Fix possible NULL pointer dereference in JBD2
Subsystems (3)
Jan Kara
reviewercover Fix possible NULL pointer dereference in JBD2
Hi,
the three patches below fix a bug in jbd2_journal_begin_ordered_truncate() that
could possibly lead to NULL pointer dereference. The question is how to merge the fix as it changes the prototype of the function and both ext4 and ocfs2 use it. Mark, Mingming, any suggestions?
Honza
Jan Kara
reviewerPATCH 1/3 jbd2: Fix possible NULL pointer 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 journal pointer (journal->j_list_lock) we don't have. So we have to change the prototype of the function so that filesystem passes us the journal pointer. Also add more detailed comment about why function does what it does how it should be used.
Thanks to Dan Carpenter <[email protected]> for pointing to the suspitious code.
--- fs/jbd2/transaction.c | 41 ++++++++++++++++++++++++++++++----------- include/linux/jbd2.h | 3 ++- 2 files changed, 32 insertions(+), 12 deletions(-)
fs/jbd2/transaction.c +30 -11
@@ -2129,26 +2129,45 @@ done: }/* -* This function must be called when inode is journaled in ordered mode -* before truncation happens. It starts writeout of truncated part in -* case it is in the committing transaction so that we stand to ordered -* mode consistency guarantees. +* File truncate and transaction commit interact with each other in a +* non-trivial way. If a transaction writing data block A is committing, +* we cannot discard the data by truncate until we have written them. +* Otherwise if we crashed after the transaction with write has committed +* but before the transaction with truncate has committed, we could see +* stale data in block A. This function is a helper to solve this problem. +* It starts writeout of the truncated part in case it is in the committing +* transaction. +* +* Filesystem must call this function when inode is journaled in ordered mode +* before truncation happens and after the inode has been placed on orphan list +* with the new inode size. The second condition avoids the race that someone +* writes new data and we start committing the transaction after this function +* has been called but before a transaction for truncate is started (and +* furthermore it allows us to optimize the case where addition to orphan list +* happens in the same transaction as write - we don't have to write any data +* in such case). */ -int jbd2_journal_begin_ordered_truncate (struct jbd2_inode *inode, +int jbd2_journal_begin_ordered_truncate (journal_t *journal, +struct jbd2_inode *jinode,loff_t new_size) { -journal_t *journal; -transaction_t *commit_trans; +transaction_t *inode_trans, *commit_trans;int ret =0 ; -if (!inode->i_transaction && !inode->i_next_transaction) +/* This is a quick check to avoid locking if not necessary */ +if (!jinode->i_transaction) goto out; - journal = inode->i_transaction->t_journal; +/* Locks are here just to force reading of recent values, it is +* enough that the transaction was not committing before we started +* a transaction adding the inode to orphan list */ spin_lock (&journal->j_state_lock); commit_trans = journal->j_committing_transaction;spin_unlock (&journal->j_state_lock); -if (inode->i_transaction == commit_trans) { - ret =filemap_fdatawrite_range (inode->i_vfs_inode->i_mapping, +spin_lock (&journal->j_list_lock); + inode_trans = jinode->i_transaction; +spin_unlock (&journal->j_list_lock); +if (inode_trans == commit_trans) { + ret =filemap_fdatawrite_range (jinode->i_vfs_inode->i_mapping, new_size,LLONG_MAX );if (ret)jbd2_journal_abort (journal, ret);
include/linux/jbd2.h +2 -1
@@ -1150,7 +1150,8 @@ extern int jbd2_journal_clear_err (journal_t *);extern int jbd2_journal_bmap (journal_t *,unsigned long ,unsigned long long *);extern int jbd2_journal_force_commit (journal_t *);extern int jbd2_journal_file_inode (handle_t *handle,struct jbd2_inode *inode); -extern int jbd2_journal_begin_ordered_truncate (struct jbd2_inode *inode,loff_t new_size); +extern int jbd2_journal_begin_ordered_truncate (journal_t *journal, +struct jbd2_inode *inode,loff_t new_size);extern void jbd2_journal_init_jbd_inode (struct jbd2_inode *jinode,struct inode *inode);extern void jbd2_journal_release_jbd_inode (journal_t *journal,struct jbd2_inode *jinode);
Signature · 1 line
1.6.0.2
Jan Kara
reviewerPATCH 2/3 ext4: Add journal parameter to jbd2_journal_begin_ordered_truncate()
Add new parameter of the function.
--- fs/ext4/inode.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-)
fs/ext4/inode.c +4 -2
@@ -47,8 +47,10 @@static inline int ext4_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&EXT4_I (inode)->jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +EXT4_SB (inode->i_sb)->s_journal, + &EXT4_I (inode)->jinode, + new_size); }static void ext4_invalidatepage (struct page *page,unsigned long offset);
Signature · 1 line
1.6.0.2
Jan Kara
reviewerPATCH 3/3 ocfs2: Add journal parameter to jbd2_journal_begin_ordered_truncate()
Add new parameter of the function.
--- fs/ocfs2/journal.h | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-)
fs/ocfs2/journal.h +4 -2
@@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)static inline int ocfs2_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&OCFS2_I (inode)->ip_jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +OCFS2_SB (inode->i_sb)->journal->j_journal, + &OCFS2_I (inode)->ip_jinode, + new_size); }#endif /* OCFS2_JOURNAL_H */
Signature · 1 line
1.6.0.2
Re: [Ocfs2-devel] [PATCH 3/3] ocfs2: Add journal parameter to jbd2_journal_begin_ordered_truncate()
Quoted 5 lines · Jan Kara
Add new parameter of the function.
Quoted 29 lines
--- fs/ocfs2/journal.h | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-)fs/ocfs2/journal.h +4
-2@@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)static inline int ocfs2_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&OCFS2_I (inode)->ip_jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +OCFS2_SB (inode->i_sb)->journal->j_journal, + &OCFS2_I (inode)->ip_jinode, + new_size); }#endif /* OCFS2_JOURNAL_H */ Signature · 7 lines
1.6.0.2 _______________________________________________ Ocfs2-devel mailing list Ocfs2-devel at oss.oracle.com http://oss.oracle.com/mailman/listinfo/ocfs2-devel
Signature · 10 lines
Life's Little Instruction Book #464 "Don't miss the magic of the moment by focusing on what's to come." Joel Becker Principal Software Developer Oracle E-mail: joel.becker at oracle.com Phone: (650) 506-8127
Re: [PATCH 3/3] ocfs2: Add journal parameter to jbd2_journal_begin_ordered_truncate()
In reply to Jan KaraThe problem with separating out the patch into three separate commits is that it can potentially break a "git bisect". To fix this, I believe the right thing to do is to combine the three patches into a single commit, and to push this via the ext4 tree; any objections?
- Ted
jbd2: Avoid possible NULL dereference in jbd2_journal_begin_ordered_truncate()
From: Jan Kara <[email protected]>
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 <[email protected]> for pointing to the suspitious code.
--- fs/ext4/inode.c | 6 ++++-- fs/jbd2/transaction.c | 42 +++++++++++++++++++++++++++++++----------- fs/ocfs2/journal.h | 6 ++++-- include/linux/jbd2.h | 3 ++- 4 files changed, 41 insertions(+), 16 deletions(-)
fs/ext4/inode.c +4 -2
@@ -47,8 +47,10 @@static inline int ext4_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&EXT4_I (inode)->jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +EXT4_SB (inode->i_sb)->s_journal, + &EXT4_I (inode)->jinode, + new_size); }static void ext4_invalidatepage (struct page *page,unsigned long offset);
fs/jbd2/transaction.c +31 -11
57 lines not on this page · raw message
fs/ocfs2/journal.h +4 -2
@@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)static inline int ocfs2_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&OCFS2_I (inode)->ip_jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +OCFS2_SB (inode->i_sb)->journal->j_journal, + &OCFS2_I (inode)->ip_jinode, + new_size); }#endif /* OCFS2_JOURNAL_H */
include/linux/jbd2.h +2 -1
@@ -1150,7 +1150,8 @@ extern int jbd2_journal_clear_err (journal_t *);extern int jbd2_journal_bmap (journal_t *,unsigned long ,unsigned long long *);extern int jbd2_journal_force_commit (journal_t *);extern int jbd2_journal_file_inode (handle_t *handle,struct jbd2_inode *inode); -extern int jbd2_journal_begin_ordered_truncate (struct jbd2_inode *inode,loff_t new_size); +extern int jbd2_journal_begin_ordered_truncate (journal_t *journal, +struct jbd2_inode *inode,loff_t new_size);extern void jbd2_journal_init_jbd_inode (struct jbd2_inode *jinode,struct inode *inode);extern void jbd2_journal_release_jbd_inode (journal_t *journal,struct jbd2_inode *jinode);
Jan Kara
reviewerRe: [PATCH 3/3] ocfs2: Add journal parameter to jbd2_journal_begin_ordered_truncate()
Quoted 4 lines · Theodore Tso
The problem with separating out the patch into three separate commits is that it can potentially break a "git bisect". To fix this, I believe the right thing to do is to combine the three patches into a single commit, and to push this via the ext4 tree; any objections?
On a second thought, this is probably the best for such simple change.
So please go ahead. Thanks.
Honza
Quoted 138 lines
jbd2: Avoid possible NULL dereference in
jbd2_journal_begin_ordered_truncate()From: Jan Kara <[email protected]>
If we race with commit code setting
i_transactionto NULL, we could possibly dereference it. Proper locking requires the journal pointer (to accessjournal->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 functionjbd2_journal_begin_ordered_truncate()does what it does and how it should be used.Thanks to Dan Carpenter <[email protected]> for pointing to the suspitious code.
Signed-off-by Jan Kara <[email protected]>Signed-off-by "Theodore Ts'o" <[email protected]>Acked-by Joel Becker <[email protected]>CC linux-ext4 at vger.kernel.orgCC ocfs2-devel at oss.oracle.comCC mfasheh at suse.deCC Dan Carpenter <[email protected]>--- fs/ext4/inode.c | 6 ++++-- fs/jbd2/transaction.c | 42 +++++++++++++++++++++++++++++++----------- fs/ocfs2/journal.h | 6 ++++-- include/linux/jbd2.h | 3 ++- 4 files changed, 41 insertions(+), 16 deletions(-)fs/ext4/inode.c +4
-2@@ -47,8 +47,10 @@static inline int ext4_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&EXT4_I (inode)->jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +EXT4_SB (inode->i_sb)->s_journal, + &EXT4_I (inode)->jinode, + new_size); }static void ext4_invalidatepage (struct page *page,unsigned long offset);fs/jbd2/transaction.c +31
-11@@ -2129,26 +2129,46 @@ done: }/* -* This function must be called when inode is journaled in ordered mode -* before truncation happens. It starts writeout of truncated part in -* case it is in the committing transaction so that we stand to ordered -* mode consistency guarantees. +* File truncate and transaction commit interact with each other in a +* non-trivial way. If a transaction writing data block A is +* committing, we cannot discard the data by truncate until we have +* written them. Otherwise if we crashed after the transaction with +* write has committed but before the transaction with truncate has +* committed, we could see stale data in block A. This function is a +* helper to solve this problem. It starts writeout of the truncated +* part in case it is in the committing transaction. +* +* Filesystem code must call this function when inode is journaled in +* ordered mode before truncation happens and after the inode has been +* placed on orphan list with the new inode size. The second condition +* avoids the race that someone writes new data and we start +* committing the transaction after this function has been called but +* before a transaction for truncate is started (and furthermore it +* allows us to optimize the case where the addition to orphan list +* happens in the same transaction as write --- we don't have to write +* any data in such case). */ -int jbd2_journal_begin_ordered_truncate (struct jbd2_inode *inode, +int jbd2_journal_begin_ordered_truncate (journal_t *journal, +struct jbd2_inode *jinode,loff_t new_size) { -journal_t *journal; -transaction_t *commit_trans; +transaction_t *inode_trans, *commit_trans;int ret =0 ; -if (!inode->i_transaction && !inode->i_next_transaction) +/* This is a quick check to avoid locking if not necessary */ +if (!jinode->i_transaction) goto out; - journal = inode->i_transaction->t_journal; +/* Locks are here just to force reading of recent values, it is +* enough that the transaction was not committing before we started +* a transaction adding the inode to orphan list */ spin_lock (&journal->j_state_lock); commit_trans = journal->j_committing_transaction;spin_unlock (&journal->j_state_lock); -if (inode->i_transaction == commit_trans) { - ret =filemap_fdatawrite_range (inode->i_vfs_inode->i_mapping, +spin_lock (&journal->j_list_lock); + inode_trans = jinode->i_transaction; +spin_unlock (&journal->j_list_lock); +if (inode_trans == commit_trans) { + ret =filemap_fdatawrite_range (jinode->i_vfs_inode->i_mapping, new_size,LLONG_MAX );if (ret)jbd2_journal_abort (journal, ret);fs/ocfs2/journal.h +4
-2@@ -513,8 +513,10 @@ static inline int ocfs2_jbd2_file_inode(handle_t *handle, struct inode *inode)static inline int ocfs2_begin_ordered_truncate (struct inode *inode,loff_t new_size) { -return jbd2_journal_begin_ordered_truncate (&OCFS2_I (inode)->ip_jinode, - new_size); +return jbd2_journal_begin_ordered_truncate ( +OCFS2_SB (inode->i_sb)->journal->j_journal, + &OCFS2_I (inode)->ip_jinode, + new_size); }#endif /* OCFS2_JOURNAL_H */ include/linux/jbd2.h +2
-1@@ -1150,7 +1150,8 @@ extern int jbd2_journal_clear_err (journal_t *);extern int jbd2_journal_bmap (journal_t *,unsigned long ,unsigned long long *);extern int jbd2_journal_force_commit (journal_t *);extern int jbd2_journal_file_inode (handle_t *handle,struct jbd2_inode *inode); -extern int jbd2_journal_begin_ordered_truncate (struct jbd2_inode *inode,loff_t new_size); +extern int jbd2_journal_begin_ordered_truncate (journal_t *journal, +struct jbd2_inode *inode,loff_t new_size);extern void jbd2_journal_init_jbd_inode (struct jbd2_inode *jinode,struct inode *inode);extern void jbd2_journal_release_jbd_inode (journal_t *journal,struct jbd2_inode *jinode);
Signature · 2 lines
Jan Kara <[email protected]> SUSE Labs, CR
Mark Fasheh
Acked-byRe: [PATCH 3/3] ocfs2: Add journal parameter to jbd2_journal_begin_ordered_truncate()
In reply to Theodore TsoQuoted 4 lines · Theodore Tso
The problem with separating out the patch into three separate commits is that it can potentially break a "git bisect". To fix this, I believe the right thing to do is to combine the three patches into a single commit, and to push this via the ext4 tree; any objections?
That's fine with me. Btw, you forgot my:
--Mark
--
Mark Fasheh
Quoted 7 lines · Jan Kara
Hi,
the three patches below fix a bug in
jbd2_journal_begin_ordered_truncate()that
could possibly lead to NULL pointer dereference. The question is how to merge the fix as it changes the prototype of the function and both ext4 and ocfs2 use it. Mark, Mingming, any suggestions?
This seems like it'd be easiest for Andrew to carry. The Ocfs2 patch is unlikely to conflict with anything in my tree.
Btw, you can add:
to the Ocfs2 patch.
--Mark
--
Mark Fasheh
Theodore Tso
maintainerQuoted 7 lines · Jan Kara
the three patches below fix a bug in
jbd2_journal_begin_ordered_truncate()that could possibly lead to NULL pointer dereference. The question is how to merge the fix as it changes the prototype of the function and both ext4 and ocfs2 use it. Mark, Mingming, any suggestions?
Jan, how how easy is it to hit this? Do you believe we should try to get this in for 2.6.29? I got some final things to push to Linus so I can also push it via the ext4 tree. I doubt it will conflict with
Quoted 1 line
anything in my tree, either.
- Ted
Jan Kara
reviewerQuoted 9 lines · Theodore Tso
Quoted 7 lines · Jan Kara
the three patches below fix a bug in
jbd2_journal_begin_ordered_truncate()that could possibly lead to NULL pointer dereference. The question is how to merge the fix as it changes the prototype of the function and both ext4 and ocfs2 use it. Mark, Mingming, any suggestions?
Quoted 4 lines
Jan, how how easy is it to hit this? Do you believe we should try to get this in for 2.6.29? I got some final things to push to Linus so I can also push it via the ext4 tree. I doubt it will conflict with
Quoted 1 line
anything in my tree, either.
The possibility of hitting this is quite small. We checkinode->i_transaction for NULL just before dereferencing it so the commit would have to set it to NULL just between those two instructions.
Honza
Signature · 2 lines
Jan Kara <[email protected]> SUSE Labs, CR