aboutsummaryrefslogtreecommitdiffstats
path: root/fs/jbd2/transaction.c
diff options
context:
space:
mode:
authorJan Kara <jack@suse.cz>2009-02-10 11:15:34 -0500
committerTheodore Tso <tytso@mit.edu>2009-02-10 11:15:34 -0500
commit7f5aa215088b817add9c71914b83650bdd49f8a9 (patch)
tree9b811d1f0b41a2738fc68654ae605bf5d8cf2a72 /fs/jbd2/transaction.c
parent9eddacf9e9c03578ef2c07c9534423e823d677f8 (diff)
jbd2: Avoid possible NULL dereference in jbd2_journal_begin_ordered_truncate()
If we race with commit code setting i_transaction to NULL, we could possibly dereference it. Proper locking requires the journal pointer (to access journal->j_list_lock), which we don't have. So we have to change the prototype of the function so that filesystem passes us the journal pointer. Also add a more detailed comment about why the function jbd2_journal_begin_ordered_truncate() does what it does and how it should be used. Thanks to Dan Carpenter <error27@gmail.com> for pointing to the suspitious code. Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Acked-by: Joel Becker <joel.becker@oracle.com> CC: linux-ext4@vger.kernel.org CC: ocfs2-devel@oss.oracle.com CC: mfasheh@suse.de CC: Dan Carpenter <error27@gmail.com>
Diffstat (limited to 'fs/jbd2/transaction.c')
-rw-r--r--fs/jbd2/transaction.c42
1 files changed, 31 insertions, 11 deletions
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index 46b4e347ed7d..28ce21d8598e 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -2129,26 +2129,46 @@ done:
2129} 2129}
2130 2130
2131/* 2131/*
2132 * This function must be called when inode is journaled in ordered mode 2132 * File truncate and transaction commit interact with each other in a
2133 * before truncation happens. It starts writeout of truncated part in 2133 * non-trivial way. If a transaction writing data block A is
2134 * case it is in the committing transaction so that we stand to ordered 2134 * committing, we cannot discard the data by truncate until we have
2135 * mode consistency guarantees. 2135 * written them. Otherwise if we crashed after the transaction with
2136 * write has committed but before the transaction with truncate has
2137 * committed, we could see stale data in block A. This function is a
2138 * helper to solve this problem. It starts writeout of the truncated
2139 * part in case it is in the committing transaction.
2140 *
2141 * Filesystem code must call this function when inode is journaled in
2142 * ordered mode before truncation happens and after the inode has been
2143 * placed on orphan list with the new inode size. The second condition
2144 * avoids the race that someone writes new data and we start
2145 * committing the transaction after this function has been called but
2146 * before a transaction for truncate is started (and furthermore it
2147 * allows us to optimize the case where the addition to orphan list
2148 * happens in the same transaction as write --- we don't have to write
2149 * any data in such case).
2136 */ 2150 */
2137int jbd2_journal_begin_ordered_truncate(struct jbd2_inode *inode, 2151int jbd2_journal_begin_ordered_truncate(journal_t *journal,
2152 struct jbd2_inode *jinode,
2138 loff_t new_size) 2153 loff_t new_size)
2139{ 2154{
2140 journal_t *journal; 2155 transaction_t *inode_trans, *commit_trans;
2141 transaction_t *commit_trans;
2142 int ret = 0; 2156 int ret = 0;
2143 2157
2144 if (!inode->i_transaction && !inode->i_next_transaction) 2158 /* This is a quick check to avoid locking if not necessary */
2159 if (!jinode->i_transaction)
2145 goto out; 2160 goto out;
2146 journal = inode->i_transaction->t_journal; 2161 /* Locks are here just to force reading of recent values, it is
2162 * enough that the transaction was not committing before we started
2163 * a transaction adding the inode to orphan list */
2147 spin_lock(&journal->j_state_lock); 2164 spin_lock(&journal->j_state_lock);
2148 commit_trans = journal->j_committing_transaction; 2165 commit_trans = journal->j_committing_transaction;
2149 spin_unlock(&journal->j_state_lock); 2166 spin_unlock(&journal->j_state_lock);
2150 if (inode->i_transaction == commit_trans) { 2167 spin_lock(&journal->j_list_lock);
2151 ret = filemap_fdatawrite_range(inode->i_vfs_inode->i_mapping, 2168 inode_trans = jinode->i_transaction;
2169 spin_unlock(&journal->j_list_lock);
2170 if (inode_trans == commit_trans) {
2171 ret = filemap_fdatawrite_range(jinode->i_vfs_inode->i_mapping,
2152 new_size, LLONG_MAX); 2172 new_size, LLONG_MAX);
2153 if (ret) 2173 if (ret)
2154 jbd2_journal_abort(journal, ret); 2174 jbd2_journal_abort(journal, ret);