);
+ if (wbuf->used)
+ wbuf->inodes[wbuf->next_ino++] = inum;
+ spin_unlock(&wbuf->lock);
+}
+
+/**
+ * wbuf_has_ino - returns if the wbuf contains data from the inode.
+ * @wbuf: the write-buffer
+ * @inum: the inode number
+ *
+ * This function returns with %1 if the write-buffer contains some data from the
+ * given inode otherwise it returns with %0.
+ */
+static int wbuf_has_ino(struct ubifs_wbuf *wbuf, ino_t inum)
+{
+ int i, ret = 0;
+
+ spin_lock(&wbuf->lock);
+ for (i = 0; i < wbuf->next_ino; i++)
+ if (inum == wbuf->inodes[i]) {
+ ret = 1;
+ break;
+ }
+ spin_unlock(&wbuf->lock);
+
+ return ret;
+}
+
+/**
+ * ubifs_sync_wbufs_by_inode - synchronize write-buffers for an inode.
+ * @c: UBIFS file-system description object
+ * @inode: inode to synchronize
+ *
+ * This function synchronizes write-buffers which contain nodes belonging to
+ * @inode. Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_sync_wbufs_by_inode(struct ubifs_info *c, struct inode *inode)
+{
+ int i, err = 0;
+
+ for (i = 0; i < c->jhead_cnt; i++) {
+ struct ubifs_wbuf *wbuf = &c->jheads[i].wbuf;
+
+ if (i == GCHD)
+ /*
+ * GC head is special, do not look at it. Even if the
+ * head contains something related to this inode, it is
+ * a _copy_ of corresponding on-flash node which sits
+ * somewhere else.
+ */
+ continue;
+
+ if (!wbuf_has_ino(wbuf, inode->i_ino))
+ continue;
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ if (wbuf_has_ino(wbuf, inode->i_ino))
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ mutex_unlock(&wbuf->io_mutex);
+
+ if (err) {
+ ubifs_ro_mode(c, err);
+ return err;
+ }
+ }
+ return 0;
+}
new file mode 100644
@@ -0,0 +1,1928 @@
+// 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 journal.
+ *
+ * The journal consists of 2 parts - the log and bud LEBs. The log has fixed
+ * length and position, while a bud logical eraseblock is any LEB in the main
+ * area. Buds contain file system data - data nodes, inode nodes, etc. The log
+ * contains only references to buds and some other stuff like commit
+ * start node. The idea is that when we commit the journal, we do
+ * not copy the data, the buds just become indexed. Since after the commit the
+ * nodes in bud eraseblocks become leaf nodes of the file system index tree, we
+ * use term "bud". Analogy is obvious, bud eraseblocks contain nodes which will
+ * become leafs in the future.
+ *
+ * The journal is multi-headed because we want to write data to the journal as
+ * optimally as possible. It is nice to have nodes belonging to the same inode
+ * in one LEB, so we may write data owned by different inodes to different
+ * journal heads, although at present only one data head is used.
+ *
+ * For recovery reasons, the base head contains all inode nodes, all directory
+ * entry nodes and all truncate nodes. This means that the other heads contain
+ * only data nodes.
+ *
+ * Bud LEBs may be half-indexed. For example, if the bud was not full at the
+ * time of commit, the bud is retained to continue to be used in the journal,
+ * even though the "front" of the LEB is now indexed. In that case, the log
+ * reference contains the offset where the bud starts for the purposes of the
+ * journal.
+ *
+ * The journal size has to be limited, because the larger is the journal, the
+ * longer it takes to mount UBIFS (scanning the journal) and the more memory it
+ * takes (indexing in the TNC).
+ *
+ * All the journal write operations like 'ubifs_jnl_update()' here, which write
+ * multiple UBIFS nodes to the journal at one go, are atomic with respect to
+ * unclean reboots. Should the unclean reboot happen, the recovery code drops
+ * all the nodes.
+ */
+
+#include "ubifs.h"
+
+/**
+ * zero_ino_node_unused - zero out unused fields of an on-flash inode node.
+ * @ino: the inode to zero out
+ */
+static inline void zero_ino_node_unused(struct ubifs_ino_node *ino)
+{
+ memset(ino->padding1, 0, 4);
+ memset(ino->padding2, 0, 26);
+}
+
+/**
+ * zero_dent_node_unused - zero out unused fields of an on-flash directory
+ * entry node.
+ * @dent: the directory entry to zero out
+ */
+static inline void zero_dent_node_unused(struct ubifs_dent_node *dent)
+{
+ dent->padding1 = 0;
+}
+
+/**
+ * zero_trun_node_unused - zero out unused fields of an on-flash truncation
+ * node.
+ * @trun: the truncation node to zero out
+ */
+static inline void zero_trun_node_unused(struct ubifs_trun_node *trun)
+{
+ memset(trun->padding, 0, 12);
+}
+
+static void ubifs_add_auth_dirt(struct ubifs_info *c, int lnum)
+{
+ if (ubifs_authenticated(c))
+ ubifs_add_dirt(c, lnum, ubifs_auth_node_sz(c));
+}
+
+/**
+ * reserve_space - reserve space in the journal.
+ * @c: UBIFS file-system description object
+ * @jhead: journal head number
+ * @len: node length
+ *
+ * This function reserves space in journal head @head. If the reservation
+ * succeeded, the journal head stays locked and later has to be unlocked using
+ * 'release_head()'. Returns zero in case of success, %-EAGAIN if commit has to
+ * be done, and other negative error codes in case of other failures.
+ */
+static int reserve_space(struct ubifs_info *c, int jhead, int len)
+{
+ int err = 0, err1, retries = 0, avail, lnum, offs, squeeze;
+ struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
+
+ /*
+ * Typically, the base head has smaller nodes written to it, so it is
+ * better to try to allocate space at the ends of eraseblocks. This is
+ * what the squeeze parameter does.
+ */
+ ubifs_assert(c, !c->ro_media && !c->ro_mount);
+ squeeze = (jhead == BASEHD);
+again:
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+
+ if (c->ro_error) {
+ err = -EROFS;
+ goto out_unlock;
+ }
+
+ avail = c->leb_size - wbuf->offs - wbuf->used;
+ if (wbuf->lnum != -1 && avail >= len)
+ return 0;
+
+ /*
+ * Write buffer wasn't seek'ed or there is no enough space - look for an
+ * LEB with some empty space.
+ */
+ lnum = ubifs_find_free_space(c, len, &offs, squeeze);
+ if (lnum >= 0)
+ goto out;
+
+ err = lnum;
+ if (err != -ENOSPC)
+ goto out_unlock;
+
+ /*
+ * No free space, we have to run garbage collector to make
+ * some. But the write-buffer mutex has to be unlocked because
+ * GC also takes it.
+ */
+ dbg_jnl("no free space in jhead %s, run GC", dbg_jhead(jhead));
+ mutex_unlock(&wbuf->io_mutex);
+
+ lnum = ubifs_garbage_collect(c, 0);
+ if (lnum < 0) {
+ err = lnum;
+ if (err != -ENOSPC)
+ return err;
+
+ /*
+ * GC could not make a free LEB. But someone else may
+ * have allocated new bud for this journal head,
+ * because we dropped @wbuf->io_mutex, so try once
+ * again.
+ */
+ dbg_jnl("GC couldn't make a free LEB for jhead %s",
+ dbg_jhead(jhead));
+ if (retries++ < 2) {
+ dbg_jnl("retry (%d)", retries);
+ goto again;
+ }
+
+ dbg_jnl("return -ENOSPC");
+ return err;
+ }
+
+ mutex_lock_nested(&wbuf->io_mutex, wbuf->jhead);
+ dbg_jnl("got LEB %d for jhead %s", lnum, dbg_jhead(jhead));
+ avail = c->leb_size - wbuf->offs - wbuf->used;
+
+ if (wbuf->lnum != -1 && avail >= len) {
+ /*
+ * Someone else has switched the journal head and we have
+ * enough space now. This happens when more than one process is
+ * trying to write to the same journal head at the same time.
+ */
+ dbg_jnl("return LEB %d back, already have LEB %d:%d",
+ lnum, wbuf->lnum, wbuf->offs + wbuf->used);
+ err = ubifs_return_leb(c, lnum);
+ if (err)
+ goto out_unlock;
+ return 0;
+ }
+
+ offs = 0;
+
+out:
+ /*
+ * Make sure we synchronize the write-buffer before we add the new bud
+ * to the log. Otherwise we may have a power cut after the log
+ * reference node for the last bud (@lnum) is written but before the
+ * write-buffer data are written to the next-to-last bud
+ * (@wbuf->lnum). And the effect would be that the recovery would see
+ * that there is corruption in the next-to-last bud.
+ */
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ if (err)
+ goto out_return;
+ err = ubifs_add_bud_to_log(c, jhead, lnum, offs);
+ if (err)
+ goto out_return;
+ err = ubifs_wbuf_seek_nolock(wbuf, lnum, offs);
+ if (err)
+ goto out_unlock;
+
+ return 0;
+
+out_unlock:
+ mutex_unlock(&wbuf->io_mutex);
+ return err;
+
+out_return:
+ /* An error occurred and the LEB has to be returned to lprops */
+ ubifs_assert(c, err < 0);
+ err1 = ubifs_return_leb(c, lnum);
+ if (err1 && err == -EAGAIN)
+ /*
+ * Return original error code only if it is not %-EAGAIN,
+ * which is not really an error. Otherwise, return the error
+ * code of 'ubifs_return_leb()'.
+ */
+ err = err1;
+ mutex_unlock(&wbuf->io_mutex);
+ return err;
+}
+
+static int ubifs_hash_nodes(struct ubifs_info *c, void *node,
+ int len, struct shash_desc *hash)
+{
+ int auth_node_size = ubifs_auth_node_sz(c);
+ int err;
+
+ while (1) {
+ const struct ubifs_ch *ch = node;
+ int nodelen = le32_to_cpu(ch->len);
+
+ ubifs_assert(c, len >= auth_node_size);
+
+ if (len == auth_node_size)
+ break;
+
+ ubifs_assert(c, len > nodelen);
+ ubifs_assert(c, ch->magic == cpu_to_le32(UBIFS_NODE_MAGIC));
+
+ err = ubifs_shash_update(c, hash, (void *)node, nodelen);
+ if (err)
+ return err;
+
+ node += ALIGN(nodelen, 8);
+ len -= ALIGN(nodelen, 8);
+ }
+
+ return ubifs_prepare_auth_node(c, node, hash);
+}
+
+/**
+ * write_head - write data to a journal head.
+ * @c: UBIFS file-system description object
+ * @jhead: journal head
+ * @buf: buffer to write
+ * @len: length to write
+ * @lnum: LEB number written is returned here
+ * @offs: offset written is returned here
+ * @sync: non-zero if the write-buffer has to by synchronized
+ *
+ * This function writes data to the reserved space of journal head @jhead.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+static int write_head(struct ubifs_info *c, int jhead, void *buf, int len,
+ int *lnum, int *offs, int sync)
+{
+ int err;
+ struct ubifs_wbuf *wbuf = &c->jheads[jhead].wbuf;
+
+ ubifs_assert(c, jhead != GCHD);
+
+ *lnum = c->jheads[jhead].wbuf.lnum;
+ *offs = c->jheads[jhead].wbuf.offs + c->jheads[jhead].wbuf.used;
+ dbg_jnl("jhead %s, LEB %d:%d, len %d",
+ dbg_jhead(jhead), *lnum, *offs, len);
+
+ if (ubifs_authenticated(c)) {
+ err = ubifs_hash_nodes(c, buf, len, c->jheads[jhead].log_hash);
+ if (err)
+ return err;
+ }
+
+ err = ubifs_wbuf_write_nolock(wbuf, buf, len);
+ if (err)
+ return err;
+ if (sync)
+ err = ubifs_wbuf_sync_nolock(wbuf);
+ return err;
+}
+
+/**
+ * make_reservation - reserve journal space.
+ * @c: UBIFS file-system description object
+ * @jhead: journal head
+ * @len: how many bytes to reserve
+ *
+ * This function makes space reservation in journal head @jhead. The function
+ * takes the commit lock and locks the journal head, and the caller has to
+ * unlock the head and finish the reservation with 'finish_reservation()'.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ *
+ * Note, the journal head may be unlocked as soon as the data is written, while
+ * the commit lock has to be released after the data has been added to the
+ * TNC.
+ */
+static int make_reservation(struct ubifs_info *c, int jhead, int len)
+{
+ int err, cmt_retries = 0, nospc_retries = 0;
+
+again:
+ down_read(&c->commit_sem);
+ err = reserve_space(c, jhead, len);
+ if (!err)
+ /* c->commit_sem will get released via finish_reservation(). */
+ return 0;
+ up_read(&c->commit_sem);
+
+ if (err == -ENOSPC) {
+ /*
+ * GC could not make any progress. We should try to commit
+ * once because it could make some dirty space and GC would
+ * make progress, so make the error -EAGAIN so that the below
+ * will commit and re-try.
+ */
+ if (nospc_retries++ < 2) {
+ dbg_jnl("no space, retry");
+ err = -EAGAIN;
+ }
+
+ /*
+ * This means that the budgeting is incorrect. We always have
+ * to be able to write to the media, because all operations are
+ * budgeted. Deletions are not budgeted, though, but we reserve
+ * an extra LEB for them.
+ */
+ }
+
+ if (err != -EAGAIN)
+ goto out;
+
+ /*
+ * -EAGAIN means that the journal is full or too large, or the above
+ * code wants to do one commit. Do this and re-try.
+ */
+ if (cmt_retries > 128) {
+ /*
+ * This should not happen unless the journal size limitations
+ * are too tough.
+ */
+ ubifs_err(c, "stuck in space allocation");
+ err = -ENOSPC;
+ goto out;
+ } else if (cmt_retries > 32)
+ ubifs_warn(c, "too many space allocation re-tries (%d)",
+ cmt_retries);
+
+ dbg_jnl("-EAGAIN, commit and retry (retried %d times)",
+ cmt_retries);
+ cmt_retries += 1;
+
+ err = ubifs_run_commit(c);
+ if (err)
+ return err;
+ goto again;
+
+out:
+ ubifs_err(c, "cannot reserve %d bytes in jhead %d, error %d",
+ len, jhead, err);
+ if (err == -ENOSPC) {
+ /* This are some budgeting problems, print useful information */
+ down_write(&c->commit_sem);
+ dump_stack();
+ ubifs_dump_budg(c, &c->bi);
+ ubifs_dump_lprops(c);
+ cmt_retries = dbg_check_lprops(c);
+ up_write(&c->commit_sem);
+ }
+ return err;
+}
+
+/**
+ * release_head - release a journal head.
+ * @c: UBIFS file-system description object
+ * @jhead: journal head
+ *
+ * This function releases journal head @jhead which was locked by
+ * the 'make_reservation()' function. It has to be called after each successful
+ * 'make_reservation()' invocation.
+ */
+static inline void release_head(struct ubifs_info *c, int jhead)
+{
+ mutex_unlock(&c->jheads[jhead].wbuf.io_mutex);
+}
+
+/**
+ * finish_reservation - finish a reservation.
+ * @c: UBIFS file-system description object
+ *
+ * This function finishes journal space reservation. It must be called after
+ * 'make_reservation()'.
+ */
+static void finish_reservation(struct ubifs_info *c)
+{
+ up_read(&c->commit_sem);
+}
+
+/**
+ * get_dent_type - translate VFS inode mode to UBIFS directory entry type.
+ * @mode: inode mode
+ */
+static int get_dent_type(int mode)
+{
+ switch (mode & S_IFMT) {
+ case S_IFREG:
+ return UBIFS_ITYPE_REG;
+ case S_IFDIR:
+ return UBIFS_ITYPE_DIR;
+ case S_IFLNK:
+ return UBIFS_ITYPE_LNK;
+ case S_IFBLK:
+ return UBIFS_ITYPE_BLK;
+ case S_IFCHR:
+ return UBIFS_ITYPE_CHR;
+ case S_IFIFO:
+ return UBIFS_ITYPE_FIFO;
+ case S_IFSOCK:
+ return UBIFS_ITYPE_SOCK;
+ default:
+ BUG();
+ }
+ return 0;
+}
+
+/**
+ * pack_inode - pack an inode node.
+ * @c: UBIFS file-system description object
+ * @ino: buffer in which to pack inode node
+ * @inode: inode to pack
+ * @last: indicates the last node of the group
+ */
+static void pack_inode(struct ubifs_info *c, struct ubifs_ino_node *ino,
+ const struct inode *inode, int last)
+{
+ int data_len = 0, last_reference = !inode->i_nlink;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ ino->ch.node_type = UBIFS_INO_NODE;
+ ino_key_init_flash(c, &ino->key, inode->i_ino);
+ ino->creat_sqnum = cpu_to_le64(ui->creat_sqnum);
+ ino->atime_sec = cpu_to_le64(inode_get_atime_sec(inode));
+ ino->atime_nsec = cpu_to_le32(inode_get_atime_nsec(inode));
+ ino->ctime_sec = cpu_to_le64(inode_get_ctime_sec(inode));
+ ino->ctime_nsec = cpu_to_le32(inode_get_ctime_nsec(inode));
+ ino->mtime_sec = cpu_to_le64(inode_get_mtime_sec(inode));
+ ino->mtime_nsec = cpu_to_le32(inode_get_mtime_nsec(inode));
+ ino->uid = cpu_to_le32(i_uid_read(inode));
+ ino->gid = cpu_to_le32(i_gid_read(inode));
+ ino->mode = cpu_to_le32(inode->i_mode);
+ ino->flags = cpu_to_le32(ui->flags);
+ ino->size = cpu_to_le64(ui->ui_size);
+ ino->nlink = cpu_to_le32(inode->i_nlink);
+ ino->compr_type = cpu_to_le16(ui->compr_type);
+ ino->data_len = cpu_to_le32(ui->data_len);
+ ino->xattr_cnt = cpu_to_le32(ui->xattr_cnt);
+ ino->xattr_size = cpu_to_le32(ui->xattr_size);
+ ino->xattr_names = cpu_to_le32(ui->xattr_names);
+ zero_ino_node_unused(ino);
+
+ /*
+ * Drop the attached data if this is a deletion inode, the data is not
+ * needed anymore.
+ */
+ if (!last_reference) {
+ memcpy(ino->data, ui->data, ui->data_len);
+ data_len = ui->data_len;
+ }
+
+ ubifs_prep_grp_node(c, ino, UBIFS_INO_NODE_SZ + data_len, last);
+}
+
+/**
+ * mark_inode_clean - mark UBIFS inode as clean.
+ * @c: UBIFS file-system description object
+ * @ui: UBIFS inode to mark as clean
+ *
+ * This helper function marks UBIFS inode @ui as clean by cleaning the
+ * @ui->dirty flag and releasing its budget. Note, VFS may still treat the
+ * inode as dirty and try to write it back, but 'ubifs_write_inode()' would
+ * just do nothing.
+ */
+static void mark_inode_clean(struct ubifs_info *c, struct ubifs_inode *ui)
+{
+ if (ui->dirty)
+ ubifs_release_dirty_inode_budget(c, ui);
+ ui->dirty = 0;
+}
+
+static void set_dent_cookie(struct ubifs_info *c, struct ubifs_dent_node *dent)
+{
+ if (c->double_hash)
+ dent->cookie = (__force __le32) get_random_u32();
+ else
+ dent->cookie = 0;
+}
+
+/**
+ * ubifs_jnl_update - update inode.
+ * @c: UBIFS file-system description object
+ * @dir: parent inode or host inode in case of extended attributes
+ * @nm: directory entry name
+ * @inode: inode to update
+ * @deletion: indicates a directory entry deletion i.e unlink or rmdir
+ * @xent: non-zero if the directory entry is an extended attribute entry
+ * @in_orphan: indicates whether the @inode is in orphan list
+ *
+ * This function updates an inode by writing a directory entry (or extended
+ * attribute entry), the inode itself, and the parent directory inode (or the
+ * host inode) to the journal.
+ *
+ * The function writes the host inode @dir last, which is important in case of
+ * extended attributes. Indeed, then we guarantee that if the host inode gets
+ * synchronized (with 'fsync()'), and the write-buffer it sits in gets flushed,
+ * the extended attribute inode gets flushed too. And this is exactly what the
+ * user expects - synchronizing the host inode synchronizes its extended
+ * attributes. Similarly, this guarantees that if @dir is synchronized, its
+ * directory entry corresponding to @nm gets synchronized too.
+ *
+ * If the inode (@inode) or the parent directory (@dir) are synchronous, this
+ * function synchronizes the write-buffer.
+ *
+ * This function marks the @dir and @inode inodes as clean and returns zero on
+ * success. In case of failure, a negative error code is returned.
+ */
+int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
+ const struct fscrypt_name *nm, const struct inode *inode,
+ int deletion, int xent, int in_orphan)
+{
+ int err, dlen, ilen, len, lnum, ino_offs, dent_offs, orphan_added = 0;
+ int aligned_dlen, aligned_ilen, sync = IS_DIRSYNC(dir);
+ int last_reference = !!(deletion && inode->i_nlink == 0);
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ struct ubifs_inode *host_ui = ubifs_inode(dir);
+ struct ubifs_dent_node *dent;
+ struct ubifs_ino_node *ino;
+ union ubifs_key dent_key, ino_key;
+ u8 hash_dent[UBIFS_HASH_ARR_SZ];
+ u8 hash_ino[UBIFS_HASH_ARR_SZ];
+ u8 hash_ino_host[UBIFS_HASH_ARR_SZ];
+
+ ubifs_assert(c, mutex_is_locked(&host_ui->ui_mutex));
+
+ dlen = UBIFS_DENT_NODE_SZ + fname_len(nm) + 1;
+ ilen = UBIFS_INO_NODE_SZ;
+
+ /*
+ * If the last reference to the inode is being deleted, then there is
+ * no need to attach and write inode data, it is being deleted anyway.
+ * And if the inode is being deleted, no need to synchronize
+ * write-buffer even if the inode is synchronous.
+ */
+ if (!last_reference) {
+ ilen += ui->data_len;
+ sync |= IS_SYNC(inode);
+ }
+
+ aligned_dlen = ALIGN(dlen, 8);
+ aligned_ilen = ALIGN(ilen, 8);
+
+ len = aligned_dlen + aligned_ilen + UBIFS_INO_NODE_SZ;
+ /* Make sure to also account for extended attributes */
+ if (ubifs_authenticated(c))
+ len += ALIGN(host_ui->data_len, 8) + ubifs_auth_node_sz(c);
+ else
+ len += host_ui->data_len;
+
+ dent = kzalloc(len, GFP_NOFS);
+ if (!dent)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, len);
+ if (err)
+ goto out_free;
+
+ if (!xent) {
+ dent->ch.node_type = UBIFS_DENT_NODE;
+ if (fname_name(nm) == NULL)
+ dent_key_init_hash(c, &dent_key, dir->i_ino, nm->hash);
+ else
+ dent_key_init(c, &dent_key, dir->i_ino, nm);
+ } else {
+ dent->ch.node_type = UBIFS_XENT_NODE;
+ xent_key_init(c, &dent_key, dir->i_ino, nm);
+ }
+
+ key_write(c, &dent_key, dent->key);
+ dent->inum = deletion ? 0 : cpu_to_le64(inode->i_ino);
+ dent->type = get_dent_type(inode->i_mode);
+ dent->nlen = cpu_to_le16(fname_len(nm));
+ memcpy(dent->name, fname_name(nm), fname_len(nm));
+ dent->name[fname_len(nm)] = '\0';
+ set_dent_cookie(c, dent);
+
+ zero_dent_node_unused(dent);
+ ubifs_prep_grp_node(c, dent, dlen, 0);
+ err = ubifs_node_calc_hash(c, dent, hash_dent);
+ if (err)
+ goto out_release;
+
+ ino = (void *)dent + aligned_dlen;
+ pack_inode(c, ino, inode, 0);
+ err = ubifs_node_calc_hash(c, ino, hash_ino);
+ if (err)
+ goto out_release;
+
+ ino = (void *)ino + aligned_ilen;
+ pack_inode(c, ino, dir, 1);
+ err = ubifs_node_calc_hash(c, ino, hash_ino_host);
+ if (err)
+ goto out_release;
+
+ if (last_reference && !in_orphan) {
+ err = ubifs_add_orphan(c, inode->i_ino);
+ if (err) {
+ release_head(c, BASEHD);
+ goto out_finish;
+ }
+ ui->del_cmtno = c->cmt_no;
+ orphan_added = 1;
+ }
+
+ err = write_head(c, BASEHD, dent, len, &lnum, &dent_offs, sync);
+ if (err)
+ goto out_release;
+ if (!sync) {
+ struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
+
+ ubifs_wbuf_add_ino_nolock(wbuf, inode->i_ino);
+ ubifs_wbuf_add_ino_nolock(wbuf, dir->i_ino);
+ }
+ release_head(c, BASEHD);
+ kfree(dent);
+ ubifs_add_auth_dirt(c, lnum);
+
+ if (deletion) {
+ if (fname_name(nm) == NULL)
+ err = ubifs_tnc_remove_dh(c, &dent_key, nm->minor_hash);
+ else
+ err = ubifs_tnc_remove_nm(c, &dent_key, nm);
+ if (err)
+ goto out_ro;
+ err = ubifs_add_dirt(c, lnum, dlen);
+ } else
+ err = ubifs_tnc_add_nm(c, &dent_key, lnum, dent_offs, dlen,
+ hash_dent, nm);
+ if (err)
+ goto out_ro;
+
+ /*
+ * Note, we do not remove the inode from TNC even if the last reference
+ * to it has just been deleted, because the inode may still be opened.
+ * Instead, the inode has been added to orphan lists and the orphan
+ * subsystem will take further care about it.
+ */
+ ino_key_init(c, &ino_key, inode->i_ino);
+ ino_offs = dent_offs + aligned_dlen;
+ err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs, ilen, hash_ino);
+ if (err)
+ goto out_ro;
+
+ ino_key_init(c, &ino_key, dir->i_ino);
+ ino_offs += aligned_ilen;
+ err = ubifs_tnc_add(c, &ino_key, lnum, ino_offs,
+ UBIFS_INO_NODE_SZ + host_ui->data_len, hash_ino_host);
+ if (err)
+ goto out_ro;
+
+ if (in_orphan && inode->i_nlink)
+ ubifs_delete_orphan(c, inode->i_ino);
+
+ finish_reservation(c);
+ spin_lock(&ui->ui_lock);
+ ui->synced_i_size = ui->ui_size;
+ spin_unlock(&ui->ui_lock);
+ if (xent) {
+ spin_lock(&host_ui->ui_lock);
+ host_ui->synced_i_size = host_ui->ui_size;
+ spin_unlock(&host_ui->ui_lock);
+ }
+ mark_inode_clean(c, ui);
+ mark_inode_clean(c, host_ui);
+ return 0;
+
+out_finish:
+ finish_reservation(c);
+out_free:
+ kfree(dent);
+ return err;
+
+out_release:
+ release_head(c, BASEHD);
+ kfree(dent);
+out_ro:
+ ubifs_ro_mode(c, err);
+ if (orphan_added)
+ ubifs_delete_orphan(c, inode->i_ino);
+ finish_reservation(c);
+ return err;
+}
+
+/**
+ * ubifs_jnl_write_data - write a data node to the journal.
+ * @c: UBIFS file-system description object
+ * @inode: inode the data node belongs to
+ * @key: node key
+ * @buf: buffer to write
+ * @len: data length (must not exceed %UBIFS_BLOCK_SIZE)
+ *
+ * This function writes a data node to the journal. Returns %0 if the data node
+ * was successfully written, and a negative error code in case of failure.
+ */
+int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
+ const union ubifs_key *key, const void *buf, int len)
+{
+ struct ubifs_data_node *data;
+ int err, lnum, offs, compr_type, out_len, compr_len, auth_len;
+ int dlen = COMPRESSED_DATA_NODE_BUF_SZ, allocated = 1;
+ int write_len;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ bool encrypted = IS_ENCRYPTED(inode);
+ u8 hash[UBIFS_HASH_ARR_SZ];
+
+ dbg_jnlk(key, "ino %lu, blk %u, len %d, key ",
+ (unsigned long)key_inum(c, key), key_block(c, key), len);
+ ubifs_assert(c, len <= UBIFS_BLOCK_SIZE);
+
+ if (encrypted)
+ dlen += UBIFS_CIPHER_BLOCK_SIZE;
+
+ auth_len = ubifs_auth_node_sz(c);
+
+ data = kmalloc(dlen + auth_len, GFP_NOFS | __GFP_NOWARN);
+ if (!data) {
+ /*
+ * Fall-back to the write reserve buffer. Note, we might be
+ * currently on the memory reclaim path, when the kernel is
+ * trying to free some memory by writing out dirty pages. The
+ * write reserve buffer helps us to guarantee that we are
+ * always able to write the data.
+ */
+ allocated = 0;
+ mutex_lock(&c->write_reserve_mutex);
+ data = c->write_reserve_buf;
+ }
+
+ data->ch.node_type = UBIFS_DATA_NODE;
+ key_write(c, key, &data->key);
+ data->size = cpu_to_le32(len);
+
+ if (!(ui->flags & UBIFS_COMPR_FL))
+ /* Compression is disabled for this inode */
+ compr_type = UBIFS_COMPR_NONE;
+ else
+ compr_type = ui->compr_type;
+
+ out_len = compr_len = dlen - UBIFS_DATA_NODE_SZ;
+ ubifs_compress(c, buf, len, &data->data, &compr_len, &compr_type);
+ ubifs_assert(c, compr_len <= UBIFS_BLOCK_SIZE);
+
+ if (encrypted) {
+ err = ubifs_encrypt(inode, data, compr_len, &out_len, key_block(c, key));
+ if (err)
+ goto out_free;
+
+ } else {
+ data->compr_size = 0;
+ out_len = compr_len;
+ }
+
+ dlen = UBIFS_DATA_NODE_SZ + out_len;
+ if (ubifs_authenticated(c))
+ write_len = ALIGN(dlen, 8) + auth_len;
+ else
+ write_len = dlen;
+
+ data->compr_type = cpu_to_le16(compr_type);
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, DATAHD, write_len);
+ if (err)
+ goto out_free;
+
+ ubifs_prepare_node(c, data, dlen, 0);
+ err = write_head(c, DATAHD, data, write_len, &lnum, &offs, 0);
+ if (err)
+ goto out_release;
+
+ err = ubifs_node_calc_hash(c, data, hash);
+ if (err)
+ goto out_release;
+
+ ubifs_wbuf_add_ino_nolock(&c->jheads[DATAHD].wbuf, key_inum(c, key));
+ release_head(c, DATAHD);
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ err = ubifs_tnc_add(c, key, lnum, offs, dlen, hash);
+ if (err)
+ goto out_ro;
+
+ finish_reservation(c);
+ if (!allocated)
+ mutex_unlock(&c->write_reserve_mutex);
+ else
+ kfree(data);
+ return 0;
+
+out_release:
+ release_head(c, DATAHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+out_free:
+ if (!allocated)
+ mutex_unlock(&c->write_reserve_mutex);
+ else
+ kfree(data);
+ return err;
+}
+
+/**
+ * ubifs_jnl_write_inode - flush inode to the journal.
+ * @c: UBIFS file-system description object
+ * @inode: inode to flush
+ *
+ * This function writes inode @inode to the journal. If the inode is
+ * synchronous, it also synchronizes the write-buffer. Returns zero in case of
+ * success and a negative error code in case of failure.
+ */
+int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode)
+{
+ int err, lnum, offs;
+ struct ubifs_ino_node *ino, *ino_start;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ int sync = 0, write_len = 0, ilen = UBIFS_INO_NODE_SZ;
+ int last_reference = !inode->i_nlink;
+ int kill_xattrs = ui->xattr_cnt && last_reference;
+ u8 hash[UBIFS_HASH_ARR_SZ];
+
+ dbg_jnl("ino %lu, nlink %u", inode->i_ino, inode->i_nlink);
+
+ /*
+ * If the inode is being deleted, do not write the attached data. No
+ * need to synchronize the write-buffer either.
+ */
+ if (!last_reference) {
+ ilen += ui->data_len;
+ sync = IS_SYNC(inode);
+ } else if (kill_xattrs) {
+ write_len += UBIFS_INO_NODE_SZ * ui->xattr_cnt;
+ }
+
+ if (ubifs_authenticated(c))
+ write_len += ALIGN(ilen, 8) + ubifs_auth_node_sz(c);
+ else
+ write_len += ilen;
+
+ ino_start = ino = kmalloc(write_len, GFP_NOFS);
+ if (!ino)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, write_len);
+ if (err)
+ goto out_free;
+
+ if (kill_xattrs) {
+ union ubifs_key key;
+ struct fscrypt_name nm = {0};
+ struct inode *xino;
+ struct ubifs_dent_node *xent, *pxent = NULL;
+
+ if (ui->xattr_cnt > ubifs_xattr_max_cnt(c)) {
+ err = -EPERM;
+ ubifs_err(c, "Cannot delete inode, it has too much xattrs!");
+ goto out_release;
+ }
+
+ lowest_xent_key(c, &key, inode->i_ino);
+ while (1) {
+ xent = ubifs_tnc_next_ent(c, &key, &nm);
+ if (IS_ERR(xent)) {
+ err = PTR_ERR(xent);
+ if (err == -ENOENT)
+ break;
+
+ kfree(pxent);
+ goto out_release;
+ }
+
+ 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);
+ kfree(pxent);
+ kfree(xent);
+ goto out_release;
+ }
+ ubifs_assert(c, ubifs_inode(xino)->xattr);
+
+ clear_nlink(xino);
+ pack_inode(c, ino, xino, 0);
+ ino = (void *)ino + UBIFS_INO_NODE_SZ;
+ iput(xino);
+
+ kfree(pxent);
+ pxent = xent;
+ key_read(c, &xent->key, &key);
+ }
+ kfree(pxent);
+ }
+
+ pack_inode(c, ino, inode, 1);
+ err = ubifs_node_calc_hash(c, ino, hash);
+ if (err)
+ goto out_release;
+
+ err = write_head(c, BASEHD, ino_start, write_len, &lnum, &offs, sync);
+ if (err)
+ goto out_release;
+ if (!sync)
+ ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf,
+ inode->i_ino);
+ release_head(c, BASEHD);
+
+ if (last_reference) {
+ err = ubifs_tnc_remove_ino(c, inode->i_ino);
+ if (err)
+ goto out_ro;
+ ubifs_delete_orphan(c, inode->i_ino);
+ err = ubifs_add_dirt(c, lnum, write_len);
+ } else {
+ union ubifs_key key;
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ ino_key_init(c, &key, inode->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, ilen, hash);
+ }
+ if (err)
+ goto out_ro;
+
+ finish_reservation(c);
+ spin_lock(&ui->ui_lock);
+ ui->synced_i_size = ui->ui_size;
+ spin_unlock(&ui->ui_lock);
+ kfree(ino_start);
+ return 0;
+
+out_release:
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+out_free:
+ kfree(ino_start);
+ return err;
+}
+
+/**
+ * ubifs_jnl_delete_inode - delete an inode.
+ * @c: UBIFS file-system description object
+ * @inode: inode to delete
+ *
+ * This function deletes inode @inode which includes removing it from orphans,
+ * deleting it from TNC and, in some cases, writing a deletion inode to the
+ * journal.
+ *
+ * When regular file inodes are unlinked or a directory inode is removed, the
+ * 'ubifs_jnl_update()' function writes a corresponding deletion inode and
+ * direntry to the media, and adds the inode to orphans. After this, when the
+ * last reference to this inode has been dropped, this function is called. In
+ * general, it has to write one more deletion inode to the media, because if
+ * a commit happened between 'ubifs_jnl_update()' and
+ * 'ubifs_jnl_delete_inode()', the deletion inode is not in the journal
+ * anymore, and in fact it might not be on the flash anymore, because it might
+ * have been garbage-collected already. And for optimization reasons UBIFS does
+ * not read the orphan area if it has been unmounted cleanly, so it would have
+ * no indication in the journal that there is a deleted inode which has to be
+ * removed from TNC.
+ *
+ * However, if there was no commit between 'ubifs_jnl_update()' and
+ * 'ubifs_jnl_delete_inode()', then there is no need to write the deletion
+ * inode to the media for the second time. And this is quite a typical case.
+ *
+ * This function returns zero in case of success and a negative error code in
+ * case of failure.
+ */
+int ubifs_jnl_delete_inode(struct ubifs_info *c, const struct inode *inode)
+{
+ int err;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+
+ ubifs_assert(c, inode->i_nlink == 0);
+
+ if (ui->xattr_cnt || ui->del_cmtno != c->cmt_no)
+ /* A commit happened for sure or inode hosts xattrs */
+ return ubifs_jnl_write_inode(c, inode);
+
+ down_read(&c->commit_sem);
+ /*
+ * Check commit number again, because the first test has been done
+ * without @c->commit_sem, so a commit might have happened.
+ */
+ if (ui->del_cmtno != c->cmt_no) {
+ up_read(&c->commit_sem);
+ return ubifs_jnl_write_inode(c, inode);
+ }
+
+ err = ubifs_tnc_remove_ino(c, inode->i_ino);
+ if (err)
+ ubifs_ro_mode(c, err);
+ else
+ ubifs_delete_orphan(c, inode->i_ino);
+ up_read(&c->commit_sem);
+ return err;
+}
+
+/**
+ * ubifs_jnl_xrename - cross rename two directory entries.
+ * @c: UBIFS file-system description object
+ * @fst_dir: parent inode of 1st directory entry to exchange
+ * @fst_inode: 1st inode to exchange
+ * @fst_nm: name of 1st inode to exchange
+ * @snd_dir: parent inode of 2nd directory entry to exchange
+ * @snd_inode: 2nd inode to exchange
+ * @snd_nm: name of 2nd inode to exchange
+ * @sync: non-zero if the write-buffer has to be synchronized
+ *
+ * This function implements the cross rename operation which may involve
+ * writing 2 inodes and 2 directory entries. It marks the written inodes as clean
+ * and returns zero on success. In case of failure, a negative error code is
+ * returned.
+ */
+int ubifs_jnl_xrename(struct ubifs_info *c, const struct inode *fst_dir,
+ const struct inode *fst_inode,
+ const struct fscrypt_name *fst_nm,
+ const struct inode *snd_dir,
+ const struct inode *snd_inode,
+ const struct fscrypt_name *snd_nm, int sync)
+{
+ union ubifs_key key;
+ struct ubifs_dent_node *dent1, *dent2;
+ int err, dlen1, dlen2, lnum, offs, len, plen = UBIFS_INO_NODE_SZ;
+ int aligned_dlen1, aligned_dlen2;
+ int twoparents = (fst_dir != snd_dir);
+ void *p;
+ u8 hash_dent1[UBIFS_HASH_ARR_SZ];
+ u8 hash_dent2[UBIFS_HASH_ARR_SZ];
+ u8 hash_p1[UBIFS_HASH_ARR_SZ];
+ u8 hash_p2[UBIFS_HASH_ARR_SZ];
+
+ ubifs_assert(c, ubifs_inode(fst_dir)->data_len == 0);
+ ubifs_assert(c, ubifs_inode(snd_dir)->data_len == 0);
+ ubifs_assert(c, mutex_is_locked(&ubifs_inode(fst_dir)->ui_mutex));
+ ubifs_assert(c, mutex_is_locked(&ubifs_inode(snd_dir)->ui_mutex));
+
+ dlen1 = UBIFS_DENT_NODE_SZ + fname_len(snd_nm) + 1;
+ dlen2 = UBIFS_DENT_NODE_SZ + fname_len(fst_nm) + 1;
+ aligned_dlen1 = ALIGN(dlen1, 8);
+ aligned_dlen2 = ALIGN(dlen2, 8);
+
+ len = aligned_dlen1 + aligned_dlen2 + ALIGN(plen, 8);
+ if (twoparents)
+ len += plen;
+
+ len += ubifs_auth_node_sz(c);
+
+ dent1 = kzalloc(len, GFP_NOFS);
+ if (!dent1)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, len);
+ if (err)
+ goto out_free;
+
+ /* Make new dent for 1st entry */
+ dent1->ch.node_type = UBIFS_DENT_NODE;
+ dent_key_init_flash(c, &dent1->key, snd_dir->i_ino, snd_nm);
+ dent1->inum = cpu_to_le64(fst_inode->i_ino);
+ dent1->type = get_dent_type(fst_inode->i_mode);
+ dent1->nlen = cpu_to_le16(fname_len(snd_nm));
+ memcpy(dent1->name, fname_name(snd_nm), fname_len(snd_nm));
+ dent1->name[fname_len(snd_nm)] = '\0';
+ set_dent_cookie(c, dent1);
+ zero_dent_node_unused(dent1);
+ ubifs_prep_grp_node(c, dent1, dlen1, 0);
+ err = ubifs_node_calc_hash(c, dent1, hash_dent1);
+ if (err)
+ goto out_release;
+
+ /* Make new dent for 2nd entry */
+ dent2 = (void *)dent1 + aligned_dlen1;
+ dent2->ch.node_type = UBIFS_DENT_NODE;
+ dent_key_init_flash(c, &dent2->key, fst_dir->i_ino, fst_nm);
+ dent2->inum = cpu_to_le64(snd_inode->i_ino);
+ dent2->type = get_dent_type(snd_inode->i_mode);
+ dent2->nlen = cpu_to_le16(fname_len(fst_nm));
+ memcpy(dent2->name, fname_name(fst_nm), fname_len(fst_nm));
+ dent2->name[fname_len(fst_nm)] = '\0';
+ set_dent_cookie(c, dent2);
+ zero_dent_node_unused(dent2);
+ ubifs_prep_grp_node(c, dent2, dlen2, 0);
+ err = ubifs_node_calc_hash(c, dent2, hash_dent2);
+ if (err)
+ goto out_release;
+
+ p = (void *)dent2 + aligned_dlen2;
+ if (!twoparents) {
+ pack_inode(c, p, fst_dir, 1);
+ err = ubifs_node_calc_hash(c, p, hash_p1);
+ if (err)
+ goto out_release;
+ } else {
+ pack_inode(c, p, fst_dir, 0);
+ err = ubifs_node_calc_hash(c, p, hash_p1);
+ if (err)
+ goto out_release;
+ p += ALIGN(plen, 8);
+ pack_inode(c, p, snd_dir, 1);
+ err = ubifs_node_calc_hash(c, p, hash_p2);
+ if (err)
+ goto out_release;
+ }
+
+ err = write_head(c, BASEHD, dent1, len, &lnum, &offs, sync);
+ if (err)
+ goto out_release;
+ if (!sync) {
+ struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
+
+ ubifs_wbuf_add_ino_nolock(wbuf, fst_dir->i_ino);
+ ubifs_wbuf_add_ino_nolock(wbuf, snd_dir->i_ino);
+ }
+ release_head(c, BASEHD);
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ dent_key_init(c, &key, snd_dir->i_ino, snd_nm);
+ err = ubifs_tnc_add_nm(c, &key, lnum, offs, dlen1, hash_dent1, snd_nm);
+ if (err)
+ goto out_ro;
+
+ offs += aligned_dlen1;
+ dent_key_init(c, &key, fst_dir->i_ino, fst_nm);
+ err = ubifs_tnc_add_nm(c, &key, lnum, offs, dlen2, hash_dent2, fst_nm);
+ if (err)
+ goto out_ro;
+
+ offs += aligned_dlen2;
+
+ ino_key_init(c, &key, fst_dir->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, plen, hash_p1);
+ if (err)
+ goto out_ro;
+
+ if (twoparents) {
+ offs += ALIGN(plen, 8);
+ ino_key_init(c, &key, snd_dir->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, plen, hash_p2);
+ if (err)
+ goto out_ro;
+ }
+
+ finish_reservation(c);
+
+ mark_inode_clean(c, ubifs_inode(fst_dir));
+ if (twoparents)
+ mark_inode_clean(c, ubifs_inode(snd_dir));
+ kfree(dent1);
+ return 0;
+
+out_release:
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+out_free:
+ kfree(dent1);
+ return err;
+}
+
+/**
+ * ubifs_jnl_rename - rename a directory entry.
+ * @c: UBIFS file-system description object
+ * @old_dir: parent inode of directory entry to rename
+ * @old_inode: directory entry's inode to rename
+ * @old_nm: name of the old directory entry to rename
+ * @new_dir: parent inode of directory entry to rename
+ * @new_inode: new directory entry's inode (or directory entry's inode to
+ * replace)
+ * @new_nm: new name of the new directory entry
+ * @whiteout: whiteout inode
+ * @sync: non-zero if the write-buffer has to be synchronized
+ * @delete_orphan: indicates an orphan entry deletion for @whiteout
+ *
+ * This function implements the re-name operation which may involve writing up
+ * to 4 inodes(new inode, whiteout inode, old and new parent directory inodes)
+ * and 2 directory entries. It marks the written inodes as clean and returns
+ * zero on success. In case of failure, a negative error code is returned.
+ */
+int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
+ const struct inode *old_inode,
+ const struct fscrypt_name *old_nm,
+ const struct inode *new_dir,
+ const struct inode *new_inode,
+ const struct fscrypt_name *new_nm,
+ const struct inode *whiteout, int sync, int delete_orphan)
+{
+ void *p;
+ union ubifs_key key;
+ struct ubifs_dent_node *dent, *dent2;
+ int err, dlen1, dlen2, ilen, wlen, lnum, offs, len, orphan_added = 0;
+ int aligned_dlen1, aligned_dlen2, plen = UBIFS_INO_NODE_SZ;
+ int last_reference = !!(new_inode && new_inode->i_nlink == 0);
+ int move = (old_dir != new_dir);
+ struct ubifs_inode *new_ui, *whiteout_ui;
+ u8 hash_old_dir[UBIFS_HASH_ARR_SZ];
+ u8 hash_new_dir[UBIFS_HASH_ARR_SZ];
+ u8 hash_new_inode[UBIFS_HASH_ARR_SZ];
+ u8 hash_whiteout_inode[UBIFS_HASH_ARR_SZ];
+ u8 hash_dent1[UBIFS_HASH_ARR_SZ];
+ u8 hash_dent2[UBIFS_HASH_ARR_SZ];
+
+ ubifs_assert(c, ubifs_inode(old_dir)->data_len == 0);
+ ubifs_assert(c, ubifs_inode(new_dir)->data_len == 0);
+ ubifs_assert(c, mutex_is_locked(&ubifs_inode(old_dir)->ui_mutex));
+ ubifs_assert(c, mutex_is_locked(&ubifs_inode(new_dir)->ui_mutex));
+
+ dlen1 = UBIFS_DENT_NODE_SZ + fname_len(new_nm) + 1;
+ dlen2 = UBIFS_DENT_NODE_SZ + fname_len(old_nm) + 1;
+ if (new_inode) {
+ new_ui = ubifs_inode(new_inode);
+ ubifs_assert(c, mutex_is_locked(&new_ui->ui_mutex));
+ ilen = UBIFS_INO_NODE_SZ;
+ if (!last_reference)
+ ilen += new_ui->data_len;
+ } else
+ ilen = 0;
+
+ if (whiteout) {
+ whiteout_ui = ubifs_inode(whiteout);
+ ubifs_assert(c, mutex_is_locked(&whiteout_ui->ui_mutex));
+ ubifs_assert(c, whiteout->i_nlink == 1);
+ ubifs_assert(c, !whiteout_ui->dirty);
+ wlen = UBIFS_INO_NODE_SZ;
+ wlen += whiteout_ui->data_len;
+ } else
+ wlen = 0;
+
+ aligned_dlen1 = ALIGN(dlen1, 8);
+ aligned_dlen2 = ALIGN(dlen2, 8);
+ len = aligned_dlen1 + aligned_dlen2 + ALIGN(ilen, 8) +
+ ALIGN(wlen, 8) + ALIGN(plen, 8);
+ if (move)
+ len += plen;
+
+ len += ubifs_auth_node_sz(c);
+
+ dent = kzalloc(len, GFP_NOFS);
+ if (!dent)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, len);
+ if (err)
+ goto out_free;
+
+ /* Make new dent */
+ dent->ch.node_type = UBIFS_DENT_NODE;
+ dent_key_init_flash(c, &dent->key, new_dir->i_ino, new_nm);
+ dent->inum = cpu_to_le64(old_inode->i_ino);
+ dent->type = get_dent_type(old_inode->i_mode);
+ dent->nlen = cpu_to_le16(fname_len(new_nm));
+ memcpy(dent->name, fname_name(new_nm), fname_len(new_nm));
+ dent->name[fname_len(new_nm)] = '\0';
+ set_dent_cookie(c, dent);
+ zero_dent_node_unused(dent);
+ ubifs_prep_grp_node(c, dent, dlen1, 0);
+ err = ubifs_node_calc_hash(c, dent, hash_dent1);
+ if (err)
+ goto out_release;
+
+ dent2 = (void *)dent + aligned_dlen1;
+ dent2->ch.node_type = UBIFS_DENT_NODE;
+ dent_key_init_flash(c, &dent2->key, old_dir->i_ino, old_nm);
+
+ if (whiteout) {
+ dent2->inum = cpu_to_le64(whiteout->i_ino);
+ dent2->type = get_dent_type(whiteout->i_mode);
+ } else {
+ /* Make deletion dent */
+ dent2->inum = 0;
+ dent2->type = DT_UNKNOWN;
+ }
+ dent2->nlen = cpu_to_le16(fname_len(old_nm));
+ memcpy(dent2->name, fname_name(old_nm), fname_len(old_nm));
+ dent2->name[fname_len(old_nm)] = '\0';
+ set_dent_cookie(c, dent2);
+ zero_dent_node_unused(dent2);
+ ubifs_prep_grp_node(c, dent2, dlen2, 0);
+ err = ubifs_node_calc_hash(c, dent2, hash_dent2);
+ if (err)
+ goto out_release;
+
+ p = (void *)dent2 + aligned_dlen2;
+ if (new_inode) {
+ pack_inode(c, p, new_inode, 0);
+ err = ubifs_node_calc_hash(c, p, hash_new_inode);
+ if (err)
+ goto out_release;
+
+ p += ALIGN(ilen, 8);
+ }
+
+ if (whiteout) {
+ pack_inode(c, p, whiteout, 0);
+ err = ubifs_node_calc_hash(c, p, hash_whiteout_inode);
+ if (err)
+ goto out_release;
+
+ p += ALIGN(wlen, 8);
+ }
+
+ if (!move) {
+ pack_inode(c, p, old_dir, 1);
+ err = ubifs_node_calc_hash(c, p, hash_old_dir);
+ if (err)
+ goto out_release;
+ } else {
+ pack_inode(c, p, old_dir, 0);
+ err = ubifs_node_calc_hash(c, p, hash_old_dir);
+ if (err)
+ goto out_release;
+
+ p += ALIGN(plen, 8);
+ pack_inode(c, p, new_dir, 1);
+ err = ubifs_node_calc_hash(c, p, hash_new_dir);
+ if (err)
+ goto out_release;
+ }
+
+ if (last_reference) {
+ err = ubifs_add_orphan(c, new_inode->i_ino);
+ if (err) {
+ release_head(c, BASEHD);
+ goto out_finish;
+ }
+ new_ui->del_cmtno = c->cmt_no;
+ orphan_added = 1;
+ }
+
+ err = write_head(c, BASEHD, dent, len, &lnum, &offs, sync);
+ if (err)
+ goto out_release;
+ if (!sync) {
+ struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
+
+ ubifs_wbuf_add_ino_nolock(wbuf, new_dir->i_ino);
+ ubifs_wbuf_add_ino_nolock(wbuf, old_dir->i_ino);
+ if (new_inode)
+ ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf,
+ new_inode->i_ino);
+ if (whiteout)
+ ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf,
+ whiteout->i_ino);
+ }
+ release_head(c, BASEHD);
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ dent_key_init(c, &key, new_dir->i_ino, new_nm);
+ err = ubifs_tnc_add_nm(c, &key, lnum, offs, dlen1, hash_dent1, new_nm);
+ if (err)
+ goto out_ro;
+
+ offs += aligned_dlen1;
+ if (whiteout) {
+ dent_key_init(c, &key, old_dir->i_ino, old_nm);
+ err = ubifs_tnc_add_nm(c, &key, lnum, offs, dlen2, hash_dent2, old_nm);
+ if (err)
+ goto out_ro;
+ } else {
+ err = ubifs_add_dirt(c, lnum, dlen2);
+ if (err)
+ goto out_ro;
+
+ dent_key_init(c, &key, old_dir->i_ino, old_nm);
+ err = ubifs_tnc_remove_nm(c, &key, old_nm);
+ if (err)
+ goto out_ro;
+ }
+
+ offs += aligned_dlen2;
+ if (new_inode) {
+ ino_key_init(c, &key, new_inode->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, ilen, hash_new_inode);
+ if (err)
+ goto out_ro;
+ offs += ALIGN(ilen, 8);
+ }
+
+ if (whiteout) {
+ ino_key_init(c, &key, whiteout->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, wlen,
+ hash_whiteout_inode);
+ if (err)
+ goto out_ro;
+ offs += ALIGN(wlen, 8);
+ }
+
+ ino_key_init(c, &key, old_dir->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, plen, hash_old_dir);
+ if (err)
+ goto out_ro;
+
+ if (move) {
+ offs += ALIGN(plen, 8);
+ ino_key_init(c, &key, new_dir->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, plen, hash_new_dir);
+ if (err)
+ goto out_ro;
+ }
+
+ if (delete_orphan)
+ ubifs_delete_orphan(c, whiteout->i_ino);
+
+ finish_reservation(c);
+ if (new_inode) {
+ mark_inode_clean(c, new_ui);
+ spin_lock(&new_ui->ui_lock);
+ new_ui->synced_i_size = new_ui->ui_size;
+ spin_unlock(&new_ui->ui_lock);
+ }
+ /*
+ * No need to mark whiteout inode clean.
+ * Whiteout doesn't have non-zero size, no need to update
+ * synced_i_size for whiteout_ui.
+ */
+ mark_inode_clean(c, ubifs_inode(old_dir));
+ if (move)
+ mark_inode_clean(c, ubifs_inode(new_dir));
+ kfree(dent);
+ return 0;
+
+out_release:
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ if (orphan_added)
+ ubifs_delete_orphan(c, new_inode->i_ino);
+out_finish:
+ finish_reservation(c);
+out_free:
+ kfree(dent);
+ return err;
+}
+
+/**
+ * truncate_data_node - re-compress/encrypt a truncated data node.
+ * @c: UBIFS file-system description object
+ * @inode: inode which refers to the data node
+ * @block: data block number
+ * @dn: data node to re-compress
+ * @new_len: new length
+ * @dn_size: size of the data node @dn in memory
+ *
+ * This function is used when an inode is truncated and the last data node of
+ * the inode has to be re-compressed/encrypted and re-written.
+ */
+static int truncate_data_node(const struct ubifs_info *c, const struct inode *inode,
+ unsigned int block, struct ubifs_data_node *dn,
+ int *new_len, int dn_size)
+{
+ void *buf;
+ int err, dlen, compr_type, out_len, data_size;
+
+ out_len = le32_to_cpu(dn->size);
+ buf = kmalloc_array(out_len, WORST_COMPR_FACTOR, GFP_NOFS);
+ if (!buf)
+ return -ENOMEM;
+
+ dlen = le32_to_cpu(dn->ch.len) - UBIFS_DATA_NODE_SZ;
+ data_size = dn_size - UBIFS_DATA_NODE_SZ;
+ compr_type = le16_to_cpu(dn->compr_type);
+
+ if (IS_ENCRYPTED(inode)) {
+ err = ubifs_decrypt(inode, dn, &dlen, block);
+ if (err)
+ goto out;
+ }
+
+ if (compr_type == UBIFS_COMPR_NONE) {
+ out_len = *new_len;
+ } else {
+ err = ubifs_decompress(c, &dn->data, dlen, buf, &out_len, compr_type);
+ if (err)
+ goto out;
+
+ ubifs_compress(c, buf, *new_len, &dn->data, &out_len, &compr_type);
+ }
+
+ if (IS_ENCRYPTED(inode)) {
+ err = ubifs_encrypt(inode, dn, out_len, &data_size, block);
+ if (err)
+ goto out;
+
+ out_len = data_size;
+ } else {
+ dn->compr_size = 0;
+ }
+
+ ubifs_assert(c, out_len <= UBIFS_BLOCK_SIZE);
+ dn->compr_type = cpu_to_le16(compr_type);
+ dn->size = cpu_to_le32(*new_len);
+ *new_len = UBIFS_DATA_NODE_SZ + out_len;
+ err = 0;
+out:
+ kfree(buf);
+ return err;
+}
+
+/**
+ * ubifs_jnl_truncate - update the journal for a truncation.
+ * @c: UBIFS file-system description object
+ * @inode: inode to truncate
+ * @old_size: old size
+ * @new_size: new size
+ *
+ * When the size of a file decreases due to truncation, a truncation node is
+ * written, the journal tree is updated, and the last data block is re-written
+ * if it has been affected. The inode is also updated in order to synchronize
+ * the new inode size.
+ *
+ * This function marks the inode as clean and returns zero on success. In case
+ * of failure, a negative error code is returned.
+ */
+int ubifs_jnl_truncate(struct ubifs_info *c, const struct inode *inode,
+ loff_t old_size, loff_t new_size)
+{
+ union ubifs_key key, to_key;
+ struct ubifs_ino_node *ino;
+ struct ubifs_trun_node *trun;
+ struct ubifs_data_node *dn;
+ int err, dlen, len, lnum, offs, bit, sz, sync = IS_SYNC(inode);
+ int dn_size;
+ struct ubifs_inode *ui = ubifs_inode(inode);
+ ino_t inum = inode->i_ino;
+ unsigned int blk;
+ u8 hash_ino[UBIFS_HASH_ARR_SZ];
+ u8 hash_dn[UBIFS_HASH_ARR_SZ];
+
+ dbg_jnl("ino %lu, size %lld -> %lld",
+ (unsigned long)inum, old_size, new_size);
+ ubifs_assert(c, !ui->data_len);
+ ubifs_assert(c, S_ISREG(inode->i_mode));
+ ubifs_assert(c, mutex_is_locked(&ui->ui_mutex));
+
+ dn_size = COMPRESSED_DATA_NODE_BUF_SZ;
+
+ if (IS_ENCRYPTED(inode))
+ dn_size += UBIFS_CIPHER_BLOCK_SIZE;
+
+ sz = UBIFS_TRUN_NODE_SZ + UBIFS_INO_NODE_SZ +
+ dn_size + ubifs_auth_node_sz(c);
+
+ ino = kmalloc(sz, GFP_NOFS);
+ if (!ino)
+ return -ENOMEM;
+
+ trun = (void *)ino + UBIFS_INO_NODE_SZ;
+ trun->ch.node_type = UBIFS_TRUN_NODE;
+ trun->inum = cpu_to_le32(inum);
+ trun->old_size = cpu_to_le64(old_size);
+ trun->new_size = cpu_to_le64(new_size);
+ zero_trun_node_unused(trun);
+
+ dlen = new_size & (UBIFS_BLOCK_SIZE - 1);
+ if (dlen) {
+ /* Get last data block so it can be truncated */
+ dn = (void *)trun + UBIFS_TRUN_NODE_SZ;
+ blk = new_size >> UBIFS_BLOCK_SHIFT;
+ data_key_init(c, &key, inum, blk);
+ dbg_jnlk(&key, "last block key ");
+ err = ubifs_tnc_lookup(c, &key, dn);
+ if (err == -ENOENT)
+ dlen = 0; /* Not found (so it is a hole) */
+ else if (err)
+ goto out_free;
+ else {
+ int dn_len = le32_to_cpu(dn->size);
+
+ if (dn_len <= 0 || dn_len > UBIFS_BLOCK_SIZE) {
+ ubifs_err(c, "bad data node (block %u, inode %lu)",
+ blk, inode->i_ino);
+ ubifs_dump_node(c, dn, dn_size);
+ err = -EUCLEAN;
+ goto out_free;
+ }
+
+ if (dn_len <= dlen)
+ dlen = 0; /* Nothing to do */
+ else {
+ err = truncate_data_node(c, inode, blk, dn,
+ &dlen, dn_size);
+ if (err)
+ goto out_free;
+ }
+ }
+ }
+
+ /* Must make reservation before allocating sequence numbers */
+ len = UBIFS_TRUN_NODE_SZ + UBIFS_INO_NODE_SZ;
+
+ if (ubifs_authenticated(c))
+ len += ALIGN(dlen, 8) + ubifs_auth_node_sz(c);
+ else
+ len += dlen;
+
+ err = make_reservation(c, BASEHD, len);
+ if (err)
+ goto out_free;
+
+ pack_inode(c, ino, inode, 0);
+ err = ubifs_node_calc_hash(c, ino, hash_ino);
+ if (err)
+ goto out_release;
+
+ ubifs_prep_grp_node(c, trun, UBIFS_TRUN_NODE_SZ, dlen ? 0 : 1);
+ if (dlen) {
+ ubifs_prep_grp_node(c, dn, dlen, 1);
+ err = ubifs_node_calc_hash(c, dn, hash_dn);
+ if (err)
+ goto out_release;
+ }
+
+ err = write_head(c, BASEHD, ino, len, &lnum, &offs, sync);
+ if (err)
+ goto out_release;
+ if (!sync)
+ ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf, inum);
+ release_head(c, BASEHD);
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ if (dlen) {
+ sz = offs + UBIFS_INO_NODE_SZ + UBIFS_TRUN_NODE_SZ;
+ err = ubifs_tnc_add(c, &key, lnum, sz, dlen, hash_dn);
+ if (err)
+ goto out_ro;
+ }
+
+ ino_key_init(c, &key, inum);
+ err = ubifs_tnc_add(c, &key, lnum, offs, UBIFS_INO_NODE_SZ, hash_ino);
+ if (err)
+ goto out_ro;
+
+ err = ubifs_add_dirt(c, lnum, UBIFS_TRUN_NODE_SZ);
+ if (err)
+ goto out_ro;
+
+ bit = new_size & (UBIFS_BLOCK_SIZE - 1);
+ blk = (new_size >> UBIFS_BLOCK_SHIFT) + (bit ? 1 : 0);
+ data_key_init(c, &key, inum, blk);
+
+ bit = old_size & (UBIFS_BLOCK_SIZE - 1);
+ blk = (old_size >> UBIFS_BLOCK_SHIFT) - (bit ? 0 : 1);
+ data_key_init(c, &to_key, inum, blk);
+
+ err = ubifs_tnc_remove_range(c, &key, &to_key);
+ if (err)
+ goto out_ro;
+
+ finish_reservation(c);
+ spin_lock(&ui->ui_lock);
+ ui->synced_i_size = ui->ui_size;
+ spin_unlock(&ui->ui_lock);
+ mark_inode_clean(c, ui);
+ kfree(ino);
+ return 0;
+
+out_release:
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+out_free:
+ kfree(ino);
+ return err;
+}
+
+
+/**
+ * ubifs_jnl_delete_xattr - delete an extended attribute.
+ * @c: UBIFS file-system description object
+ * @host: host inode
+ * @inode: extended attribute inode
+ * @nm: extended attribute entry name
+ *
+ * This function delete an extended attribute which is very similar to
+ * un-linking regular files - it writes a deletion xentry, a deletion inode and
+ * updates the target inode. Returns zero in case of success and a negative
+ * error code in case of failure.
+ */
+int ubifs_jnl_delete_xattr(struct ubifs_info *c, const struct inode *host,
+ const struct inode *inode,
+ const struct fscrypt_name *nm)
+{
+ int err, xlen, hlen, len, lnum, xent_offs, aligned_xlen, write_len;
+ struct ubifs_dent_node *xent;
+ struct ubifs_ino_node *ino;
+ union ubifs_key xent_key, key1, key2;
+ int sync = IS_DIRSYNC(host);
+ struct ubifs_inode *host_ui = ubifs_inode(host);
+ u8 hash[UBIFS_HASH_ARR_SZ];
+
+ ubifs_assert(c, inode->i_nlink == 0);
+ ubifs_assert(c, mutex_is_locked(&host_ui->ui_mutex));
+
+ /*
+ * Since we are deleting the inode, we do not bother to attach any data
+ * to it and assume its length is %UBIFS_INO_NODE_SZ.
+ */
+ xlen = UBIFS_DENT_NODE_SZ + fname_len(nm) + 1;
+ aligned_xlen = ALIGN(xlen, 8);
+ hlen = host_ui->data_len + UBIFS_INO_NODE_SZ;
+ len = aligned_xlen + UBIFS_INO_NODE_SZ + ALIGN(hlen, 8);
+
+ write_len = len + ubifs_auth_node_sz(c);
+
+ xent = kzalloc(write_len, GFP_NOFS);
+ if (!xent)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, write_len);
+ if (err) {
+ kfree(xent);
+ return err;
+ }
+
+ xent->ch.node_type = UBIFS_XENT_NODE;
+ xent_key_init(c, &xent_key, host->i_ino, nm);
+ key_write(c, &xent_key, xent->key);
+ xent->inum = 0;
+ xent->type = get_dent_type(inode->i_mode);
+ xent->nlen = cpu_to_le16(fname_len(nm));
+ memcpy(xent->name, fname_name(nm), fname_len(nm));
+ xent->name[fname_len(nm)] = '\0';
+ zero_dent_node_unused(xent);
+ ubifs_prep_grp_node(c, xent, xlen, 0);
+
+ ino = (void *)xent + aligned_xlen;
+ pack_inode(c, ino, inode, 0);
+ ino = (void *)ino + UBIFS_INO_NODE_SZ;
+ pack_inode(c, ino, host, 1);
+ err = ubifs_node_calc_hash(c, ino, hash);
+ if (err)
+ goto out_release;
+
+ err = write_head(c, BASEHD, xent, write_len, &lnum, &xent_offs, sync);
+ if (!sync && !err)
+ ubifs_wbuf_add_ino_nolock(&c->jheads[BASEHD].wbuf, host->i_ino);
+ release_head(c, BASEHD);
+
+ ubifs_add_auth_dirt(c, lnum);
+ kfree(xent);
+ if (err)
+ goto out_ro;
+
+ /* Remove the extended attribute entry from TNC */
+ err = ubifs_tnc_remove_nm(c, &xent_key, nm);
+ if (err)
+ goto out_ro;
+ err = ubifs_add_dirt(c, lnum, xlen);
+ if (err)
+ goto out_ro;
+
+ /*
+ * Remove all nodes belonging to the extended attribute inode from TNC.
+ * Well, there actually must be only one node - the inode itself.
+ */
+ lowest_ino_key(c, &key1, inode->i_ino);
+ highest_ino_key(c, &key2, inode->i_ino);
+ err = ubifs_tnc_remove_range(c, &key1, &key2);
+ if (err)
+ goto out_ro;
+ err = ubifs_add_dirt(c, lnum, UBIFS_INO_NODE_SZ);
+ if (err)
+ goto out_ro;
+
+ /* And update TNC with the new host inode position */
+ ino_key_init(c, &key1, host->i_ino);
+ err = ubifs_tnc_add(c, &key1, lnum, xent_offs + len - hlen, hlen, hash);
+ if (err)
+ goto out_ro;
+
+ finish_reservation(c);
+ spin_lock(&host_ui->ui_lock);
+ host_ui->synced_i_size = host_ui->ui_size;
+ spin_unlock(&host_ui->ui_lock);
+ mark_inode_clean(c, host_ui);
+ return 0;
+
+out_release:
+ kfree(xent);
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+ return err;
+}
+
+/**
+ * ubifs_jnl_change_xattr - change an extended attribute.
+ * @c: UBIFS file-system description object
+ * @inode: extended attribute inode
+ * @host: host inode
+ *
+ * This function writes the updated version of an extended attribute inode and
+ * the host inode to the journal (to the base head). The host inode is written
+ * after the extended attribute inode in order to guarantee that the extended
+ * attribute will be flushed when the inode is synchronized by 'fsync()' and
+ * consequently, the write-buffer is synchronized. This function returns zero
+ * in case of success and a negative error code in case of failure.
+ */
+int ubifs_jnl_change_xattr(struct ubifs_info *c, const struct inode *inode,
+ const struct inode *host)
+{
+ int err, len1, len2, aligned_len, aligned_len1, lnum, offs;
+ struct ubifs_inode *host_ui = ubifs_inode(host);
+ struct ubifs_ino_node *ino;
+ union ubifs_key key;
+ int sync = IS_DIRSYNC(host);
+ u8 hash_host[UBIFS_HASH_ARR_SZ];
+ u8 hash[UBIFS_HASH_ARR_SZ];
+
+ dbg_jnl("ino %lu, ino %lu", host->i_ino, inode->i_ino);
+ ubifs_assert(c, inode->i_nlink > 0);
+ ubifs_assert(c, mutex_is_locked(&host_ui->ui_mutex));
+
+ len1 = UBIFS_INO_NODE_SZ + host_ui->data_len;
+ len2 = UBIFS_INO_NODE_SZ + ubifs_inode(inode)->data_len;
+ aligned_len1 = ALIGN(len1, 8);
+ aligned_len = aligned_len1 + ALIGN(len2, 8);
+
+ aligned_len += ubifs_auth_node_sz(c);
+
+ ino = kzalloc(aligned_len, GFP_NOFS);
+ if (!ino)
+ return -ENOMEM;
+
+ /* Make reservation before allocating sequence numbers */
+ err = make_reservation(c, BASEHD, aligned_len);
+ if (err)
+ goto out_free;
+
+ pack_inode(c, ino, host, 0);
+ err = ubifs_node_calc_hash(c, ino, hash_host);
+ if (err)
+ goto out_release;
+ pack_inode(c, (void *)ino + aligned_len1, inode, 1);
+ err = ubifs_node_calc_hash(c, (void *)ino + aligned_len1, hash);
+ if (err)
+ goto out_release;
+
+ err = write_head(c, BASEHD, ino, aligned_len, &lnum, &offs, 0);
+ if (!sync && !err) {
+ struct ubifs_wbuf *wbuf = &c->jheads[BASEHD].wbuf;
+
+ ubifs_wbuf_add_ino_nolock(wbuf, host->i_ino);
+ ubifs_wbuf_add_ino_nolock(wbuf, inode->i_ino);
+ }
+ release_head(c, BASEHD);
+ if (err)
+ goto out_ro;
+
+ ubifs_add_auth_dirt(c, lnum);
+
+ ino_key_init(c, &key, host->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs, len1, hash_host);
+ if (err)
+ goto out_ro;
+
+ ino_key_init(c, &key, inode->i_ino);
+ err = ubifs_tnc_add(c, &key, lnum, offs + aligned_len1, len2, hash);
+ if (err)
+ goto out_ro;
+
+ finish_reservation(c);
+ spin_lock(&host_ui->ui_lock);
+ host_ui->synced_i_size = host_ui->ui_size;
+ spin_unlock(&host_ui->ui_lock);
+ mark_inode_clean(c, host_ui);
+ kfree(ino);
+ return 0;
+
+out_release:
+ release_head(c, BASEHD);
+out_ro:
+ ubifs_ro_mode(c, err);
+ finish_reservation(c);
+out_free:
+ kfree(ino);
+ return err;
+}
+
new file mode 100644
@@ -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__ */
new file mode 100644
@@ -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(sizeof(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(sizeof(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;
+}
new file mode 100644
@@ -0,0 +1,1307 @@
+// 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 functions that access LEB properties and their
+ * categories. LEBs are categorized based on the needs of UBIFS, and the
+ * categories are stored as either heaps or lists to provide a fast way of
+ * finding a LEB in a particular category. For example, UBIFS may need to find
+ * an empty LEB for the journal, or a very dirty LEB for garbage collection.
+ */
+
+#include "ubifs.h"
+
+/**
+ * get_heap_comp_val - get the LEB properties value for heap comparisons.
+ * @lprops: LEB properties
+ * @cat: LEB category
+ */
+static int get_heap_comp_val(struct ubifs_lprops *lprops, int cat)
+{
+ switch (cat) {
+ case LPROPS_FREE:
+ return lprops->free;
+ case LPROPS_DIRTY_IDX:
+ return lprops->free + lprops->dirty;
+ default:
+ return lprops->dirty;
+ }
+}
+
+/**
+ * move_up_lpt_heap - move a new heap entry up as far as possible.
+ * @c: UBIFS file-system description object
+ * @heap: LEB category heap
+ * @lprops: LEB properties to move
+ * @cat: LEB category
+ *
+ * New entries to a heap are added at the bottom and then moved up until the
+ * parent's value is greater. In the case of LPT's category heaps, the value
+ * is either the amount of free space or the amount of dirty space, depending
+ * on the category.
+ */
+static void move_up_lpt_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
+ struct ubifs_lprops *lprops, int cat)
+{
+ int val1, val2, hpos;
+
+ hpos = lprops->hpos;
+ if (!hpos)
+ return; /* Already top of the heap */
+ val1 = get_heap_comp_val(lprops, cat);
+ /* Compare to parent and, if greater, move up the heap */
+ do {
+ int ppos = (hpos - 1) / 2;
+
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val2 >= val1)
+ return;
+ /* Greater than parent so move up */
+ heap->arr[ppos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[ppos];
+ heap->arr[ppos] = lprops;
+ lprops->hpos = ppos;
+ hpos = ppos;
+ } while (hpos);
+}
+
+/**
+ * adjust_lpt_heap - move a changed heap entry up or down the heap.
+ * @c: UBIFS file-system description object
+ * @heap: LEB category heap
+ * @lprops: LEB properties to move
+ * @hpos: heap position of @lprops
+ * @cat: LEB category
+ *
+ * Changed entries in a heap are moved up or down until the parent's value is
+ * greater. In the case of LPT's category heaps, the value is either the amount
+ * of free space or the amount of dirty space, depending on the category.
+ */
+static void adjust_lpt_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap,
+ struct ubifs_lprops *lprops, int hpos, int cat)
+{
+ int val1, val2, val3, cpos;
+
+ val1 = get_heap_comp_val(lprops, cat);
+ /* Compare to parent and, if greater than parent, move up the heap */
+ if (hpos) {
+ int ppos = (hpos - 1) / 2;
+
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val1 > val2) {
+ /* Greater than parent so move up */
+ while (1) {
+ heap->arr[ppos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[ppos];
+ heap->arr[ppos] = lprops;
+ lprops->hpos = ppos;
+ hpos = ppos;
+ if (!hpos)
+ return;
+ ppos = (hpos - 1) / 2;
+ val2 = get_heap_comp_val(heap->arr[ppos], cat);
+ if (val1 <= val2)
+ return;
+ /* Still greater than parent so keep going */
+ }
+ }
+ }
+
+ /* Not greater than parent, so compare to children */
+ while (1) {
+ /* Compare to left child */
+ cpos = hpos * 2 + 1;
+ if (cpos >= heap->cnt)
+ return;
+ val2 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 < val2) {
+ /* Less than left child, so promote biggest child */
+ if (cpos + 1 < heap->cnt) {
+ val3 = get_heap_comp_val(heap->arr[cpos + 1],
+ cat);
+ if (val3 > val2)
+ cpos += 1; /* Right child is bigger */
+ }
+ heap->arr[cpos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[cpos];
+ heap->arr[cpos] = lprops;
+ lprops->hpos = cpos;
+ hpos = cpos;
+ continue;
+ }
+ /* Compare to right child */
+ cpos += 1;
+ if (cpos >= heap->cnt)
+ return;
+ val3 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 < val3) {
+ /* Less than right child, so promote right child */
+ heap->arr[cpos]->hpos = hpos;
+ heap->arr[hpos] = heap->arr[cpos];
+ heap->arr[cpos] = lprops;
+ lprops->hpos = cpos;
+ hpos = cpos;
+ continue;
+ }
+ return;
+ }
+}
+
+/**
+ * add_to_lpt_heap - add LEB properties to a LEB category heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to add
+ * @cat: LEB category
+ *
+ * This function returns %1 if @lprops is added to the heap for LEB category
+ * @cat, otherwise %0 is returned because the heap is full.
+ */
+static int add_to_lpt_heap(struct ubifs_info *c, struct ubifs_lprops *lprops,
+ int cat)
+{
+ struct ubifs_lpt_heap *heap = &c->lpt_heap[cat - 1];
+
+ if (heap->cnt >= heap->max_cnt) {
+ const int b = LPT_HEAP_SZ / 2 - 1;
+ int cpos, val1, val2;
+
+ /* Compare to some other LEB on the bottom of heap */
+ /* Pick a position kind of randomly */
+ cpos = (((size_t)lprops >> 4) & b) + b;
+ ubifs_assert(c, cpos >= b);
+ ubifs_assert(c, cpos < LPT_HEAP_SZ);
+ ubifs_assert(c, cpos < heap->cnt);
+
+ val1 = get_heap_comp_val(lprops, cat);
+ val2 = get_heap_comp_val(heap->arr[cpos], cat);
+ if (val1 > val2) {
+ struct ubifs_lprops *lp;
+
+ lp = heap->arr[cpos];
+ lp->flags &= ~LPROPS_CAT_MASK;
+ lp->flags |= LPROPS_UNCAT;
+ list_add(&lp->list, &c->uncat_list);
+ lprops->hpos = cpos;
+ heap->arr[cpos] = lprops;
+ move_up_lpt_heap(c, heap, lprops, cat);
+ dbg_check_heap(c, heap, cat, lprops->hpos);
+ return 1; /* Added to heap */
+ }
+ dbg_check_heap(c, heap, cat, -1);
+ return 0; /* Not added to heap */
+ } else {
+ lprops->hpos = heap->cnt++;
+ heap->arr[lprops->hpos] = lprops;
+ move_up_lpt_heap(c, heap, lprops, cat);
+ dbg_check_heap(c, heap, cat, lprops->hpos);
+ return 1; /* Added to heap */
+ }
+}
+
+/**
+ * remove_from_lpt_heap - remove LEB properties from a LEB category heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to remove
+ * @cat: LEB category
+ */
+static void remove_from_lpt_heap(struct ubifs_info *c,
+ struct ubifs_lprops *lprops, int cat)
+{
+ struct ubifs_lpt_heap *heap;
+ int hpos = lprops->hpos;
+
+ heap = &c->lpt_heap[cat - 1];
+ ubifs_assert(c, hpos >= 0 && hpos < heap->cnt);
+ ubifs_assert(c, heap->arr[hpos] == lprops);
+ heap->cnt -= 1;
+ if (hpos < heap->cnt) {
+ heap->arr[hpos] = heap->arr[heap->cnt];
+ heap->arr[hpos]->hpos = hpos;
+ adjust_lpt_heap(c, heap, heap->arr[hpos], hpos, cat);
+ }
+ dbg_check_heap(c, heap, cat, -1);
+}
+
+/**
+ * lpt_heap_replace - replace lprops in a category heap.
+ * @c: UBIFS file-system description object
+ * @new_lprops: LEB properties with which to replace
+ * @cat: LEB category
+ *
+ * During commit it is sometimes necessary to copy a pnode (see dirty_cow_pnode)
+ * and the lprops that the pnode contains. When that happens, references in
+ * the category heaps to those lprops must be updated to point to the new
+ * lprops. This function does that.
+ */
+static void lpt_heap_replace(struct ubifs_info *c,
+ struct ubifs_lprops *new_lprops, int cat)
+{
+ struct ubifs_lpt_heap *heap;
+ int hpos = new_lprops->hpos;
+
+ heap = &c->lpt_heap[cat - 1];
+ heap->arr[hpos] = new_lprops;
+}
+
+/**
+ * ubifs_add_to_cat - add LEB properties to a category list or heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to add
+ * @cat: LEB category to which to add
+ *
+ * LEB properties are categorized to enable fast find operations.
+ */
+void ubifs_add_to_cat(struct ubifs_info *c, struct ubifs_lprops *lprops,
+ int cat)
+{
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ if (add_to_lpt_heap(c, lprops, cat))
+ break;
+ /* No more room on heap so make it un-categorized */
+ cat = LPROPS_UNCAT;
+ fallthrough;
+ case LPROPS_UNCAT:
+ list_add(&lprops->list, &c->uncat_list);
+ break;
+ case LPROPS_EMPTY:
+ list_add(&lprops->list, &c->empty_list);
+ break;
+ case LPROPS_FREEABLE:
+ list_add(&lprops->list, &c->freeable_list);
+ c->freeable_cnt += 1;
+ break;
+ case LPROPS_FRDI_IDX:
+ list_add(&lprops->list, &c->frdi_idx_list);
+ break;
+ default:
+ ubifs_assert(c, 0);
+ }
+
+ lprops->flags &= ~LPROPS_CAT_MASK;
+ lprops->flags |= cat;
+ c->in_a_category_cnt += 1;
+ ubifs_assert(c, c->in_a_category_cnt <= c->main_lebs);
+}
+
+/**
+ * ubifs_remove_from_cat - remove LEB properties from a category list or heap.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to remove
+ * @cat: LEB category from which to remove
+ *
+ * LEB properties are categorized to enable fast find operations.
+ */
+static void ubifs_remove_from_cat(struct ubifs_info *c,
+ struct ubifs_lprops *lprops, int cat)
+{
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ remove_from_lpt_heap(c, lprops, cat);
+ break;
+ case LPROPS_FREEABLE:
+ c->freeable_cnt -= 1;
+ ubifs_assert(c, c->freeable_cnt >= 0);
+ fallthrough;
+ case LPROPS_UNCAT:
+ case LPROPS_EMPTY:
+ case LPROPS_FRDI_IDX:
+ ubifs_assert(c, !list_empty(&lprops->list));
+ list_del(&lprops->list);
+ break;
+ default:
+ ubifs_assert(c, 0);
+ }
+
+ c->in_a_category_cnt -= 1;
+ ubifs_assert(c, c->in_a_category_cnt >= 0);
+}
+
+/**
+ * ubifs_replace_cat - replace lprops in a category list or heap.
+ * @c: UBIFS file-system description object
+ * @old_lprops: LEB properties to replace
+ * @new_lprops: LEB properties with which to replace
+ *
+ * During commit it is sometimes necessary to copy a pnode (see dirty_cow_pnode)
+ * and the lprops that the pnode contains. When that happens, references in
+ * category lists and heaps must be replaced. This function does that.
+ */
+void ubifs_replace_cat(struct ubifs_info *c, struct ubifs_lprops *old_lprops,
+ struct ubifs_lprops *new_lprops)
+{
+ int cat;
+
+ cat = new_lprops->flags & LPROPS_CAT_MASK;
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ lpt_heap_replace(c, new_lprops, cat);
+ break;
+ case LPROPS_UNCAT:
+ case LPROPS_EMPTY:
+ case LPROPS_FREEABLE:
+ case LPROPS_FRDI_IDX:
+ list_replace(&old_lprops->list, &new_lprops->list);
+ break;
+ default:
+ ubifs_assert(c, 0);
+ }
+}
+
+/**
+ * ubifs_ensure_cat - ensure LEB properties are categorized.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties
+ *
+ * A LEB may have fallen off of the bottom of a heap, and ended up as
+ * un-categorized even though it has enough space for us now. If that is the
+ * case this function will put the LEB back onto a heap.
+ */
+void ubifs_ensure_cat(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ int cat = lprops->flags & LPROPS_CAT_MASK;
+
+ if (cat != LPROPS_UNCAT)
+ return;
+ cat = ubifs_categorize_lprops(c, lprops);
+ if (cat == LPROPS_UNCAT)
+ return;
+ ubifs_remove_from_cat(c, lprops, LPROPS_UNCAT);
+ ubifs_add_to_cat(c, lprops, cat);
+}
+
+/**
+ * ubifs_categorize_lprops - categorize LEB properties.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to categorize
+ *
+ * LEB properties are categorized to enable fast find operations. This function
+ * returns the LEB category to which the LEB properties belong. Note however
+ * that if the LEB category is stored as a heap and the heap is full, the
+ * LEB properties may have their category changed to %LPROPS_UNCAT.
+ */
+int ubifs_categorize_lprops(const struct ubifs_info *c,
+ const struct ubifs_lprops *lprops)
+{
+ if (lprops->flags & LPROPS_TAKEN)
+ return LPROPS_UNCAT;
+
+ if (lprops->free == c->leb_size) {
+ ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+ return LPROPS_EMPTY;
+ }
+
+ if (lprops->free + lprops->dirty == c->leb_size) {
+ if (lprops->flags & LPROPS_INDEX)
+ return LPROPS_FRDI_IDX;
+ else
+ return LPROPS_FREEABLE;
+ }
+
+ if (lprops->flags & LPROPS_INDEX) {
+ if (lprops->dirty + lprops->free >= c->min_idx_node_sz)
+ return LPROPS_DIRTY_IDX;
+ } else {
+ if (lprops->dirty >= c->dead_wm &&
+ lprops->dirty > lprops->free)
+ return LPROPS_DIRTY;
+ if (lprops->free > 0)
+ return LPROPS_FREE;
+ }
+
+ return LPROPS_UNCAT;
+}
+
+/**
+ * change_category - change LEB properties category.
+ * @c: UBIFS file-system description object
+ * @lprops: LEB properties to re-categorize
+ *
+ * LEB properties are categorized to enable fast find operations. When the LEB
+ * properties change they must be re-categorized.
+ */
+static void change_category(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ int old_cat = lprops->flags & LPROPS_CAT_MASK;
+ int new_cat = ubifs_categorize_lprops(c, lprops);
+
+ if (old_cat == new_cat) {
+ struct ubifs_lpt_heap *heap;
+
+ /* lprops on a heap now must be moved up or down */
+ if (new_cat < 1 || new_cat > LPROPS_HEAP_CNT)
+ return; /* Not on a heap */
+ heap = &c->lpt_heap[new_cat - 1];
+ adjust_lpt_heap(c, heap, lprops, lprops->hpos, new_cat);
+ } else {
+ ubifs_remove_from_cat(c, lprops, old_cat);
+ ubifs_add_to_cat(c, lprops, new_cat);
+ }
+}
+
+/**
+ * ubifs_calc_dark - calculate LEB dark space size.
+ * @c: the UBIFS file-system description object
+ * @spc: amount of free and dirty space in the LEB
+ *
+ * This function calculates and returns amount of dark space in an LEB which
+ * has @spc bytes of free and dirty space.
+ *
+ * UBIFS is trying to account the space which might not be usable, and this
+ * space is called "dark space". For example, if an LEB has only %512 free
+ * bytes, it is dark space, because it cannot fit a large data node.
+ */
+int ubifs_calc_dark(const struct ubifs_info *c, int spc)
+{
+ ubifs_assert(c, !(spc & 7));
+
+ if (spc < c->dark_wm)
+ return spc;
+
+ /*
+ * If we have slightly more space then the dark space watermark, we can
+ * anyway safely assume it we'll be able to write a node of the
+ * smallest size there.
+ */
+ if (spc - c->dark_wm < MIN_WRITE_SZ)
+ return spc - MIN_WRITE_SZ;
+
+ return c->dark_wm;
+}
+
+/**
+ * is_lprops_dirty - determine if LEB properties are dirty.
+ * @c: the UBIFS file-system description object
+ * @lprops: LEB properties to test
+ */
+static int is_lprops_dirty(struct ubifs_info *c, struct ubifs_lprops *lprops)
+{
+ struct ubifs_pnode *pnode;
+ int pos;
+
+ pos = (lprops->lnum - c->main_first) & (UBIFS_LPT_FANOUT - 1);
+ pnode = (struct ubifs_pnode *)container_of(lprops - pos,
+ struct ubifs_pnode,
+ lprops[0]);
+ return !test_bit(COW_CNODE, &pnode->flags) &&
+ test_bit(DIRTY_CNODE, &pnode->flags);
+}
+
+/**
+ * ubifs_change_lp - change LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lp: LEB properties to change
+ * @free: new free space amount
+ * @dirty: new dirty space amount
+ * @flags: new flags
+ * @idx_gc_cnt: change to the count of @idx_gc list
+ *
+ * This function changes LEB properties (@free, @dirty or @flag). However, the
+ * property which has the %LPROPS_NC value is not changed. Returns a pointer to
+ * the updated LEB properties on success and a negative error code on failure.
+ *
+ * Note, the LEB properties may have had to be copied (due to COW) and
+ * consequently the pointer returned may not be the same as the pointer
+ * passed.
+ */
+const struct ubifs_lprops *ubifs_change_lp(struct ubifs_info *c,
+ const struct ubifs_lprops *lp,
+ int free, int dirty, int flags,
+ int idx_gc_cnt)
+{
+ /*
+ * This is the only function that is allowed to change lprops, so we
+ * discard the "const" qualifier.
+ */
+ struct ubifs_lprops *lprops = (struct ubifs_lprops *)lp;
+
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d",
+ lprops->lnum, free, dirty, flags);
+
+ ubifs_assert(c, mutex_is_locked(&c->lp_mutex));
+ ubifs_assert(c, c->lst.empty_lebs >= 0 &&
+ c->lst.empty_lebs <= c->main_lebs);
+ ubifs_assert(c, c->freeable_cnt >= 0);
+ ubifs_assert(c, c->freeable_cnt <= c->main_lebs);
+ ubifs_assert(c, c->lst.taken_empty_lebs >= 0);
+ ubifs_assert(c, c->lst.taken_empty_lebs <= c->lst.empty_lebs);
+ ubifs_assert(c, !(c->lst.total_free & 7) && !(c->lst.total_dirty & 7));
+ ubifs_assert(c, !(c->lst.total_dead & 7) && !(c->lst.total_dark & 7));
+ ubifs_assert(c, !(c->lst.total_used & 7));
+ ubifs_assert(c, free == LPROPS_NC || free >= 0);
+ ubifs_assert(c, dirty == LPROPS_NC || dirty >= 0);
+
+ if (!is_lprops_dirty(c, lprops)) {
+ lprops = ubifs_lpt_lookup_dirty(c, lprops->lnum);
+ if (IS_ERR(lprops))
+ return lprops;
+ } else
+ ubifs_assert(c, lprops == ubifs_lpt_lookup_dirty(c, lprops->lnum));
+
+ ubifs_assert(c, !(lprops->free & 7) && !(lprops->dirty & 7));
+
+ spin_lock(&c->space_lock);
+ if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size)
+ c->lst.taken_empty_lebs -= 1;
+
+ if (!(lprops->flags & LPROPS_INDEX)) {
+ int old_spc;
+
+ old_spc = lprops->free + lprops->dirty;
+ if (old_spc < c->dead_wm)
+ c->lst.total_dead -= old_spc;
+ else
+ c->lst.total_dark -= ubifs_calc_dark(c, old_spc);
+
+ c->lst.total_used -= c->leb_size - old_spc;
+ }
+
+ if (free != LPROPS_NC) {
+ free = ALIGN(free, 8);
+ c->lst.total_free += free - lprops->free;
+
+ /* Increase or decrease empty LEBs counter if needed */
+ if (free == c->leb_size) {
+ if (lprops->free != c->leb_size)
+ c->lst.empty_lebs += 1;
+ } else if (lprops->free == c->leb_size)
+ c->lst.empty_lebs -= 1;
+ lprops->free = free;
+ }
+
+ if (dirty != LPROPS_NC) {
+ dirty = ALIGN(dirty, 8);
+ c->lst.total_dirty += dirty - lprops->dirty;
+ lprops->dirty = dirty;
+ }
+
+ if (flags != LPROPS_NC) {
+ /* Take care about indexing LEBs counter if needed */
+ if ((lprops->flags & LPROPS_INDEX)) {
+ if (!(flags & LPROPS_INDEX))
+ c->lst.idx_lebs -= 1;
+ } else if (flags & LPROPS_INDEX)
+ c->lst.idx_lebs += 1;
+ lprops->flags = flags;
+ }
+
+ if (!(lprops->flags & LPROPS_INDEX)) {
+ int new_spc;
+
+ new_spc = lprops->free + lprops->dirty;
+ if (new_spc < c->dead_wm)
+ c->lst.total_dead += new_spc;
+ else
+ c->lst.total_dark += ubifs_calc_dark(c, new_spc);
+
+ c->lst.total_used += c->leb_size - new_spc;
+ }
+
+ if ((lprops->flags & LPROPS_TAKEN) && lprops->free == c->leb_size)
+ c->lst.taken_empty_lebs += 1;
+
+ change_category(c, lprops);
+ c->idx_gc_cnt += idx_gc_cnt;
+ spin_unlock(&c->space_lock);
+ return lprops;
+}
+
+/**
+ * ubifs_get_lp_stats - get lprops statistics.
+ * @c: UBIFS file-system description object
+ * @lst: return statistics
+ */
+void ubifs_get_lp_stats(struct ubifs_info *c, struct ubifs_lp_stats *lst)
+{
+ spin_lock(&c->space_lock);
+ memcpy(lst, &c->lst, sizeof(struct ubifs_lp_stats));
+ spin_unlock(&c->space_lock);
+}
+
+/**
+ * ubifs_change_one_lp - change LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to change properties for
+ * @free: amount of free space
+ * @dirty: amount of dirty space
+ * @flags_set: flags to set
+ * @flags_clean: flags to clean
+ * @idx_gc_cnt: change to the count of idx_gc list
+ *
+ * This function changes properties of LEB @lnum. It is a helper wrapper over
+ * 'ubifs_change_lp()' which hides lprops get/release. The arguments are the
+ * same as in case of 'ubifs_change_lp()'. Returns zero in case of success and
+ * a negative error code in case of failure.
+ */
+int ubifs_change_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean, int idx_gc_cnt)
+{
+ int err = 0, flags;
+ const struct ubifs_lprops *lp;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ flags = (lp->flags | flags_set) & ~flags_clean;
+ lp = ubifs_change_lp(c, lp, free, dirty, flags, idx_gc_cnt);
+ if (IS_ERR(lp))
+ err = PTR_ERR(lp);
+
+out:
+ ubifs_release_lprops(c);
+ if (err)
+ ubifs_err(c, "cannot change properties of LEB %d, error %d",
+ lnum, err);
+ return err;
+}
+
+/**
+ * ubifs_update_one_lp - update LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to change properties for
+ * @free: amount of free space
+ * @dirty: amount of dirty space to add
+ * @flags_set: flags to set
+ * @flags_clean: flags to clean
+ *
+ * This function is the same as 'ubifs_change_one_lp()' but @dirty is added to
+ * current dirty space, not substitutes it.
+ */
+int ubifs_update_one_lp(struct ubifs_info *c, int lnum, int free, int dirty,
+ int flags_set, int flags_clean)
+{
+ int err = 0, flags;
+ const struct ubifs_lprops *lp;
+
+ ubifs_get_lprops(c);
+
+ lp = ubifs_lpt_lookup_dirty(c, lnum);
+ if (IS_ERR(lp)) {
+ err = PTR_ERR(lp);
+ goto out;
+ }
+
+ flags = (lp->flags | flags_set) & ~flags_clean;
+ lp = ubifs_change_lp(c, lp, free, lp->dirty + dirty, flags, 0);
+ if (IS_ERR(lp))
+ err = PTR_ERR(lp);
+
+out:
+ ubifs_release_lprops(c);
+ if (err)
+ ubifs_err(c, "cannot update properties of LEB %d, error %d",
+ lnum, err);
+ return err;
+}
+
+/**
+ * ubifs_read_one_lp - read LEB properties.
+ * @c: the UBIFS file-system description object
+ * @lnum: LEB to read properties for
+ * @lp: where to store read properties
+ *
+ * This helper function reads properties of a LEB @lnum and stores them in @lp.
+ * Returns zero in case of success and a negative error code in case of
+ * failure.
+ */
+int ubifs_read_one_lp(struct ubifs_info *c, int lnum, struct ubifs_lprops *lp)
+{
+ int err = 0;
+ const struct ubifs_lprops *lpp;
+
+ ubifs_get_lprops(c);
+
+ lpp = ubifs_lpt_lookup(c, lnum);
+ if (IS_ERR(lpp)) {
+ err = PTR_ERR(lpp);
+ ubifs_err(c, "cannot read properties of LEB %d, error %d",
+ lnum, err);
+ goto out;
+ }
+
+ memcpy(lp, lpp, sizeof(struct ubifs_lprops));
+
+out:
+ ubifs_release_lprops(c);
+ return err;
+}
+
+/**
+ * ubifs_fast_find_free - try to find a LEB with free space quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a LEB with free space or %NULL if
+ * the function is unable to find a LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_free(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+ struct ubifs_lpt_heap *heap;
+
+ ubifs_assert(c, mutex_is_locked(&c->lp_mutex));
+
+ heap = &c->lpt_heap[LPROPS_FREE - 1];
+ if (heap->cnt == 0)
+ return NULL;
+
+ lprops = heap->arr[0];
+ ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_empty - try to find an empty LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for an empty LEB or %NULL if the
+ * function is unable to find an empty LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_empty(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(c, mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->empty_list))
+ return NULL;
+
+ lprops = list_entry(c->empty_list.next, struct ubifs_lprops, list);
+ ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+ ubifs_assert(c, lprops->free == c->leb_size);
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_freeable - try to find a freeable LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a freeable LEB or %NULL if the
+ * function is unable to find a freeable LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_freeable(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(c, mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->freeable_list))
+ return NULL;
+
+ lprops = list_entry(c->freeable_list.next, struct ubifs_lprops, list);
+ ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(c, !(lprops->flags & LPROPS_INDEX));
+ ubifs_assert(c, lprops->free + lprops->dirty == c->leb_size);
+ ubifs_assert(c, c->freeable_cnt > 0);
+ return lprops;
+}
+
+/**
+ * ubifs_fast_find_frdi_idx - try to find a freeable index LEB quickly.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns LEB properties for a freeable index LEB or %NULL if the
+ * function is unable to find a freeable index LEB quickly.
+ */
+const struct ubifs_lprops *ubifs_fast_find_frdi_idx(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+
+ ubifs_assert(c, mutex_is_locked(&c->lp_mutex));
+
+ if (list_empty(&c->frdi_idx_list))
+ return NULL;
+
+ lprops = list_entry(c->frdi_idx_list.next, struct ubifs_lprops, list);
+ ubifs_assert(c, !(lprops->flags & LPROPS_TAKEN));
+ ubifs_assert(c, (lprops->flags & LPROPS_INDEX));
+ ubifs_assert(c, lprops->free + lprops->dirty == c->leb_size);
+ return lprops;
+}
+
+/*
+ * Everything below is related to debugging.
+ */
+
+/**
+ * dbg_check_cats - check category heaps and lists.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int dbg_check_cats(struct ubifs_info *c)
+{
+ struct ubifs_lprops *lprops;
+ struct list_head *pos;
+ int i, cat;
+
+ if (!dbg_is_chk_gen(c) && !dbg_is_chk_lprops(c))
+ return 0;
+
+ list_for_each_entry(lprops, &c->empty_list, list) {
+ if (lprops->free != c->leb_size) {
+ ubifs_err(c, "non-empty LEB %d on empty list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ if (lprops->flags & LPROPS_TAKEN) {
+ ubifs_err(c, "taken LEB %d on empty list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ }
+
+ i = 0;
+ list_for_each_entry(lprops, &c->freeable_list, list) {
+ if (lprops->free + lprops->dirty != c->leb_size) {
+ ubifs_err(c, "non-freeable LEB %d on freeable list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ if (lprops->flags & LPROPS_TAKEN) {
+ ubifs_err(c, "taken LEB %d on freeable list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ i += 1;
+ }
+ if (i != c->freeable_cnt) {
+ ubifs_err(c, "freeable list count %d expected %d", i,
+ c->freeable_cnt);
+ return -EINVAL;
+ }
+
+ i = 0;
+ list_for_each(pos, &c->idx_gc)
+ i += 1;
+ if (i != c->idx_gc_cnt) {
+ ubifs_err(c, "idx_gc list count %d expected %d", i,
+ c->idx_gc_cnt);
+ return -EINVAL;
+ }
+
+ list_for_each_entry(lprops, &c->frdi_idx_list, list) {
+ if (lprops->free + lprops->dirty != c->leb_size) {
+ ubifs_err(c, "non-freeable LEB %d on frdi_idx list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ if (lprops->flags & LPROPS_TAKEN) {
+ ubifs_err(c, "taken LEB %d on frdi_idx list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ if (!(lprops->flags & LPROPS_INDEX)) {
+ ubifs_err(c, "non-index LEB %d on frdi_idx list (free %d dirty %d flags %d)",
+ lprops->lnum, lprops->free, lprops->dirty,
+ lprops->flags);
+ return -EINVAL;
+ }
+ }
+
+ for (cat = 1; cat <= LPROPS_HEAP_CNT; cat++) {
+ struct ubifs_lpt_heap *heap = &c->lpt_heap[cat - 1];
+
+ for (i = 0; i < heap->cnt; i++) {
+ lprops = heap->arr[i];
+ if (!lprops) {
+ ubifs_err(c, "null ptr in LPT heap cat %d", cat);
+ return -EINVAL;
+ }
+ if (lprops->hpos != i) {
+ ubifs_err(c, "bad ptr in LPT heap cat %d", cat);
+ return -EINVAL;
+ }
+ if (lprops->flags & LPROPS_TAKEN) {
+ ubifs_err(c, "taken LEB in LPT heap cat %d", cat);
+ return -EINVAL;
+ }
+ }
+ }
+
+ return 0;
+}
+
+void dbg_check_heap(struct ubifs_info *c, struct ubifs_lpt_heap *heap, int cat,
+ int add_pos)
+{
+ int i = 0, j, err = 0;
+
+ if (!dbg_is_chk_gen(c) && !dbg_is_chk_lprops(c))
+ return;
+
+ for (i = 0; i < heap->cnt; i++) {
+ struct ubifs_lprops *lprops = heap->arr[i];
+ struct ubifs_lprops *lp;
+
+ if (i != add_pos)
+ if ((lprops->flags & LPROPS_CAT_MASK) != cat) {
+ err = 1;
+ goto out;
+ }
+ if (lprops->hpos != i) {
+ err = 2;
+ goto out;
+ }
+ lp = ubifs_lpt_lookup(c, lprops->lnum);
+ if (IS_ERR(lp)) {
+ err = 3;
+ goto out;
+ }
+ if (lprops != lp) {
+ ubifs_err(c, "lprops %zx lp %zx lprops->lnum %d lp->lnum %d",
+ (size_t)lprops, (size_t)lp, lprops->lnum,
+ lp->lnum);
+ err = 4;
+ goto out;
+ }
+ for (j = 0; j < i; j++) {
+ lp = heap->arr[j];
+ if (lp == lprops) {
+ err = 5;
+ goto out;
+ }
+ if (lp->lnum == lprops->lnum) {
+ err = 6;
+ goto out;
+ }
+ }
+ }
+out:
+ if (err) {
+ ubifs_err(c, "failed cat %d hpos %d err %d", cat, i, err);
+ dump_stack();
+ ubifs_dump_heap(c, heap, cat);
+ }
+}
+
+/**
+ * scan_check_cb - scan callback.
+ * @c: the UBIFS file-system description object
+ * @lp: LEB properties to scan
+ * @in_tree: whether the LEB properties are in main memory
+ * @lst: lprops statistics to update
+ *
+ * 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_check_cb(struct ubifs_info *c,
+ const struct ubifs_lprops *lp, int in_tree,
+ struct ubifs_lp_stats *lst)
+{
+ struct ubifs_scan_leb *sleb;
+ struct ubifs_scan_node *snod;
+ int cat, lnum = lp->lnum, is_idx = 0, used = 0, free, dirty, ret;
+ void *buf = NULL;
+
+ cat = lp->flags & LPROPS_CAT_MASK;
+ if (cat != LPROPS_UNCAT) {
+ cat = ubifs_categorize_lprops(c, lp);
+ if (cat != (lp->flags & LPROPS_CAT_MASK)) {
+ ubifs_err(c, "bad LEB category %d expected %d",
+ (lp->flags & LPROPS_CAT_MASK), cat);
+ return -EINVAL;
+ }
+ }
+
+ /* Check lp is on its category list (if it has one) */
+ if (in_tree) {
+ struct list_head *list = NULL;
+
+ switch (cat) {
+ case LPROPS_EMPTY:
+ list = &c->empty_list;
+ break;
+ case LPROPS_FREEABLE:
+ list = &c->freeable_list;
+ break;
+ case LPROPS_FRDI_IDX:
+ list = &c->frdi_idx_list;
+ break;
+ case LPROPS_UNCAT:
+ list = &c->uncat_list;
+ break;
+ }
+ if (list) {
+ struct ubifs_lprops *lprops;
+ int found = 0;
+
+ list_for_each_entry(lprops, list, list) {
+ if (lprops == lp) {
+ found = 1;
+ break;
+ }
+ }
+ if (!found) {
+ ubifs_err(c, "bad LPT list (category %d)", cat);
+ return -EINVAL;
+ }
+ }
+ }
+
+ /* Check lp is on its category heap (if it has one) */
+ if (in_tree && cat > 0 && cat <= LPROPS_HEAP_CNT) {
+ struct ubifs_lpt_heap *heap = &c->lpt_heap[cat - 1];
+
+ if ((lp->hpos != -1 && heap->arr[lp->hpos]->lnum != lnum) ||
+ lp != heap->arr[lp->hpos]) {
+ ubifs_err(c, "bad LPT heap (category %d)", cat);
+ return -EINVAL;
+ }
+ }
+
+ /*
+ * After an unclean unmount, empty and freeable LEBs
+ * may contain garbage - do not scan them.
+ */
+ if (lp->free == c->leb_size) {
+ lst->empty_lebs += 1;
+ lst->total_free += c->leb_size;
+ lst->total_dark += ubifs_calc_dark(c, c->leb_size);
+ return LPT_SCAN_CONTINUE;
+ }
+ if (lp->free + lp->dirty == c->leb_size &&
+ !(lp->flags & LPROPS_INDEX)) {
+ lst->total_free += lp->free;
+ lst->total_dirty += lp->dirty;
+ lst->total_dark += ubifs_calc_dark(c, c->leb_size);
+ return LPT_SCAN_CONTINUE;
+ }
+
+ buf = __vmalloc(c->leb_size, GFP_NOFS);
+ if (!buf)
+ return -ENOMEM;
+
+ sleb = ubifs_scan(c, lnum, 0, buf, 0);
+ if (IS_ERR(sleb)) {
+ ret = PTR_ERR(sleb);
+ if (ret == -EUCLEAN) {
+ ubifs_dump_lprops(c);
+ ubifs_dump_budg(c, &c->bi);
+ }
+ goto out;
+ }
+
+ is_idx = -1;
+ list_for_each_entry(snod, &sleb->nodes, list) {
+ int found, level = 0;
+
+ cond_resched();
+
+ if (is_idx == -1)
+ is_idx = (snod->type == UBIFS_IDX_NODE) ? 1 : 0;
+
+ if (is_idx && snod->type != UBIFS_IDX_NODE) {
+ ubifs_err(c, "indexing node in data LEB %d:%d",
+ lnum, snod->offs);
+ goto out_destroy;
+ }
+
+ if (snod->type == UBIFS_IDX_NODE) {
+ struct ubifs_idx_node *idx = snod->node;
+
+ key_read(c, ubifs_idx_key(c, idx), &snod->key);
+ level = le16_to_cpu(idx->level);
+ }
+
+ found = ubifs_tnc_has_node(c, &snod->key, level, lnum,
+ snod->offs, is_idx);
+ if (found) {
+ if (found < 0)
+ goto out_destroy;
+ used += ALIGN(snod->len, 8);
+ }
+ }
+
+ free = c->leb_size - sleb->endpt;
+ dirty = sleb->endpt - used;
+
+ if (free > c->leb_size || free < 0 || dirty > c->leb_size ||
+ dirty < 0) {
+ ubifs_err(c, "bad calculated accounting for LEB %d: free %d, dirty %d",
+ lnum, free, dirty);
+ goto out_destroy;
+ }
+
+ if (lp->free + lp->dirty == c->leb_size &&
+ free + dirty == c->leb_size)
+ if ((is_idx && !(lp->flags & LPROPS_INDEX)) ||
+ (!is_idx && free == c->leb_size) ||
+ lp->free == c->leb_size) {
+ /*
+ * Empty or freeable LEBs could contain index
+ * nodes from an uncompleted commit due to an
+ * unclean unmount. Or they could be empty for
+ * the same reason. Or it may simply not have been
+ * unmapped.
+ */
+ free = lp->free;
+ dirty = lp->dirty;
+ is_idx = 0;
+ }
+
+ if (is_idx && lp->free + lp->dirty == free + dirty &&
+ lnum != c->ihead_lnum) {
+ /*
+ * After an unclean unmount, an index LEB could have a different
+ * amount of free space than the value recorded by lprops. That
+ * is because the in-the-gaps method may use free space or
+ * create free space (as a side-effect of using ubi_leb_change
+ * and not writing the whole LEB). The incorrect free space
+ * value is not a problem because the index is only ever
+ * allocated empty LEBs, so there will never be an attempt to
+ * write to the free space at the end of an index LEB - except
+ * by the in-the-gaps method for which it is not a problem.
+ */
+ free = lp->free;
+ dirty = lp->dirty;
+ }
+
+ if (lp->free != free || lp->dirty != dirty)
+ goto out_print;
+
+ if (is_idx && !(lp->flags & LPROPS_INDEX)) {
+ if (free == c->leb_size)
+ /* Free but not unmapped LEB, it's fine */
+ is_idx = 0;
+ else {
+ ubifs_err(c, "indexing node without indexing flag");
+ goto out_print;
+ }
+ }
+
+ if (!is_idx && (lp->flags & LPROPS_INDEX)) {
+ ubifs_err(c, "data node with indexing flag");
+ goto out_print;
+ }
+
+ if (free == c->leb_size)
+ lst->empty_lebs += 1;
+
+ if (is_idx)
+ lst->idx_lebs += 1;
+
+ if (!(lp->flags & LPROPS_INDEX))
+ lst->total_used += c->leb_size - free - dirty;
+ lst->total_free += free;
+ lst->total_dirty += dirty;
+
+ if (!(lp->flags & LPROPS_INDEX)) {
+ int spc = free + dirty;
+
+ if (spc < c->dead_wm)
+ lst->total_dead += spc;
+ else
+ lst->total_dark += ubifs_calc_dark(c, spc);
+ }
+
+ ubifs_scan_destroy(sleb);
+ vfree(buf);
+ return LPT_SCAN_CONTINUE;
+
+out_print:
+ ubifs_err(c, "bad accounting of LEB %d: free %d, dirty %d flags %#x, should be free %d, dirty %d",
+ lnum, lp->free, lp->dirty, lp->flags, free, dirty);
+ ubifs_dump_leb(c, lnum);
+out_destroy:
+ ubifs_scan_destroy(sleb);
+ ret = -EINVAL;
+out:
+ vfree(buf);
+ return ret;
+}
+
+/**
+ * dbg_check_lprops - check all LEB properties.
+ * @c: UBIFS file-system description object
+ *
+ * This function checks all LEB properties and makes sure they are all correct.
+ * It returns zero if everything is fine, %-EINVAL if there is an inconsistency
+ * and other negative error codes in case of other errors. This function is
+ * called while the file system is locked (because of commit start), so no
+ * additional locking is required. Note that locking the LPT mutex would cause
+ * a circular lock dependency with the TNC mutex.
+ */
+int dbg_check_lprops(struct ubifs_info *c)
+{
+ int i, err;
+ struct ubifs_lp_stats lst;
+
+ if (!dbg_is_chk_lprops(c))
+ return 0;
+
+ /*
+ * As we are going to scan the media, the write buffers have to be
+ * synchronized.
+ */
+ for (i = 0; i < c->jhead_cnt; i++) {
+ err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
+ if (err)
+ return err;
+ }
+
+ memset(&lst, 0, sizeof(struct ubifs_lp_stats));
+ err = ubifs_lpt_scan_nolock(c, c->main_first, c->leb_cnt - 1,
+ (ubifs_lpt_scan_callback)scan_check_cb,
+ &lst);
+ if (err && err != -ENOSPC)
+ goto out;
+
+ if (lst.empty_lebs != c->lst.empty_lebs ||
+ lst.idx_lebs != c->lst.idx_lebs ||
+ lst.total_free != c->lst.total_free ||
+ lst.total_dirty != c->lst.total_dirty ||
+ lst.total_used != c->lst.total_used) {
+ ubifs_err(c, "bad overall accounting");
+ ubifs_err(c, "calculated: empty_lebs %d, idx_lebs %d, total_free %lld, total_dirty %lld, total_used %lld",
+ lst.empty_lebs, lst.idx_lebs, lst.total_free,
+ lst.total_dirty, lst.total_used);
+ ubifs_err(c, "read from lprops: empty_lebs %d, idx_lebs %d, total_free %lld, total_dirty %lld, total_used %lld",
+ c->lst.empty_lebs, c->lst.idx_lebs, c->lst.total_free,
+ c->lst.total_dirty, c->lst.total_used);
+ err = -EINVAL;
+ goto out;
+ }
+
+ if (lst.total_dead != c->lst.total_dead ||
+ lst.total_dark != c->lst.total_dark) {
+ ubifs_err(c, "bad dead/dark space accounting");
+ ubifs_err(c, "calculated: total_dead %lld, total_dark %lld",
+ lst.total_dead, lst.total_dark);
+ ubifs_err(c, "read from lprops: total_dead %lld, total_dark %lld",
+ c->lst.total_dead, c->lst.total_dark);
+ err = -EINVAL;
+ goto out;
+ }
+
+ err = dbg_check_cats(c);
+out:
+ return err;
+}
new file mode 100644
@@ -0,0 +1,2451 @@
+// 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 LEB properties tree (LPT) area. The LPT area
+ * contains the LEB properties tree, a table of LPT area eraseblocks (ltab), and
+ * (for the "big" model) a table of saved LEB numbers (lsave). The LPT area sits
+ * between the log and the orphan area.
+ *
+ * The LPT area is like a miniature self-contained file system. It is required
+ * that it never runs out of space, is fast to access and update, and scales
+ * logarithmically. The LEB properties tree is implemented as a wandering tree
+ * much like the TNC, and the LPT area has its own garbage collection.
+ *
+ * The LPT has two slightly different forms called the "small model" and the
+ * "big model". The small model is used when the entire LEB properties table
+ * can be written into a single eraseblock. In that case, garbage collection
+ * consists of just writing the whole table, which therefore makes all other
+ * eraseblocks reusable. In the case of the big model, dirty eraseblocks are
+ * selected for garbage collection, which consists of marking the clean nodes in
+ * that LEB as dirty, and then only the dirty nodes are written out. Also, in
+ * the case of the big model, a table of LEB numbers is saved so that the entire
+ * LPT does not to be scanned looking for empty eraseblocks when UBIFS is first
+ * mounted.
+ */
+
+#include "ubifs.h"
+#include <linux/crc16.h>
+#include <linux/math64.h>
+#include <linux/slab.h>
+
+/**
+ * do_calc_lpt_geom - calculate sizes for the LPT area.
+ * @c: the UBIFS file-system description object
+ *
+ * Calculate the sizes of LPT bit fields, nodes, and tree, based on the
+ * properties of the flash and whether LPT is "big" (c->big_lpt).
+ */
+static void do_calc_lpt_geom(struct ubifs_info *c)
+{
+ int i, n, bits, per_leb_wastage, max_pnode_cnt;
+ long long sz, tot_wastage;
+
+ n = c->main_lebs + c->max_leb_cnt - c->leb_cnt;
+ max_pnode_cnt = DIV_ROUND_UP(n, UBIFS_LPT_FANOUT);
+
+ c->lpt_hght = 1;
+ n = UBIFS_LPT_FANOUT;
+ while (n < max_pnode_cnt) {
+ c->lpt_hght += 1;
+ n <<= UBIFS_LPT_FANOUT_SHIFT;
+ }
+
+ c->pnode_cnt = DIV_ROUND_UP(c->main_lebs, UBIFS_LPT_FANOUT);
+
+ n = DIV_ROUND_UP(c->pnode_cnt, UBIFS_LPT_FANOUT);
+ c->nnode_cnt = n;
+ for (i = 1; i < c->lpt_hght; i++) {
+ n = DIV_ROUND_UP(n, UBIFS_LPT_FANOUT);
+ c->nnode_cnt += n;
+ }
+
+ c->space_bits = fls(c->leb_size) - 3;
+ c->lpt_lnum_bits = fls(c->lpt_lebs);
+ c->lpt_offs_bits = fls(c->leb_size - 1);
+ c->lpt_spc_bits = fls(c->leb_size);
+
+ n = DIV_ROUND_UP(c->max_leb_cnt, UBIFS_LPT_FANOUT);
+ c->pcnt_bits = fls(n - 1);
+
+ c->lnum_bits = fls(c->max_leb_cnt - 1);
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ (c->big_lpt ? c->pcnt_bits : 0) +
+ (c->space_bits * 2 + 1) * UBIFS_LPT_FANOUT;
+ c->pnode_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ (c->big_lpt ? c->pcnt_bits : 0) +
+ (c->lpt_lnum_bits + c->lpt_offs_bits) * UBIFS_LPT_FANOUT;
+ c->nnode_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ c->lpt_lebs * c->lpt_spc_bits * 2;
+ c->ltab_sz = (bits + 7) / 8;
+
+ bits = UBIFS_LPT_CRC_BITS + UBIFS_LPT_TYPE_BITS +
+ c->lnum_bits * c->lsave_cnt;
+ c->lsave_sz = (bits + 7) / 8;
+
+ /* Calculate the minimum LPT size */
+ c->lpt_sz = (long long)c->pnode_cnt * c->pnode_sz;
+ c->lpt_sz += (long long)c->nnode_cnt * c->nnode_sz;
+ c->lpt_sz += c->ltab_sz;
+ if (c->big_lpt)
+ c->lpt_sz += c->lsave_sz;
+
+ /* Add wastage */
+ sz = c->lpt_sz;
+ per_leb_wastage = max_t(int, c->pnode_sz, c->nnode_sz);
+ sz += per_leb_wastage;
+ tot_wastage = per_leb_wastage;
+ while (sz > c->leb_size) {
+ sz += per_leb_wastage;
+ sz -= c->leb_size;
+ tot_wastage += per_leb_wastage;
+ }
+ tot_wastage += ALIGN(sz, c->min_io_size) - sz;
+ c->lpt_sz += tot_wastage;
+}
+
+/**
+ * ubifs_calc_lpt_geom - calculate and check sizes for the LPT area.
+ * @c: the UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_calc_lpt_geom(struct ubifs_info *c)
+{
+ int lebs_needed;
+ long long sz;
+
+ do_calc_lpt_geom(c);
+
+ /* Verify that lpt_lebs is big enough */
+ sz = c->lpt_sz * 2; /* Must have at least 2 times the size */
+ lebs_needed = div_u64(sz + c->leb_size - 1, c->leb_size);
+ if (lebs_needed > c->lpt_lebs) {
+ ubifs_err(c, "too few LPT LEBs");
+ return -EINVAL;
+ }
+
+ /* Verify that ltab fits in a single LEB (since ltab is a single node */
+ if (c->ltab_sz > c->leb_size) {
+ ubifs_err(c, "LPT ltab too big");
+ return -EINVAL;
+ }
+
+ c->check_lpt_free = c->big_lpt;
+ return 0;
+}
+
+/**
+ * calc_dflt_lpt_geom - calculate default LPT geometry.
+ * @c: the UBIFS file-system description object
+ * @main_lebs: number of main area LEBs is passed and returned here
+ * @big_lpt: whether the LPT area is "big" is returned here
+ *
+ * The size of the LPT area depends on parameters that themselves are dependent
+ * on the size of the LPT area. This function, successively recalculates the LPT
+ * area geometry until the parameters and resultant geometry are consistent.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int calc_dflt_lpt_geom(struct ubifs_info *c, int *main_lebs,
+ int *big_lpt)
+{
+ int i, lebs_needed;
+ long long sz;
+
+ /* Start by assuming the minimum number of LPT LEBs */
+ c->lpt_lebs = UBIFS_MIN_LPT_LEBS;
+ c->main_lebs = *main_lebs - c->lpt_lebs;
+ if (c->main_lebs <= 0)
+ return -EINVAL;
+
+ /* And assume we will use the small LPT model */
+ c->big_lpt = 0;
+
+ /*
+ * Calculate the geometry based on assumptions above and then see if it
+ * makes sense
+ */
+ do_calc_lpt_geom(c);
+
+ /* Small LPT model must have lpt_sz < leb_size */
+ if (c->lpt_sz > c->leb_size) {
+ /* Nope, so try again using big LPT model */
+ c->big_lpt = 1;
+ do_calc_lpt_geom(c);
+ }
+
+ /* Now check there are enough LPT LEBs */
+ for (i = 0; i < 64 ; i++) {
+ sz = c->lpt_sz * 4; /* Allow 4 times the size */
+ lebs_needed = div_u64(sz + c->leb_size - 1, c->leb_size);
+ if (lebs_needed > c->lpt_lebs) {
+ /* Not enough LPT LEBs so try again with more */
+ c->lpt_lebs = lebs_needed;
+ c->main_lebs = *main_lebs - c->lpt_lebs;
+ if (c->main_lebs <= 0)
+ return -EINVAL;
+ do_calc_lpt_geom(c);
+ continue;
+ }
+ if (c->ltab_sz > c->leb_size) {
+ ubifs_err(c, "LPT ltab too big");
+ return -EINVAL;
+ }
+ *main_lebs = c->main_lebs;
+ *big_lpt = c->big_lpt;
+ return 0;
+ }
+ return -EINVAL;
+}
+
+/**
+ * pack_bits - pack bit fields end-to-end.
+ * @c: UBIFS file-system description object
+ * @addr: address at which to pack (passed and next address returned)
+ * @pos: bit position at which to pack (passed and next position returned)
+ * @val: value to pack
+ * @nrbits: number of bits of value to pack (1-32)
+ */
+static void pack_bits(const struct ubifs_info *c, uint8_t **addr, int *pos, uint32_t val, int nrbits)
+{
+ uint8_t *p = *addr;
+ int b = *pos;
+
+ ubifs_assert(c, nrbits > 0);
+ ubifs_assert(c, nrbits <= 32);
+ ubifs_assert(c, *pos >= 0);
+ ubifs_assert(c, *pos < 8);
+ ubifs_assert(c, (val >> nrbits) == 0 || nrbits == 32);
+ if (b) {
+ *p |= ((uint8_t)val) << b;
+ nrbits += b;
+ if (nrbits > 8) {
+ *++p = (uint8_t)(val >>= (8 - b));
+ if (nrbits > 16) {
+ *++p = (uint8_t)(val >>= 8);
+ if (nrbits > 24) {
+ *++p = (uint8_t)(val >>= 8);
+ if (nrbits > 32)
+ *++p = (uint8_t)(val >>= 8);
+ }
+ }
+ }
+ } else {
+ *p = (uint8_t)val;
+ if (nrbits > 8) {
+ *++p = (uint8_t)(val >>= 8);
+ if (nrbits > 16) {
+ *++p = (uint8_t)(val >>= 8);
+ if (nrbits > 24)
+ *++p = (uint8_t)(val >>= 8);
+ }
+ }
+ }
+ b = nrbits & 7;
+ if (b == 0)
+ p++;
+ *addr = p;
+ *pos = b;
+}
+
+/**
+ * ubifs_unpack_bits - unpack bit fields.
+ * @c: UBIFS file-system description object
+ * @addr: address at which to unpack (passed and next address returned)
+ * @pos: bit position at which to unpack (passed and next position returned)
+ * @nrbits: number of bits of value to unpack (1-32)
+ *
+ * This functions returns the value unpacked.
+ */
+uint32_t ubifs_unpack_bits(const struct ubifs_info *c, uint8_t **addr, int *pos, int nrbits)
+{
+ const int k = 32 - nrbits;
+ uint8_t *p = *addr;
+ int b = *pos;
+ uint32_t val;
+ const int bytes = (nrbits + b + 7) >> 3;
+
+ ubifs_assert(c, nrbits > 0);
+ ubifs_assert(c, nrbits <= 32);
+ ubifs_assert(c, *pos >= 0);
+ ubifs_assert(c, *pos < 8);
+ if (b) {
+ switch (bytes) {
+ case 2:
+ val = p[1];
+ break;
+ case 3:
+ val = p[1] | ((uint32_t)p[2] << 8);
+ break;
+ case 4:
+ val = p[1] | ((uint32_t)p[2] << 8) |
+ ((uint32_t)p[3] << 16);
+ break;
+ case 5:
+ val = p[1] | ((uint32_t)p[2] << 8) |
+ ((uint32_t)p[3] << 16) |
+ ((uint32_t)p[4] << 24);
+ }
+ val <<= (8 - b);
+ val |= *p >> b;
+ nrbits += b;
+ } else {
+ switch (bytes) {
+ case 1:
+ val = p[0];
+ break;
+ case 2:
+ val = p[0] | ((uint32_t)p[1] << 8);
+ break;
+ case 3:
+ val = p[0] | ((uint32_t)p[1] << 8) |
+ ((uint32_t)p[2] << 16);
+ break;
+ case 4:
+ val = p[0] | ((uint32_t)p[1] << 8) |
+ ((uint32_t)p[2] << 16) |
+ ((uint32_t)p[3] << 24);
+ break;
+ }
+ }
+ val <<= k;
+ val >>= k;
+ b = nrbits & 7;
+ p += nrbits >> 3;
+ *addr = p;
+ *pos = b;
+ ubifs_assert(c, (val >> nrbits) == 0 || nrbits - b == 32);
+ return val;
+}
+
+/**
+ * ubifs_pack_pnode - pack all the bit fields of a pnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer into which to pack
+ * @pnode: pnode to pack
+ */
+void ubifs_pack_pnode(struct ubifs_info *c, void *buf,
+ struct ubifs_pnode *pnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0;
+ uint16_t crc;
+
+ pack_bits(c, &addr, &pos, UBIFS_LPT_PNODE, UBIFS_LPT_TYPE_BITS);
+ if (c->big_lpt)
+ pack_bits(c, &addr, &pos, pnode->num, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ pack_bits(c, &addr, &pos, pnode->lprops[i].free >> 3,
+ c->space_bits);
+ pack_bits(c, &addr, &pos, pnode->lprops[i].dirty >> 3,
+ c->space_bits);
+ if (pnode->lprops[i].flags & LPROPS_INDEX)
+ pack_bits(c, &addr, &pos, 1, 1);
+ else
+ pack_bits(c, &addr, &pos, 0, 1);
+ }
+ crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ c->pnode_sz - UBIFS_LPT_CRC_BYTES);
+ addr = buf;
+ pos = 0;
+ pack_bits(c, &addr, &pos, crc, UBIFS_LPT_CRC_BITS);
+}
+
+/**
+ * ubifs_pack_nnode - pack all the bit fields of a nnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer into which to pack
+ * @nnode: nnode to pack
+ */
+void ubifs_pack_nnode(struct ubifs_info *c, void *buf,
+ struct ubifs_nnode *nnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0;
+ uint16_t crc;
+
+ pack_bits(c, &addr, &pos, UBIFS_LPT_NNODE, UBIFS_LPT_TYPE_BITS);
+ if (c->big_lpt)
+ pack_bits(c, &addr, &pos, nnode->num, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int lnum = nnode->nbranch[i].lnum;
+
+ if (lnum == 0)
+ lnum = c->lpt_last + 1;
+ pack_bits(c, &addr, &pos, lnum - c->lpt_first, c->lpt_lnum_bits);
+ pack_bits(c, &addr, &pos, nnode->nbranch[i].offs,
+ c->lpt_offs_bits);
+ }
+ crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ c->nnode_sz - UBIFS_LPT_CRC_BYTES);
+ addr = buf;
+ pos = 0;
+ pack_bits(c, &addr, &pos, crc, UBIFS_LPT_CRC_BITS);
+}
+
+/**
+ * ubifs_pack_ltab - pack the LPT's own lprops table.
+ * @c: UBIFS file-system description object
+ * @buf: buffer into which to pack
+ * @ltab: LPT's own lprops table to pack
+ */
+void ubifs_pack_ltab(struct ubifs_info *c, void *buf,
+ struct ubifs_lpt_lprops *ltab)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0;
+ uint16_t crc;
+
+ pack_bits(c, &addr, &pos, UBIFS_LPT_LTAB, UBIFS_LPT_TYPE_BITS);
+ for (i = 0; i < c->lpt_lebs; i++) {
+ pack_bits(c, &addr, &pos, ltab[i].free, c->lpt_spc_bits);
+ pack_bits(c, &addr, &pos, ltab[i].dirty, c->lpt_spc_bits);
+ }
+ crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ c->ltab_sz - UBIFS_LPT_CRC_BYTES);
+ addr = buf;
+ pos = 0;
+ pack_bits(c, &addr, &pos, crc, UBIFS_LPT_CRC_BITS);
+}
+
+/**
+ * ubifs_pack_lsave - pack the LPT's save table.
+ * @c: UBIFS file-system description object
+ * @buf: buffer into which to pack
+ * @lsave: LPT's save table to pack
+ */
+void ubifs_pack_lsave(struct ubifs_info *c, void *buf, int *lsave)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0;
+ uint16_t crc;
+
+ pack_bits(c, &addr, &pos, UBIFS_LPT_LSAVE, UBIFS_LPT_TYPE_BITS);
+ for (i = 0; i < c->lsave_cnt; i++)
+ pack_bits(c, &addr, &pos, lsave[i], c->lnum_bits);
+ crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ c->lsave_sz - UBIFS_LPT_CRC_BYTES);
+ addr = buf;
+ pos = 0;
+ pack_bits(c, &addr, &pos, crc, UBIFS_LPT_CRC_BITS);
+}
+
+/**
+ * ubifs_add_lpt_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to which to add dirty space
+ * @dirty: amount of dirty space to add
+ */
+void ubifs_add_lpt_dirt(struct ubifs_info *c, int lnum, int dirty)
+{
+ if (!dirty || !lnum)
+ return;
+ dbg_lp("LEB %d add %d to %d",
+ lnum, dirty, c->ltab[lnum - c->lpt_first].dirty);
+ ubifs_assert(c, lnum >= c->lpt_first && lnum <= c->lpt_last);
+ c->ltab[lnum - c->lpt_first].dirty += dirty;
+}
+
+/**
+ * set_ltab - set LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number
+ * @free: amount of free space
+ * @dirty: amount of dirty space
+ */
+static void set_ltab(struct ubifs_info *c, int lnum, int free, int dirty)
+{
+ dbg_lp("LEB %d free %d dirty %d to %d %d",
+ lnum, c->ltab[lnum - c->lpt_first].free,
+ c->ltab[lnum - c->lpt_first].dirty, free, dirty);
+ ubifs_assert(c, lnum >= c->lpt_first && lnum <= c->lpt_last);
+ c->ltab[lnum - c->lpt_first].free = free;
+ c->ltab[lnum - c->lpt_first].dirty = dirty;
+}
+
+/**
+ * ubifs_add_nnode_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode for which to add dirt
+ */
+void ubifs_add_nnode_dirt(struct ubifs_info *c, struct ubifs_nnode *nnode)
+{
+ struct ubifs_nnode *np = nnode->parent;
+
+ if (np)
+ ubifs_add_lpt_dirt(c, np->nbranch[nnode->iip].lnum,
+ c->nnode_sz);
+ else {
+ ubifs_add_lpt_dirt(c, c->lpt_lnum, c->nnode_sz);
+ if (!(c->lpt_drty_flgs & LTAB_DIRTY)) {
+ c->lpt_drty_flgs |= LTAB_DIRTY;
+ ubifs_add_lpt_dirt(c, c->ltab_lnum, c->ltab_sz);
+ }
+ }
+}
+
+/**
+ * add_pnode_dirt - add dirty space to LPT LEB properties.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode for which to add dirt
+ */
+static void add_pnode_dirt(struct ubifs_info *c, struct ubifs_pnode *pnode)
+{
+ ubifs_add_lpt_dirt(c, pnode->parent->nbranch[pnode->iip].lnum,
+ c->pnode_sz);
+}
+
+/**
+ * calc_nnode_num - calculate nnode number.
+ * @row: the row in the tree (root is zero)
+ * @col: the column in the row (leftmost is zero)
+ *
+ * The nnode number is a number that uniquely identifies a nnode and can be used
+ * easily to traverse the tree from the root to that nnode.
+ *
+ * This function calculates and returns the nnode number for the nnode at @row
+ * and @col.
+ */
+static int calc_nnode_num(int row, int col)
+{
+ int num, bits;
+
+ num = 1;
+ while (row--) {
+ bits = (col & (UBIFS_LPT_FANOUT - 1));
+ col >>= UBIFS_LPT_FANOUT_SHIFT;
+ num <<= UBIFS_LPT_FANOUT_SHIFT;
+ num |= bits;
+ }
+ return num;
+}
+
+/**
+ * calc_nnode_num_from_parent - calculate nnode number.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * The nnode number is a number that uniquely identifies a nnode and can be used
+ * easily to traverse the tree from the root to that nnode.
+ *
+ * This function calculates and returns the nnode number based on the parent's
+ * nnode number and the index in parent.
+ */
+static int calc_nnode_num_from_parent(const struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ int num, shft;
+
+ if (!parent)
+ return 1;
+ shft = (c->lpt_hght - parent->level) * UBIFS_LPT_FANOUT_SHIFT;
+ num = parent->num ^ (1 << shft);
+ num |= (UBIFS_LPT_FANOUT + iip) << shft;
+ return num;
+}
+
+/**
+ * calc_pnode_num_from_parent - calculate pnode number.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * The pnode number is a number that uniquely identifies a pnode and can be used
+ * easily to traverse the tree from the root to that pnode.
+ *
+ * This function calculates and returns the pnode number based on the parent's
+ * nnode number and the index in parent.
+ */
+static int calc_pnode_num_from_parent(const struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i, n = c->lpt_hght - 1, pnum = parent->num, num = 0;
+
+ for (i = 0; i < n; i++) {
+ num <<= UBIFS_LPT_FANOUT_SHIFT;
+ num |= pnum & (UBIFS_LPT_FANOUT - 1);
+ pnum >>= UBIFS_LPT_FANOUT_SHIFT;
+ }
+ num <<= UBIFS_LPT_FANOUT_SHIFT;
+ num |= iip;
+ return num;
+}
+
+/**
+ * ubifs_create_lpt - create lpt acccording to lprops array.
+ * @c: UBIFS file-system description object
+ * @lps: array of logical eraseblock properties
+ * @lp_cnt: the length of @lps
+ * @hash: hash of the LPT is returned here
+ *
+ * This function creates lpt, the pnode will be initialized based on
+ * corresponding elements in @lps. If there are no corresponding lprops
+ * (eg. @lp_cnt is smaller than @c->main_lebs), the LEB property is set
+ * as free state.
+ */
+int ubifs_create_lpt(struct ubifs_info *c, struct ubifs_lprops *lps, int lp_cnt,
+ u8 *hash)
+{
+ int lnum, err = 0, i, j, cnt, len, alen, row;
+ int blnum, boffs, bsz, bcnt;
+ struct ubifs_pnode *pnode = NULL;
+ struct ubifs_nnode *nnode = NULL;
+ void *buf = NULL, *p;
+ struct ubifs_lpt_lprops *ltab = NULL;
+ int *lsave = NULL;
+ struct shash_desc *desc;
+
+ desc = ubifs_hash_get_desc(c);
+ if (IS_ERR(desc))
+ return PTR_ERR(desc);
+
+ lsave = kmalloc_array(c->lsave_cnt, sizeof(int), GFP_KERNEL);
+ pnode = kzalloc(sizeof(struct ubifs_pnode), GFP_KERNEL);
+ nnode = kzalloc(sizeof(struct ubifs_nnode), GFP_KERNEL);
+ buf = vmalloc(c->leb_size);
+ ltab = vmalloc(array_size(sizeof(struct ubifs_lpt_lprops),
+ c->lpt_lebs));
+ if (!pnode || !nnode || !buf || !ltab || !lsave) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ ubifs_assert(c, !c->ltab);
+ c->ltab = ltab; /* Needed by set_ltab */
+
+ /* Initialize LPT's own lprops */
+ for (i = 0; i < c->lpt_lebs; i++) {
+ ltab[i].free = c->leb_size;
+ ltab[i].dirty = 0;
+ ltab[i].tgc = 0;
+ ltab[i].cmt = 0;
+ }
+
+ lnum = c->lpt_first;
+ p = buf;
+ len = 0;
+ /* Number of leaf nodes (pnodes) */
+ cnt = c->pnode_cnt;
+
+ /*
+ * To calculate the internal node branches, we keep information about
+ * the level below.
+ */
+ blnum = lnum; /* LEB number of level below */
+ boffs = 0; /* Offset of level below */
+ bcnt = cnt; /* Number of nodes in level below */
+ bsz = c->pnode_sz; /* Size of nodes in level below */
+
+ /* Add all pnodes */
+ for (i = 0; i < cnt; i++) {
+ if (len + c->pnode_sz > c->leb_size) {
+ alen = ALIGN(len, c->min_io_size);
+ set_ltab(c, lnum, c->leb_size - alen, alen - len);
+ memset(p, 0xff, alen - len);
+ err = ubifs_leb_change(c, lnum++, buf, alen);
+ if (err)
+ goto out;
+ p = buf;
+ len = 0;
+ }
+ /* Fill in the pnode */
+ for (j = 0; j < UBIFS_LPT_FANOUT; j++) {
+ int k = (i << UBIFS_LPT_FANOUT_SHIFT) + j;
+
+ if (k < lp_cnt) {
+ pnode->lprops[j].free = lps[k].free;
+ pnode->lprops[j].dirty = lps[k].dirty;
+ pnode->lprops[j].flags = lps[k].flags;
+ } else {
+ pnode->lprops[j].free = c->leb_size;
+ pnode->lprops[j].dirty = 0;
+ pnode->lprops[j].flags = 0;
+ }
+ }
+ ubifs_pack_pnode(c, p, pnode);
+ err = ubifs_shash_update(c, desc, p, c->pnode_sz);
+ if (err)
+ goto out;
+
+ p += c->pnode_sz;
+ len += c->pnode_sz;
+ /*
+ * pnodes are simply numbered left to right starting at zero,
+ * which means the pnode number can be used easily to traverse
+ * down the tree to the corresponding pnode.
+ */
+ pnode->num += 1;
+ }
+
+ row = 0;
+ for (i = UBIFS_LPT_FANOUT; cnt > i; i <<= UBIFS_LPT_FANOUT_SHIFT)
+ row += 1;
+ /* Add all nnodes, one level at a time */
+ while (1) {
+ /* Number of internal nodes (nnodes) at next level */
+ cnt = DIV_ROUND_UP(cnt, UBIFS_LPT_FANOUT);
+ for (i = 0; i < cnt; i++) {
+ if (len + c->nnode_sz > c->leb_size) {
+ alen = ALIGN(len, c->min_io_size);
+ set_ltab(c, lnum, c->leb_size - alen,
+ alen - len);
+ memset(p, 0xff, alen - len);
+ err = ubifs_leb_change(c, lnum++, buf, alen);
+ if (err)
+ goto out;
+ p = buf;
+ len = 0;
+ }
+ /* Only 1 nnode at this level, so it is the root */
+ if (cnt == 1) {
+ c->lpt_lnum = lnum;
+ c->lpt_offs = len;
+ }
+ /* Set branches to the level below */
+ for (j = 0; j < UBIFS_LPT_FANOUT; j++) {
+ if (bcnt) {
+ if (boffs + bsz > c->leb_size) {
+ blnum += 1;
+ boffs = 0;
+ }
+ nnode->nbranch[j].lnum = blnum;
+ nnode->nbranch[j].offs = boffs;
+ boffs += bsz;
+ bcnt--;
+ } else {
+ nnode->nbranch[j].lnum = 0;
+ nnode->nbranch[j].offs = 0;
+ }
+ }
+ nnode->num = calc_nnode_num(row, i);
+ ubifs_pack_nnode(c, p, nnode);
+ p += c->nnode_sz;
+ len += c->nnode_sz;
+ }
+ /* Only 1 nnode at this level, so it is the root */
+ if (cnt == 1)
+ break;
+ /* Update the information about the level below */
+ bcnt = cnt;
+ bsz = c->nnode_sz;
+ row -= 1;
+ }
+
+ if (c->big_lpt) {
+ /* Need to add LPT's save table */
+ if (len + c->lsave_sz > c->leb_size) {
+ alen = ALIGN(len, c->min_io_size);
+ set_ltab(c, lnum, c->leb_size - alen, alen - len);
+ memset(p, 0xff, alen - len);
+ err = ubifs_leb_change(c, lnum++, buf, alen);
+ if (err)
+ goto out;
+ p = buf;
+ len = 0;
+ }
+
+ c->lsave_lnum = lnum;
+ c->lsave_offs = len;
+
+ for (i = 0; i < c->lsave_cnt && i < c->main_lebs; i++)
+ lsave[i] = c->main_first + i;
+ for (; i < c->lsave_cnt; i++)
+ lsave[i] = c->main_first;
+
+ ubifs_pack_lsave(c, p, lsave);
+ p += c->lsave_sz;
+ len += c->lsave_sz;
+ }
+
+ /* Need to add LPT's own LEB properties table */
+ if (len + c->ltab_sz > c->leb_size) {
+ alen = ALIGN(len, c->min_io_size);
+ set_ltab(c, lnum, c->leb_size - alen, alen - len);
+ memset(p, 0xff, alen - len);
+ err = ubifs_leb_change(c, lnum++, buf, alen);
+ if (err)
+ goto out;
+ p = buf;
+ len = 0;
+ }
+
+ c->ltab_lnum = lnum;
+ c->ltab_offs = len;
+
+ /* Update ltab before packing it */
+ len += c->ltab_sz;
+ alen = ALIGN(len, c->min_io_size);
+ set_ltab(c, lnum, c->leb_size - alen, alen - len);
+
+ ubifs_pack_ltab(c, p, ltab);
+ p += c->ltab_sz;
+
+ /* Write remaining buffer */
+ memset(p, 0xff, alen - len);
+ err = ubifs_leb_change(c, lnum, buf, alen);
+ if (err)
+ goto out;
+
+ err = ubifs_shash_final(c, desc, hash);
+ if (err)
+ goto out;
+
+ c->nhead_lnum = lnum;
+ c->nhead_offs = ALIGN(len, c->min_io_size);
+
+ dbg_lp("space_bits %d", c->space_bits);
+ dbg_lp("lpt_lnum_bits %d", c->lpt_lnum_bits);
+ dbg_lp("lpt_offs_bits %d", c->lpt_offs_bits);
+ dbg_lp("lpt_spc_bits %d", c->lpt_spc_bits);
+ dbg_lp("pcnt_bits %d", c->pcnt_bits);
+ dbg_lp("lnum_bits %d", c->lnum_bits);
+ dbg_lp("pnode_sz %d", c->pnode_sz);
+ dbg_lp("nnode_sz %d", c->nnode_sz);
+ dbg_lp("ltab_sz %d", c->ltab_sz);
+ dbg_lp("lsave_sz %d", c->lsave_sz);
+ dbg_lp("lsave_cnt %d", c->lsave_cnt);
+ dbg_lp("lpt_hght %d", c->lpt_hght);
+ dbg_lp("big_lpt %u", c->big_lpt);
+ dbg_lp("LPT root is at %d:%d", c->lpt_lnum, c->lpt_offs);
+ dbg_lp("LPT head is at %d:%d", c->nhead_lnum, c->nhead_offs);
+ dbg_lp("LPT ltab is at %d:%d", c->ltab_lnum, c->ltab_offs);
+ if (c->big_lpt)
+ dbg_lp("LPT lsave is at %d:%d", c->lsave_lnum, c->lsave_offs);
+out:
+ c->ltab = NULL;
+ kfree(desc);
+ kfree(lsave);
+ vfree(ltab);
+ vfree(buf);
+ kfree(nnode);
+ kfree(pnode);
+ return err;
+}
+
+/**
+ * ubifs_create_dflt_lpt - create default LPT.
+ * @c: UBIFS file-system description object
+ * @main_lebs: number of main area LEBs is passed and returned here
+ * @lpt_first: LEB number of first LPT LEB
+ * @lpt_lebs: number of LEBs for LPT is passed and returned here
+ * @big_lpt: use big LPT model is passed and returned here
+ * @hash: hash of the LPT is returned here
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_create_dflt_lpt(struct ubifs_info *c, int *main_lebs, int lpt_first,
+ int *lpt_lebs, int *big_lpt, u8 *hash)
+{
+ int node_sz, iopos, err = 0;
+ struct ubifs_lprops lps[2];
+
+ err = calc_dflt_lpt_geom(c, main_lebs, big_lpt);
+ if (err)
+ return err;
+ *lpt_lebs = c->lpt_lebs;
+
+ /* Needed by 'ubifs_pack_nnode()' and 'set_ltab()' */
+ c->lpt_first = lpt_first;
+ /* Needed by 'set_ltab()' */
+ c->lpt_last = lpt_first + c->lpt_lebs - 1;
+ /* Needed by 'ubifs_pack_lsave()' */
+ c->main_first = c->leb_cnt - *main_lebs;
+
+ /*
+ * The first pnode contains the LEB properties for the LEBs that contain
+ * the root inode node and the root index node of the index tree.
+ */
+ node_sz = ALIGN(ubifs_idx_node_sz(c, 1), 8);
+ iopos = ALIGN(node_sz, c->min_io_size);
+ lps[0].free = c->leb_size - iopos;
+ lps[0].dirty = iopos - node_sz;
+ lps[0].flags = LPROPS_INDEX;
+
+ node_sz = UBIFS_INO_NODE_SZ;
+ iopos = ALIGN(node_sz, c->min_io_size);
+ lps[1].free = c->leb_size - iopos;
+ lps[1].dirty = iopos - node_sz;
+ lps[1].flags = 0;
+
+ return ubifs_create_lpt(c, lps, 2, hash);
+}
+
+/**
+ * update_cats - add LEB properties of a pnode to LEB category lists and heaps.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode
+ *
+ * When a pnode is loaded into memory, the LEB properties it contains are added,
+ * by this function, to the LEB category lists and heaps.
+ */
+static void update_cats(struct ubifs_info *c, struct ubifs_pnode *pnode)
+{
+ int i;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int cat = pnode->lprops[i].flags & LPROPS_CAT_MASK;
+ int lnum = pnode->lprops[i].lnum;
+
+ if (!lnum)
+ return;
+ ubifs_add_to_cat(c, &pnode->lprops[i], cat);
+ }
+}
+
+/**
+ * replace_cats - add LEB properties of a pnode to LEB category lists and heaps.
+ * @c: UBIFS file-system description object
+ * @old_pnode: pnode copied
+ * @new_pnode: pnode copy
+ *
+ * During commit it is sometimes necessary to copy a pnode
+ * (see dirty_cow_pnode). When that happens, references in
+ * category lists and heaps must be replaced. This function does that.
+ */
+static void replace_cats(struct ubifs_info *c, struct ubifs_pnode *old_pnode,
+ struct ubifs_pnode *new_pnode)
+{
+ int i;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (!new_pnode->lprops[i].lnum)
+ return;
+ ubifs_replace_cat(c, &old_pnode->lprops[i],
+ &new_pnode->lprops[i]);
+ }
+}
+
+/**
+ * check_lpt_crc - check LPT node crc is correct.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing node
+ * @len: length of node
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int check_lpt_crc(const struct ubifs_info *c, void *buf, int len)
+{
+ int pos = 0;
+ uint8_t *addr = buf;
+ uint16_t crc, calc_crc;
+
+ crc = ubifs_unpack_bits(c, &addr, &pos, UBIFS_LPT_CRC_BITS);
+ calc_crc = crc16(-1, buf + UBIFS_LPT_CRC_BYTES,
+ len - UBIFS_LPT_CRC_BYTES);
+ if (crc != calc_crc) {
+ ubifs_err(c, "invalid crc in LPT node: crc %hx calc %hx",
+ crc, calc_crc);
+ dump_stack();
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * check_lpt_type - check LPT node type is correct.
+ * @c: UBIFS file-system description object
+ * @addr: address of type bit field is passed and returned updated here
+ * @pos: position of type bit field is passed and returned updated here
+ * @type: expected type
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int check_lpt_type(const struct ubifs_info *c, uint8_t **addr,
+ int *pos, int type)
+{
+ int node_type;
+
+ node_type = ubifs_unpack_bits(c, addr, pos, UBIFS_LPT_TYPE_BITS);
+ if (node_type != type) {
+ ubifs_err(c, "invalid type (%d) in LPT node type %d",
+ node_type, type);
+ dump_stack();
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * unpack_pnode - unpack a pnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing packed pnode to unpack
+ * @pnode: pnode structure to fill
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int unpack_pnode(const struct ubifs_info *c, void *buf,
+ struct ubifs_pnode *pnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(c, &addr, &pos, UBIFS_LPT_PNODE);
+ if (err)
+ return err;
+ if (c->big_lpt)
+ pnode->num = ubifs_unpack_bits(c, &addr, &pos, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops * const lprops = &pnode->lprops[i];
+
+ lprops->free = ubifs_unpack_bits(c, &addr, &pos, c->space_bits);
+ lprops->free <<= 3;
+ lprops->dirty = ubifs_unpack_bits(c, &addr, &pos, c->space_bits);
+ lprops->dirty <<= 3;
+
+ if (ubifs_unpack_bits(c, &addr, &pos, 1))
+ lprops->flags = LPROPS_INDEX;
+ else
+ lprops->flags = 0;
+ lprops->flags |= ubifs_categorize_lprops(c, lprops);
+ }
+ err = check_lpt_crc(c, buf, c->pnode_sz);
+ return err;
+}
+
+/**
+ * ubifs_unpack_nnode - unpack a nnode.
+ * @c: UBIFS file-system description object
+ * @buf: buffer containing packed nnode to unpack
+ * @nnode: nnode structure to fill
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_unpack_nnode(const struct ubifs_info *c, void *buf,
+ struct ubifs_nnode *nnode)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(c, &addr, &pos, UBIFS_LPT_NNODE);
+ if (err)
+ return err;
+ if (c->big_lpt)
+ nnode->num = ubifs_unpack_bits(c, &addr, &pos, c->pcnt_bits);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int lnum;
+
+ lnum = ubifs_unpack_bits(c, &addr, &pos, c->lpt_lnum_bits) +
+ c->lpt_first;
+ if (lnum == c->lpt_last + 1)
+ lnum = 0;
+ nnode->nbranch[i].lnum = lnum;
+ nnode->nbranch[i].offs = ubifs_unpack_bits(c, &addr, &pos,
+ c->lpt_offs_bits);
+ }
+ err = check_lpt_crc(c, buf, c->nnode_sz);
+ return err;
+}
+
+/**
+ * unpack_ltab - unpack the LPT's own lprops table.
+ * @c: UBIFS file-system description object
+ * @buf: buffer from which to unpack
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int unpack_ltab(const struct ubifs_info *c, void *buf)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(c, &addr, &pos, UBIFS_LPT_LTAB);
+ if (err)
+ return err;
+ for (i = 0; i < c->lpt_lebs; i++) {
+ int free = ubifs_unpack_bits(c, &addr, &pos, c->lpt_spc_bits);
+ int dirty = ubifs_unpack_bits(c, &addr, &pos, c->lpt_spc_bits);
+
+ if (free < 0 || free > c->leb_size || dirty < 0 ||
+ dirty > c->leb_size || free + dirty > c->leb_size)
+ return -EINVAL;
+
+ c->ltab[i].free = free;
+ c->ltab[i].dirty = dirty;
+ c->ltab[i].tgc = 0;
+ c->ltab[i].cmt = 0;
+ }
+ err = check_lpt_crc(c, buf, c->ltab_sz);
+ return err;
+}
+
+/**
+ * unpack_lsave - unpack the LPT's save table.
+ * @c: UBIFS file-system description object
+ * @buf: buffer from which to unpack
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int unpack_lsave(const struct ubifs_info *c, void *buf)
+{
+ uint8_t *addr = buf + UBIFS_LPT_CRC_BYTES;
+ int i, pos = 0, err;
+
+ err = check_lpt_type(c, &addr, &pos, UBIFS_LPT_LSAVE);
+ if (err)
+ return err;
+ for (i = 0; i < c->lsave_cnt; i++) {
+ int lnum = ubifs_unpack_bits(c, &addr, &pos, c->lnum_bits);
+
+ if (lnum < c->main_first || lnum >= c->leb_cnt)
+ return -EINVAL;
+ c->lsave[i] = lnum;
+ }
+ err = check_lpt_crc(c, buf, c->lsave_sz);
+ return err;
+}
+
+/**
+ * validate_nnode - validate a nnode.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode to validate
+ * @parent: parent nnode (or NULL for the root nnode)
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int validate_nnode(const struct ubifs_info *c, struct ubifs_nnode *nnode,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i, lvl, max_offs;
+
+ if (c->big_lpt) {
+ int num = calc_nnode_num_from_parent(c, parent, iip);
+
+ if (nnode->num != num)
+ return -EINVAL;
+ }
+ lvl = parent ? parent->level - 1 : c->lpt_hght;
+ if (lvl < 1)
+ return -EINVAL;
+ if (lvl == 1)
+ max_offs = c->leb_size - c->pnode_sz;
+ else
+ max_offs = c->leb_size - c->nnode_sz;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int lnum = nnode->nbranch[i].lnum;
+ int offs = nnode->nbranch[i].offs;
+
+ if (lnum == 0) {
+ if (offs != 0)
+ return -EINVAL;
+ continue;
+ }
+ if (lnum < c->lpt_first || lnum > c->lpt_last)
+ return -EINVAL;
+ if (offs < 0 || offs > max_offs)
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * validate_pnode - validate a pnode.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to validate
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int validate_pnode(const struct ubifs_info *c, struct ubifs_pnode *pnode,
+ struct ubifs_nnode *parent, int iip)
+{
+ int i;
+
+ if (c->big_lpt) {
+ int num = calc_pnode_num_from_parent(c, parent, iip);
+
+ if (pnode->num != num)
+ return -EINVAL;
+ }
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ int free = pnode->lprops[i].free;
+ int dirty = pnode->lprops[i].dirty;
+
+ if (free < 0 || free > c->leb_size || free % c->min_io_size ||
+ (free & 7))
+ return -EINVAL;
+ if (dirty < 0 || dirty > c->leb_size || (dirty & 7))
+ return -EINVAL;
+ if (dirty + free > c->leb_size)
+ return -EINVAL;
+ }
+ return 0;
+}
+
+/**
+ * set_pnode_lnum - set LEB numbers on a pnode.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to update
+ *
+ * This function calculates the LEB numbers for the LEB properties it contains
+ * based on the pnode number.
+ */
+static void set_pnode_lnum(const struct ubifs_info *c,
+ struct ubifs_pnode *pnode)
+{
+ int i, lnum;
+
+ lnum = (pnode->num << UBIFS_LPT_FANOUT_SHIFT) + c->main_first;
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (lnum >= c->leb_cnt)
+ return;
+ pnode->lprops[i].lnum = lnum++;
+ }
+}
+
+/**
+ * ubifs_read_nnode - read a nnode from flash and link it to the tree in memory.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode (or NULL for the root)
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_read_nnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch = NULL;
+ struct ubifs_nnode *nnode = NULL;
+ void *buf = c->lpt_nod_buf;
+ int err, lnum, offs;
+
+ if (parent) {
+ branch = &parent->nbranch[iip];
+ lnum = branch->lnum;
+ offs = branch->offs;
+ } else {
+ lnum = c->lpt_lnum;
+ offs = c->lpt_offs;
+ }
+ nnode = kzalloc(sizeof(struct ubifs_nnode), GFP_NOFS);
+ if (!nnode) {
+ err = -ENOMEM;
+ goto out;
+ }
+ if (lnum == 0) {
+ /*
+ * This nnode was not written which just means that the LEB
+ * properties in the subtree below it describe empty LEBs. We
+ * make the nnode as though we had read it, which in fact means
+ * doing almost nothing.
+ */
+ if (c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ } else {
+ err = ubifs_leb_read(c, lnum, buf, offs, c->nnode_sz, 1);
+ if (err)
+ goto out;
+ err = ubifs_unpack_nnode(c, buf, nnode);
+ if (err)
+ goto out;
+ }
+ err = validate_nnode(c, nnode, parent, iip);
+ if (err)
+ goto out;
+ if (!c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ if (parent) {
+ branch->nnode = nnode;
+ nnode->level = parent->level - 1;
+ } else {
+ c->nroot = nnode;
+ nnode->level = c->lpt_hght;
+ }
+ nnode->parent = parent;
+ nnode->iip = iip;
+ return 0;
+
+out:
+ ubifs_err(c, "error %d reading nnode at %d:%d", err, lnum, offs);
+ dump_stack();
+ kfree(nnode);
+ return err;
+}
+
+/**
+ * read_pnode - read a pnode from flash and link it to the tree in memory.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int read_pnode(struct ubifs_info *c, struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_pnode *pnode = NULL;
+ void *buf = c->lpt_nod_buf;
+ int err, lnum, offs;
+
+ branch = &parent->nbranch[iip];
+ lnum = branch->lnum;
+ offs = branch->offs;
+ pnode = kzalloc(sizeof(struct ubifs_pnode), GFP_NOFS);
+ if (!pnode)
+ return -ENOMEM;
+
+ if (lnum == 0) {
+ /*
+ * This pnode was not written which just means that the LEB
+ * properties in it describe empty LEBs. We make the pnode as
+ * though we had read it.
+ */
+ int i;
+
+ if (c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops * const lprops = &pnode->lprops[i];
+
+ lprops->free = c->leb_size;
+ lprops->flags = ubifs_categorize_lprops(c, lprops);
+ }
+ } else {
+ err = ubifs_leb_read(c, lnum, buf, offs, c->pnode_sz, 1);
+ if (err)
+ goto out;
+ err = unpack_pnode(c, buf, pnode);
+ if (err)
+ goto out;
+ }
+ err = validate_pnode(c, pnode, parent, iip);
+ if (err)
+ goto out;
+ if (!c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ branch->pnode = pnode;
+ pnode->parent = parent;
+ pnode->iip = iip;
+ set_pnode_lnum(c, pnode);
+ c->pnodes_have += 1;
+ return 0;
+
+out:
+ ubifs_err(c, "error %d reading pnode at %d:%d", err, lnum, offs);
+ ubifs_dump_pnode(c, pnode, parent, iip);
+ dump_stack();
+ ubifs_err(c, "calc num: %d", calc_pnode_num_from_parent(c, parent, iip));
+ kfree(pnode);
+ return err;
+}
+
+/**
+ * read_ltab - read LPT's own lprops table.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int read_ltab(struct ubifs_info *c)
+{
+ int err;
+ void *buf;
+
+ buf = vmalloc(c->ltab_sz);
+ if (!buf)
+ return -ENOMEM;
+ err = ubifs_leb_read(c, c->ltab_lnum, buf, c->ltab_offs, c->ltab_sz, 1);
+ if (err)
+ goto out;
+ err = unpack_ltab(c, buf);
+out:
+ vfree(buf);
+ return err;
+}
+
+/**
+ * read_lsave - read LPT's save table.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int read_lsave(struct ubifs_info *c)
+{
+ int err, i;
+ void *buf;
+
+ buf = vmalloc(c->lsave_sz);
+ if (!buf)
+ return -ENOMEM;
+ err = ubifs_leb_read(c, c->lsave_lnum, buf, c->lsave_offs,
+ c->lsave_sz, 1);
+ if (err)
+ goto out;
+ err = unpack_lsave(c, buf);
+ if (err)
+ goto out;
+ for (i = 0; i < c->lsave_cnt; i++) {
+ int lnum = c->lsave[i];
+ struct ubifs_lprops *lprops;
+
+ /*
+ * Due to automatic resizing, the values in the lsave table
+ * could be beyond the volume size - just ignore them.
+ */
+ if (lnum >= c->leb_cnt)
+ continue;
+ lprops = ubifs_lpt_lookup(c, lnum);
+ if (IS_ERR(lprops)) {
+ err = PTR_ERR(lprops);
+ goto out;
+ }
+ }
+out:
+ vfree(buf);
+ return err;
+}
+
+/**
+ * ubifs_get_nnode - get a nnode.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode (or NULL for the root)
+ * @iip: index in parent
+ *
+ * This function returns a pointer to the nnode on success or a negative error
+ * code on failure.
+ */
+struct ubifs_nnode *ubifs_get_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_nnode *nnode;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ nnode = branch->nnode;
+ if (nnode)
+ return nnode;
+ err = ubifs_read_nnode(c, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ return branch->nnode;
+}
+
+/**
+ * ubifs_get_pnode - get a pnode.
+ * @c: UBIFS file-system description object
+ * @parent: parent nnode
+ * @iip: index in parent
+ *
+ * This function returns a pointer to the pnode on success or a negative error
+ * code on failure.
+ */
+struct ubifs_pnode *ubifs_get_pnode(struct ubifs_info *c,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_pnode *pnode;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ pnode = branch->pnode;
+ if (pnode)
+ return pnode;
+ err = read_pnode(c, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ update_cats(c, branch->pnode);
+ return branch->pnode;
+}
+
+/**
+ * ubifs_pnode_lookup - lookup a pnode in the LPT.
+ * @c: UBIFS file-system description object
+ * @i: pnode number (0 to (main_lebs - 1) / UBIFS_LPT_FANOUT)
+ *
+ * This function returns a pointer to the pnode on success or a negative
+ * error code on failure.
+ */
+struct ubifs_pnode *ubifs_pnode_lookup(struct ubifs_info *c, int i)
+{
+ int err, h, iip, shft;
+ struct ubifs_nnode *nnode;
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return ERR_PTR(err);
+ }
+ i <<= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = c->nroot;
+ shft = c->lpt_hght * UBIFS_LPT_FANOUT_SHIFT;
+ for (h = 1; h < c->lpt_hght; h++) {
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = ubifs_get_nnode(c, nnode, iip);
+ if (IS_ERR(nnode))
+ return ERR_CAST(nnode);
+ }
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ return ubifs_get_pnode(c, nnode, iip);
+}
+
+/**
+ * ubifs_lpt_lookup - lookup LEB properties in the LPT.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to lookup
+ *
+ * This function returns a pointer to the LEB properties on success or a
+ * negative error code on failure.
+ */
+struct ubifs_lprops *ubifs_lpt_lookup(struct ubifs_info *c, int lnum)
+{
+ int i, iip;
+ struct ubifs_pnode *pnode;
+
+ i = lnum - c->main_first;
+ pnode = ubifs_pnode_lookup(c, i >> UBIFS_LPT_FANOUT_SHIFT);
+ if (IS_ERR(pnode))
+ return ERR_CAST(pnode);
+ iip = (i & (UBIFS_LPT_FANOUT - 1));
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d", lnum,
+ pnode->lprops[iip].free, pnode->lprops[iip].dirty,
+ pnode->lprops[iip].flags);
+ return &pnode->lprops[iip];
+}
+
+/**
+ * dirty_cow_nnode - ensure a nnode is not being committed.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode to check
+ *
+ * Returns dirtied nnode on success or negative error code on failure.
+ */
+static struct ubifs_nnode *dirty_cow_nnode(struct ubifs_info *c,
+ struct ubifs_nnode *nnode)
+{
+ struct ubifs_nnode *n;
+ int i;
+
+ if (!test_bit(COW_CNODE, &nnode->flags)) {
+ /* nnode is not being committed */
+ if (!test_and_set_bit(DIRTY_CNODE, &nnode->flags)) {
+ c->dirty_nn_cnt += 1;
+ ubifs_add_nnode_dirt(c, nnode);
+ }
+ return nnode;
+ }
+
+ /* nnode is being committed, so copy it */
+ n = kmemdup(nnode, sizeof(struct ubifs_nnode), GFP_NOFS);
+ if (unlikely(!n))
+ return ERR_PTR(-ENOMEM);
+
+ n->cnext = NULL;
+ __set_bit(DIRTY_CNODE, &n->flags);
+ __clear_bit(COW_CNODE, &n->flags);
+
+ /* The children now have new parent */
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_nbranch *branch = &n->nbranch[i];
+
+ if (branch->cnode)
+ branch->cnode->parent = n;
+ }
+
+ ubifs_assert(c, !test_bit(OBSOLETE_CNODE, &nnode->flags));
+ __set_bit(OBSOLETE_CNODE, &nnode->flags);
+
+ c->dirty_nn_cnt += 1;
+ ubifs_add_nnode_dirt(c, nnode);
+ if (nnode->parent)
+ nnode->parent->nbranch[n->iip].nnode = n;
+ else
+ c->nroot = n;
+ return n;
+}
+
+/**
+ * dirty_cow_pnode - ensure a pnode is not being committed.
+ * @c: UBIFS file-system description object
+ * @pnode: pnode to check
+ *
+ * Returns dirtied pnode on success or negative error code on failure.
+ */
+static struct ubifs_pnode *dirty_cow_pnode(struct ubifs_info *c,
+ struct ubifs_pnode *pnode)
+{
+ struct ubifs_pnode *p;
+
+ if (!test_bit(COW_CNODE, &pnode->flags)) {
+ /* pnode is not being committed */
+ if (!test_and_set_bit(DIRTY_CNODE, &pnode->flags)) {
+ c->dirty_pn_cnt += 1;
+ add_pnode_dirt(c, pnode);
+ }
+ return pnode;
+ }
+
+ /* pnode is being committed, so copy it */
+ p = kmemdup(pnode, sizeof(struct ubifs_pnode), GFP_NOFS);
+ if (unlikely(!p))
+ return ERR_PTR(-ENOMEM);
+
+ p->cnext = NULL;
+ __set_bit(DIRTY_CNODE, &p->flags);
+ __clear_bit(COW_CNODE, &p->flags);
+ replace_cats(c, pnode, p);
+
+ ubifs_assert(c, !test_bit(OBSOLETE_CNODE, &pnode->flags));
+ __set_bit(OBSOLETE_CNODE, &pnode->flags);
+
+ c->dirty_pn_cnt += 1;
+ add_pnode_dirt(c, pnode);
+ pnode->parent->nbranch[p->iip].pnode = p;
+ return p;
+}
+
+/**
+ * ubifs_lpt_lookup_dirty - lookup LEB properties in the LPT.
+ * @c: UBIFS file-system description object
+ * @lnum: LEB number to lookup
+ *
+ * This function returns a pointer to the LEB properties on success or a
+ * negative error code on failure.
+ */
+struct ubifs_lprops *ubifs_lpt_lookup_dirty(struct ubifs_info *c, int lnum)
+{
+ int err, i, h, iip, shft;
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return ERR_PTR(err);
+ }
+ nnode = c->nroot;
+ nnode = dirty_cow_nnode(c, nnode);
+ if (IS_ERR(nnode))
+ return ERR_CAST(nnode);
+ i = lnum - c->main_first;
+ shft = c->lpt_hght * UBIFS_LPT_FANOUT_SHIFT;
+ for (h = 1; h < c->lpt_hght; h++) {
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = ubifs_get_nnode(c, nnode, iip);
+ if (IS_ERR(nnode))
+ return ERR_CAST(nnode);
+ nnode = dirty_cow_nnode(c, nnode);
+ if (IS_ERR(nnode))
+ return ERR_CAST(nnode);
+ }
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ pnode = ubifs_get_pnode(c, nnode, iip);
+ if (IS_ERR(pnode))
+ return ERR_CAST(pnode);
+ pnode = dirty_cow_pnode(c, pnode);
+ if (IS_ERR(pnode))
+ return ERR_CAST(pnode);
+ iip = (i & (UBIFS_LPT_FANOUT - 1));
+ dbg_lp("LEB %d, free %d, dirty %d, flags %d", lnum,
+ pnode->lprops[iip].free, pnode->lprops[iip].dirty,
+ pnode->lprops[iip].flags);
+ ubifs_assert(c, test_bit(DIRTY_CNODE, &pnode->flags));
+ return &pnode->lprops[iip];
+}
+
+/**
+ * ubifs_lpt_calc_hash - Calculate hash of the LPT pnodes
+ * @c: UBIFS file-system description object
+ * @hash: the returned hash of the LPT pnodes
+ *
+ * This function iterates over the LPT pnodes and creates a hash over them.
+ * Returns 0 for success or a negative error code otherwise.
+ */
+int ubifs_lpt_calc_hash(struct ubifs_info *c, u8 *hash)
+{
+ struct ubifs_nnode *nnode, *nn;
+ struct ubifs_cnode *cnode;
+ struct shash_desc *desc;
+ int iip = 0, i;
+ int bufsiz = max_t(int, c->nnode_sz, c->pnode_sz);
+ void *buf;
+ int err;
+
+ if (!ubifs_authenticated(c))
+ return 0;
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return err;
+ }
+
+ desc = ubifs_hash_get_desc(c);
+ if (IS_ERR(desc))
+ return PTR_ERR(desc);
+
+ buf = kmalloc(bufsiz, GFP_NOFS);
+ if (!buf) {
+ err = -ENOMEM;
+ goto out;
+ }
+
+ cnode = (struct ubifs_cnode *)c->nroot;
+
+ while (cnode) {
+ nnode = cnode->parent;
+ nn = (struct ubifs_nnode *)cnode;
+ if (cnode->level > 1) {
+ while (iip < UBIFS_LPT_FANOUT) {
+ if (nn->nbranch[iip].lnum == 0) {
+ /* Go right */
+ iip++;
+ continue;
+ }
+
+ nnode = ubifs_get_nnode(c, nn, iip);
+ if (IS_ERR(nnode)) {
+ err = PTR_ERR(nnode);
+ goto out;
+ }
+
+ /* Go down */
+ iip = 0;
+ cnode = (struct ubifs_cnode *)nnode;
+ break;
+ }
+ if (iip < UBIFS_LPT_FANOUT)
+ continue;
+ } else {
+ struct ubifs_pnode *pnode;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ if (nn->nbranch[i].lnum == 0)
+ continue;
+ pnode = ubifs_get_pnode(c, nn, i);
+ if (IS_ERR(pnode)) {
+ err = PTR_ERR(pnode);
+ goto out;
+ }
+
+ ubifs_pack_pnode(c, buf, pnode);
+ err = ubifs_shash_update(c, desc, buf,
+ c->pnode_sz);
+ if (err)
+ goto out;
+ }
+ }
+ /* Go up and to the right */
+ iip = cnode->iip + 1;
+ cnode = (struct ubifs_cnode *)nnode;
+ }
+
+ err = ubifs_shash_final(c, desc, hash);
+out:
+ kfree(desc);
+ kfree(buf);
+
+ return err;
+}
+
+/**
+ * lpt_check_hash - check the hash of the LPT.
+ * @c: UBIFS file-system description object
+ *
+ * This function calculates a hash over all pnodes in the LPT and compares it with
+ * the hash stored in the master node. Returns %0 on success and a negative error
+ * code on failure.
+ */
+static int lpt_check_hash(struct ubifs_info *c)
+{
+ int err;
+ u8 hash[UBIFS_HASH_ARR_SZ];
+
+ if (!ubifs_authenticated(c))
+ return 0;
+
+ err = ubifs_lpt_calc_hash(c, hash);
+ if (err)
+ return err;
+
+ if (ubifs_check_hash(c, c->mst_node->hash_lpt, hash)) {
+ err = -EPERM;
+ ubifs_err(c, "Failed to authenticate LPT");
+ } else {
+ err = 0;
+ }
+
+ return err;
+}
+
+/**
+ * lpt_init_rd - initialize the LPT for reading.
+ * @c: UBIFS file-system description object
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int lpt_init_rd(struct ubifs_info *c)
+{
+ int err, i;
+
+ c->ltab = vmalloc(array_size(sizeof(struct ubifs_lpt_lprops),
+ c->lpt_lebs));
+ if (!c->ltab)
+ return -ENOMEM;
+
+ i = max_t(int, c->nnode_sz, c->pnode_sz);
+ c->lpt_nod_buf = kmalloc(i, GFP_KERNEL);
+ if (!c->lpt_nod_buf)
+ return -ENOMEM;
+
+ for (i = 0; i < LPROPS_HEAP_CNT; i++) {
+ c->lpt_heap[i].arr = kmalloc_array(LPT_HEAP_SZ,
+ sizeof(void *),
+ GFP_KERNEL);
+ if (!c->lpt_heap[i].arr)
+ return -ENOMEM;
+ c->lpt_heap[i].cnt = 0;
+ c->lpt_heap[i].max_cnt = LPT_HEAP_SZ;
+ }
+
+ c->dirty_idx.arr = kmalloc_array(LPT_HEAP_SZ, sizeof(void *),
+ GFP_KERNEL);
+ if (!c->dirty_idx.arr)
+ return -ENOMEM;
+ c->dirty_idx.cnt = 0;
+ c->dirty_idx.max_cnt = LPT_HEAP_SZ;
+
+ err = read_ltab(c);
+ if (err)
+ return err;
+
+ err = lpt_check_hash(c);
+ if (err)
+ return err;
+
+ dbg_lp("space_bits %d", c->space_bits);
+ dbg_lp("lpt_lnum_bits %d", c->lpt_lnum_bits);
+ dbg_lp("lpt_offs_bits %d", c->lpt_offs_bits);
+ dbg_lp("lpt_spc_bits %d", c->lpt_spc_bits);
+ dbg_lp("pcnt_bits %d", c->pcnt_bits);
+ dbg_lp("lnum_bits %d", c->lnum_bits);
+ dbg_lp("pnode_sz %d", c->pnode_sz);
+ dbg_lp("nnode_sz %d", c->nnode_sz);
+ dbg_lp("ltab_sz %d", c->ltab_sz);
+ dbg_lp("lsave_sz %d", c->lsave_sz);
+ dbg_lp("lsave_cnt %d", c->lsave_cnt);
+ dbg_lp("lpt_hght %d", c->lpt_hght);
+ dbg_lp("big_lpt %u", c->big_lpt);
+ dbg_lp("LPT root is at %d:%d", c->lpt_lnum, c->lpt_offs);
+ dbg_lp("LPT head is at %d:%d", c->nhead_lnum, c->nhead_offs);
+ dbg_lp("LPT ltab is at %d:%d", c->ltab_lnum, c->ltab_offs);
+ if (c->big_lpt)
+ dbg_lp("LPT lsave is at %d:%d", c->lsave_lnum, c->lsave_offs);
+
+ return 0;
+}
+
+/**
+ * lpt_init_wr - initialize the LPT for writing.
+ * @c: UBIFS file-system description object
+ *
+ * 'lpt_init_rd()' must have been called already.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int lpt_init_wr(struct ubifs_info *c)
+{
+ int err, i;
+
+ c->ltab_cmt = vmalloc(array_size(sizeof(struct ubifs_lpt_lprops),
+ c->lpt_lebs));
+ if (!c->ltab_cmt)
+ return -ENOMEM;
+
+ c->lpt_buf = vmalloc(c->leb_size);
+ if (!c->lpt_buf)
+ return -ENOMEM;
+
+ if (c->big_lpt) {
+ c->lsave = kmalloc_array(c->lsave_cnt, sizeof(int), GFP_NOFS);
+ if (!c->lsave)
+ return -ENOMEM;
+ err = read_lsave(c);
+ if (err)
+ return err;
+ }
+
+ for (i = 0; i < c->lpt_lebs; i++)
+ if (c->ltab[i].free == c->leb_size) {
+ err = ubifs_leb_unmap(c, i + c->lpt_first);
+ if (err)
+ return err;
+ }
+
+ return 0;
+}
+
+/**
+ * ubifs_lpt_init - initialize the LPT.
+ * @c: UBIFS file-system description object
+ * @rd: whether to initialize lpt for reading
+ * @wr: whether to initialize lpt for writing
+ *
+ * For mounting 'rw', @rd and @wr are both true. For mounting 'ro', @rd is true
+ * and @wr is false. For mounting from 'ro' to 'rw', @rd is false and @wr is
+ * true.
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_lpt_init(struct ubifs_info *c, int rd, int wr)
+{
+ int err;
+
+ if (rd) {
+ err = lpt_init_rd(c);
+ if (err)
+ goto out_err;
+ }
+
+ if (wr) {
+ err = lpt_init_wr(c);
+ if (err)
+ goto out_err;
+ }
+
+ return 0;
+
+out_err:
+ if (wr)
+ ubifs_lpt_free(c, 1);
+ if (rd)
+ ubifs_lpt_free(c, 0);
+ return err;
+}
+
+/**
+ * struct lpt_scan_node - somewhere to put nodes while we scan LPT.
+ * @nnode: where to keep a nnode
+ * @pnode: where to keep a pnode
+ * @cnode: where to keep a cnode
+ * @in_tree: is the node in the tree in memory
+ * @ptr.nnode: pointer to the nnode (if it is an nnode) which may be here or in
+ * the tree
+ * @ptr.pnode: ditto for pnode
+ * @ptr.cnode: ditto for cnode
+ */
+struct lpt_scan_node {
+ union {
+ struct ubifs_nnode nnode;
+ struct ubifs_pnode pnode;
+ struct ubifs_cnode cnode;
+ };
+ int in_tree;
+ union {
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+ struct ubifs_cnode *cnode;
+ } ptr;
+};
+
+/**
+ * scan_get_nnode - for the scan, get a nnode from either the tree or flash.
+ * @c: the UBIFS file-system description object
+ * @path: where to put the nnode
+ * @parent: parent of the nnode
+ * @iip: index in parent of the nnode
+ *
+ * This function returns a pointer to the nnode on success or a negative error
+ * code on failure.
+ */
+static struct ubifs_nnode *scan_get_nnode(struct ubifs_info *c,
+ struct lpt_scan_node *path,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_nnode *nnode;
+ void *buf = c->lpt_nod_buf;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ nnode = branch->nnode;
+ if (nnode) {
+ path->in_tree = 1;
+ path->ptr.nnode = nnode;
+ return nnode;
+ }
+ nnode = &path->nnode;
+ path->in_tree = 0;
+ path->ptr.nnode = nnode;
+ memset(nnode, 0, sizeof(struct ubifs_nnode));
+ if (branch->lnum == 0) {
+ /*
+ * This nnode was not written which just means that the LEB
+ * properties in the subtree below it describe empty LEBs. We
+ * make the nnode as though we had read it, which in fact means
+ * doing almost nothing.
+ */
+ if (c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ } else {
+ err = ubifs_leb_read(c, branch->lnum, buf, branch->offs,
+ c->nnode_sz, 1);
+ if (err)
+ return ERR_PTR(err);
+ err = ubifs_unpack_nnode(c, buf, nnode);
+ if (err)
+ return ERR_PTR(err);
+ }
+ err = validate_nnode(c, nnode, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ if (!c->big_lpt)
+ nnode->num = calc_nnode_num_from_parent(c, parent, iip);
+ nnode->level = parent->level - 1;
+ nnode->parent = parent;
+ nnode->iip = iip;
+ return nnode;
+}
+
+/**
+ * scan_get_pnode - for the scan, get a pnode from either the tree or flash.
+ * @c: the UBIFS file-system description object
+ * @path: where to put the pnode
+ * @parent: parent of the pnode
+ * @iip: index in parent of the pnode
+ *
+ * This function returns a pointer to the pnode on success or a negative error
+ * code on failure.
+ */
+static struct ubifs_pnode *scan_get_pnode(struct ubifs_info *c,
+ struct lpt_scan_node *path,
+ struct ubifs_nnode *parent, int iip)
+{
+ struct ubifs_nbranch *branch;
+ struct ubifs_pnode *pnode;
+ void *buf = c->lpt_nod_buf;
+ int err;
+
+ branch = &parent->nbranch[iip];
+ pnode = branch->pnode;
+ if (pnode) {
+ path->in_tree = 1;
+ path->ptr.pnode = pnode;
+ return pnode;
+ }
+ pnode = &path->pnode;
+ path->in_tree = 0;
+ path->ptr.pnode = pnode;
+ memset(pnode, 0, sizeof(struct ubifs_pnode));
+ if (branch->lnum == 0) {
+ /*
+ * This pnode was not written which just means that the LEB
+ * properties in it describe empty LEBs. We make the pnode as
+ * though we had read it.
+ */
+ int i;
+
+ if (c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops * const lprops = &pnode->lprops[i];
+
+ lprops->free = c->leb_size;
+ lprops->flags = ubifs_categorize_lprops(c, lprops);
+ }
+ } else {
+ ubifs_assert(c, branch->lnum >= c->lpt_first &&
+ branch->lnum <= c->lpt_last);
+ ubifs_assert(c, branch->offs >= 0 && branch->offs < c->leb_size);
+ err = ubifs_leb_read(c, branch->lnum, buf, branch->offs,
+ c->pnode_sz, 1);
+ if (err)
+ return ERR_PTR(err);
+ err = unpack_pnode(c, buf, pnode);
+ if (err)
+ return ERR_PTR(err);
+ }
+ err = validate_pnode(c, pnode, parent, iip);
+ if (err)
+ return ERR_PTR(err);
+ if (!c->big_lpt)
+ pnode->num = calc_pnode_num_from_parent(c, parent, iip);
+ pnode->parent = parent;
+ pnode->iip = iip;
+ set_pnode_lnum(c, pnode);
+ return pnode;
+}
+
+/**
+ * ubifs_lpt_scan_nolock - scan the LPT.
+ * @c: the UBIFS file-system description object
+ * @start_lnum: LEB number from which to start scanning
+ * @end_lnum: LEB number at which to stop scanning
+ * @scan_cb: callback function called for each lprops
+ * @data: data to be passed to the callback function
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int ubifs_lpt_scan_nolock(struct ubifs_info *c, int start_lnum, int end_lnum,
+ ubifs_lpt_scan_callback scan_cb, void *data)
+{
+ int err = 0, i, h, iip, shft;
+ struct ubifs_nnode *nnode;
+ struct ubifs_pnode *pnode;
+ struct lpt_scan_node *path;
+
+ if (start_lnum == -1) {
+ start_lnum = end_lnum + 1;
+ if (start_lnum >= c->leb_cnt)
+ start_lnum = c->main_first;
+ }
+
+ ubifs_assert(c, start_lnum >= c->main_first && start_lnum < c->leb_cnt);
+ ubifs_assert(c, end_lnum >= c->main_first && end_lnum < c->leb_cnt);
+
+ if (!c->nroot) {
+ err = ubifs_read_nnode(c, NULL, 0);
+ if (err)
+ return err;
+ }
+
+ path = kmalloc_array(c->lpt_hght + 1, sizeof(struct lpt_scan_node),
+ GFP_NOFS);
+ if (!path)
+ return -ENOMEM;
+
+ path[0].ptr.nnode = c->nroot;
+ path[0].in_tree = 1;
+again:
+ /* Descend to the pnode containing start_lnum */
+ nnode = c->nroot;
+ i = start_lnum - c->main_first;
+ shft = c->lpt_hght * UBIFS_LPT_FANOUT_SHIFT;
+ for (h = 1; h < c->lpt_hght; h++) {
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ shft -= UBIFS_LPT_FANOUT_SHIFT;
+ nnode = scan_get_nnode(c, path + h, nnode, iip);
+ if (IS_ERR(nnode)) {
+ err = PTR_ERR(nnode);
+ goto out;
+ }
+ }
+ iip = ((i >> shft) & (UBIFS_LPT_FANOUT - 1));
+ pnode = scan_get_pnode(c, path + h, nnode, iip);
+ if (IS_ERR(pnode)) {
+ err = PTR_ERR(pnode);
+ goto out;
+ }
+ iip = (i & (UBIFS_LPT_FANOUT - 1));
+
+ /* Loop for each lprops */
+ while (1) {
+ struct ubifs_lprops *lprops = &pnode->lprops[iip];
+ int ret, lnum = lprops->lnum;
+
+ ret = scan_cb(c, lprops, path[h].in_tree, data);
+ if (ret < 0) {
+ err = ret;
+ goto out;
+ }
+ if (ret & LPT_SCAN_ADD) {
+ /* Add all the nodes in path to the tree in memory */
+ for (h = 1; h < c->lpt_hght; h++) {
+ const size_t sz = sizeof(struct ubifs_nnode);
+ struct ubifs_nnode *parent;
+
+ if (path[h].in_tree)
+ continue;
+ nnode = kmemdup(&path[h].nnode, sz, GFP_NOFS);
+ if (!nnode) {
+ err = -ENOMEM;
+ goto out;
+ }
+ parent = nnode->parent;
+ parent->nbranch[nnode->iip].nnode = nnode;
+ path[h].ptr.nnode = nnode;
+ path[h].in_tree = 1;
+ path[h + 1].cnode.parent = nnode;
+ }
+ if (path[h].in_tree)
+ ubifs_ensure_cat(c, lprops);
+ else {
+ const size_t sz = sizeof(struct ubifs_pnode);
+ struct ubifs_nnode *parent;
+
+ pnode = kmemdup(&path[h].pnode, sz, GFP_NOFS);
+ if (!pnode) {
+ err = -ENOMEM;
+ goto out;
+ }
+ parent = pnode->parent;
+ parent->nbranch[pnode->iip].pnode = pnode;
+ path[h].ptr.pnode = pnode;
+ path[h].in_tree = 1;
+ update_cats(c, pnode);
+ c->pnodes_have += 1;
+ }
+ err = dbg_check_lpt_nodes(c, (struct ubifs_cnode *)
+ c->nroot, 0, 0);
+ if (err)
+ goto out;
+ err = dbg_check_cats(c);
+ if (err)
+ goto out;
+ }
+ if (ret & LPT_SCAN_STOP) {
+ err = 0;
+ break;
+ }
+ /* Get the next lprops */
+ if (lnum == end_lnum) {
+ /*
+ * We got to the end without finding what we were
+ * looking for
+ */
+ err = -ENOSPC;
+ goto out;
+ }
+ if (lnum + 1 >= c->leb_cnt) {
+ /* Wrap-around to the beginning */
+ start_lnum = c->main_first;
+ goto again;
+ }
+ if (iip + 1 < UBIFS_LPT_FANOUT) {
+ /* Next lprops is in the same pnode */
+ iip += 1;
+ continue;
+ }
+ /* We need to get the next pnode. Go up until we can go right */
+ iip = pnode->iip;
+ while (1) {
+ h -= 1;
+ ubifs_assert(c, h >= 0);
+ nnode = path[h].ptr.nnode;
+ if (iip + 1 < UBIFS_LPT_FANOUT)
+ break;
+ iip = nnode->iip;
+ }
+ /* Go right */
+ iip += 1;
+ /* Descend to the pnode */
+ h += 1;
+ for (; h < c->lpt_hght; h++) {
+ nnode = scan_get_nnode(c, path + h, nnode, iip);
+ if (IS_ERR(nnode)) {
+ err = PTR_ERR(nnode);
+ goto out;
+ }
+ iip = 0;
+ }
+ pnode = scan_get_pnode(c, path + h, nnode, iip);
+ if (IS_ERR(pnode)) {
+ err = PTR_ERR(pnode);
+ goto out;
+ }
+ iip = 0;
+ }
+out:
+ kfree(path);
+ return err;
+}
+
+/**
+ * dbg_chk_pnode - check a pnode.
+ * @c: the UBIFS file-system description object
+ * @pnode: pnode to check
+ * @col: pnode column
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+static int dbg_chk_pnode(struct ubifs_info *c, struct ubifs_pnode *pnode,
+ int col)
+{
+ int i;
+
+ if (pnode->num != col) {
+ ubifs_err(c, "pnode num %d expected %d parent num %d iip %d",
+ pnode->num, col, pnode->parent->num, pnode->iip);
+ return -EINVAL;
+ }
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_lprops *lp, *lprops = &pnode->lprops[i];
+ int lnum = (pnode->num << UBIFS_LPT_FANOUT_SHIFT) + i +
+ c->main_first;
+ int found, cat = lprops->flags & LPROPS_CAT_MASK;
+ struct ubifs_lpt_heap *heap;
+ struct list_head *list = NULL;
+
+ if (lnum >= c->leb_cnt)
+ continue;
+ if (lprops->lnum != lnum) {
+ ubifs_err(c, "bad LEB number %d expected %d",
+ lprops->lnum, lnum);
+ return -EINVAL;
+ }
+ if (lprops->flags & LPROPS_TAKEN) {
+ if (cat != LPROPS_UNCAT) {
+ ubifs_err(c, "LEB %d taken but not uncat %d",
+ lprops->lnum, cat);
+ return -EINVAL;
+ }
+ continue;
+ }
+ if (lprops->flags & LPROPS_INDEX) {
+ switch (cat) {
+ case LPROPS_UNCAT:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FRDI_IDX:
+ break;
+ default:
+ ubifs_err(c, "LEB %d index but cat %d",
+ lprops->lnum, cat);
+ return -EINVAL;
+ }
+ } else {
+ switch (cat) {
+ case LPROPS_UNCAT:
+ case LPROPS_DIRTY:
+ case LPROPS_FREE:
+ case LPROPS_EMPTY:
+ case LPROPS_FREEABLE:
+ break;
+ default:
+ ubifs_err(c, "LEB %d not index but cat %d",
+ lprops->lnum, cat);
+ return -EINVAL;
+ }
+ }
+ switch (cat) {
+ case LPROPS_UNCAT:
+ list = &c->uncat_list;
+ break;
+ case LPROPS_EMPTY:
+ list = &c->empty_list;
+ break;
+ case LPROPS_FREEABLE:
+ list = &c->freeable_list;
+ break;
+ case LPROPS_FRDI_IDX:
+ list = &c->frdi_idx_list;
+ break;
+ }
+ found = 0;
+ switch (cat) {
+ case LPROPS_DIRTY:
+ case LPROPS_DIRTY_IDX:
+ case LPROPS_FREE:
+ heap = &c->lpt_heap[cat - 1];
+ if (lprops->hpos < heap->cnt &&
+ heap->arr[lprops->hpos] == lprops)
+ found = 1;
+ break;
+ case LPROPS_UNCAT:
+ case LPROPS_EMPTY:
+ case LPROPS_FREEABLE:
+ case LPROPS_FRDI_IDX:
+ list_for_each_entry(lp, list, list)
+ if (lprops == lp) {
+ found = 1;
+ break;
+ }
+ break;
+ }
+ if (!found) {
+ ubifs_err(c, "LEB %d cat %d not found in cat heap/list",
+ lprops->lnum, cat);
+ return -EINVAL;
+ }
+ switch (cat) {
+ case LPROPS_EMPTY:
+ if (lprops->free != c->leb_size) {
+ ubifs_err(c, "LEB %d cat %d free %d dirty %d",
+ lprops->lnum, cat, lprops->free,
+ lprops->dirty);
+ return -EINVAL;
+ }
+ break;
+ case LPROPS_FREEABLE:
+ case LPROPS_FRDI_IDX:
+ if (lprops->free + lprops->dirty != c->leb_size) {
+ ubifs_err(c, "LEB %d cat %d free %d dirty %d",
+ lprops->lnum, cat, lprops->free,
+ lprops->dirty);
+ return -EINVAL;
+ }
+ break;
+ }
+ }
+ return 0;
+}
+
+/**
+ * dbg_check_lpt_nodes - check nnodes and pnodes.
+ * @c: the UBIFS file-system description object
+ * @cnode: next cnode (nnode or pnode) to check
+ * @row: row of cnode (root is zero)
+ * @col: column of cnode (leftmost is zero)
+ *
+ * This function returns %0 on success and a negative error code on failure.
+ */
+int dbg_check_lpt_nodes(struct ubifs_info *c, struct ubifs_cnode *cnode,
+ int row, int col)
+{
+ struct ubifs_nnode *nnode, *nn;
+ struct ubifs_cnode *cn;
+ int num, iip = 0, err;
+
+ if (!dbg_is_chk_lprops(c))
+ return 0;
+
+ while (cnode) {
+ ubifs_assert(c, row >= 0);
+ nnode = cnode->parent;
+ if (cnode->level) {
+ /* cnode is a nnode */
+ num = calc_nnode_num(row, col);
+ if (cnode->num != num) {
+ ubifs_err(c, "nnode num %d expected %d parent num %d iip %d",
+ cnode->num, num,
+ (nnode ? nnode->num : 0), cnode->iip);
+ return -EINVAL;
+ }
+ nn = (struct ubifs_nnode *)cnode;
+ while (iip < UBIFS_LPT_FANOUT) {
+ cn = nn->nbranch[iip].cnode;
+ if (cn) {
+ /* Go down */
+ row += 1;
+ col <<= UBIFS_LPT_FANOUT_SHIFT;
+ col += iip;
+ iip = 0;
+ cnode = cn;
+ break;
+ }
+ /* Go right */
+ iip += 1;
+ }
+ if (iip < UBIFS_LPT_FANOUT)
+ continue;
+ } else {
+ struct ubifs_pnode *pnode;
+
+ /* cnode is a pnode */
+ pnode = (struct ubifs_pnode *)cnode;
+ err = dbg_chk_pnode(c, pnode, col);
+ if (err)
+ return err;
+ }
+ /* Go up and to the right */
+ row -= 1;
+ col >>= UBIFS_LPT_FANOUT_SHIFT;
+ iip = cnode->iip + 1;
+ cnode = (struct ubifs_cnode *)nnode;
+ }
+ return 0;
+}
new file mode 100644
@@ -0,0 +1,1997 @@
+// 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 commit-related functionality of the LEB properties
+ * subsystem.
+ */
+
+#include <linux/crc16.h>
+#include <linux/slab.h>
+#include <linux/random.h>
+#include "ubifs.h"
+
+static int dbg_populate_lsave(struct ubifs_info *c);
+
+/**
+ * first_dirty_cnode - find first dirty cnode.
+ * @c: UBIFS file-system description object
+ * @nnode: nnode at which to start
+ *
+ * This function returns the first dirty cnode or %NULL if there is not one.
+ */
+static struct ubifs_cnode *first_dirty_cnode(const struct ubifs_info *c, struct ubifs_nnode *nnode)
+{
+ ubifs_assert(c, nnode);
+ while (1) {
+ int i, cont = 0;
+
+ for (i = 0; i < UBIFS_LPT_FANOUT; i++) {
+ struct ubifs_cnode *cnode;
+
+ cnode = nnode->nbranch[i].cnode;
+ if (cnode &&
+ test_bit(DIRTY_CNODE, &cnode->flags)) {
+ if (cnode->level == 0)
+ return cnode;
+ nnode = (struct ubifs_nnode *)cnode;
+ cont = 1;
+ break;
+ }
+ }
+ if (!cont)
+ return (struct ubifs_cnode *)nnode;
Import UBIFS libs from linux kernel. Next patches will replace ubifs related source code with implementation of linux kernel, which makes userspace implementation be same with linux kernel, then fsck.ubifs can resuse the code. Notice: lpt.c is modified with [1] applied. ubifs.h and orphan.c are modified with [2] applied, journal.c is modified with [3] reverted( because fsck runs in a single thread, so waitqueue is not needed to be implemented in userspace.). [1] https://lore.kernel.org/linux-mtd/20231228014112.2836317-13-chengzhihao1@huawei.com/ [2] https://lore.kernel.org/linux-mtd/20240410073751.2522830-1-chengzhihao1@huawei.com/ [3] https://lore.kernel.org/linux-mtd/20240122063103.359501-1-chengzhihao1@huawei.com/ Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com> --- ubifs-utils/libubifs/auth.c | 545 ++++++ ubifs-utils/libubifs/budget.c | 714 ++++++++ ubifs-utils/libubifs/commit.c | 733 ++++++++ ubifs-utils/libubifs/debug.c | 3051 +++++++++++++++++++++++++++++++ ubifs-utils/libubifs/debug.h | 304 ++++ ubifs-utils/libubifs/dir.c | 1744 ++++++++++++++++++ ubifs-utils/libubifs/find.c | 963 ++++++++++ ubifs-utils/libubifs/gc.c | 1017 +++++++++++ ubifs-utils/libubifs/io.c | 1268 +++++++++++++ ubifs-utils/libubifs/journal.c | 1928 ++++++++++++++++++++ ubifs-utils/libubifs/key.h | 543 ++++++ ubifs-utils/libubifs/log.c | 762 ++++++++ ubifs-utils/libubifs/lprops.c | 1307 ++++++++++++++ ubifs-utils/libubifs/lpt.c | 2451 +++++++++++++++++++++++++ ubifs-utils/libubifs/lpt_commit.c | 1997 +++++++++++++++++++++ ubifs-utils/libubifs/master.c | 473 +++++ ubifs-utils/libubifs/misc.h | 289 +++ ubifs-utils/libubifs/orphan.c | 947 ++++++++++ ubifs-utils/libubifs/recovery.c | 1588 +++++++++++++++++ ubifs-utils/libubifs/replay.c | 1250 +++++++++++++ ubifs-utils/libubifs/sb.c | 956 ++++++++++ ubifs-utils/libubifs/scan.c | 366 ++++ ubifs-utils/libubifs/super.c | 2505 ++++++++++++++++++++++++++ ubifs-utils/libubifs/tnc.c | 3553 +++++++++++++++++++++++++++++++++++++ ubifs-utils/libubifs/tnc_commit.c | 1111 ++++++++++++ ubifs-utils/libubifs/tnc_misc.c | 524 ++++++ ubifs-utils/libubifs/ubifs.h | 2164 ++++++++++++++++++++++ 27 files changed, 35053 insertions(+) create mode 100644 ubifs-utils/libubifs/auth.c create mode 100644 ubifs-utils/libubifs/budget.c create mode 100644 ubifs-utils/libubifs/commit.c create mode 100644 ubifs-utils/libubifs/debug.c create mode 100644 ubifs-utils/libubifs/debug.h create mode 100644 ubifs-utils/libubifs/dir.c create mode 100644 ubifs-utils/libubifs/find.c create mode 100644 ubifs-utils/libubifs/gc.c create mode 100644 ubifs-utils/libubifs/io.c create mode 100644 ubifs-utils/libubifs/journal.c create mode 100644 ubifs-utils/libubifs/key.h create mode 100644 ubifs-utils/libubifs/log.c create mode 100644 ubifs-utils/libubifs/lprops.c create mode 100644 ubifs-utils/libubifs/lpt.c create mode 100644 ubifs-utils/libubifs/lpt_commit.c create mode 100644 ubifs-utils/libubifs/master.c create mode 100644 ubifs-utils/libubifs/misc.h create mode 100644 ubifs-utils/libubifs/orphan.c create mode 100644 ubifs-utils/libubifs/recovery.c create mode 100644 ubifs-utils/libubifs/replay.c create mode 100644 ubifs-utils/libubifs/sb.c create mode 100644 ubifs-utils/libubifs/scan.c create mode 100644 ubifs-utils/libubifs/super.c create mode 100644 ubifs-utils/libubifs/tnc.c create mode 100644 ubifs-utils/libubifs/tnc_commit.c create mode 100644 ubifs-utils/libubifs/tnc_misc.c create mode 100644 ubifs-utils/libubifs/ubifs.h