Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Now we're going to journal such modifications within the ocfs2_orphan_add() itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.
--- fs/ocfs2/namei.c | 28 +++++++++++++++++++++++----- 1 files changed, 23 insertions(+), 5 deletions(-)
fs/ocfs2/namei.c +23 -5
@@ -84,7 +84,7 @@ static int ocfs2_prepare_orphan_dir(struct ocfs2_super *osb,static int ocfs2_orphan_add (struct ocfs2_super *osb,handle_t *handle,struct inode *inode, -struct ocfs2_dinode *fe, +struct buffer_head *fe_bh,char *name,struct ocfs2_dir_lookup_result *lookup,struct inode *orphan_dir_inode);
@@ -877,7 +877,7 @@ static int ocfs2_unlink(struct inode *dir, fe = (struct ocfs2_dinode *) fe_bh->b_data;if (inode_is_unlinkable (inode)) { - status =ocfs2_orphan_add (osb, handle, inode, fe, orphan_name, + status =ocfs2_orphan_add (osb, handle, inode, fe_bh, orphan_name, &orphan_insert, orphan_dir);if (status <0 ) {mlog_errno (status);
@@ -1295,7 +1295,7 @@ static int ocfs2_rename(struct inode *old_dir,if (S_ISDIR (new_inode->i_mode) || (ocfs2_read_links_count (newfe) ==1 )) { status =ocfs2_orphan_add (osb, handle, new_inode, - newfe, orphan_name, + newfe_bh, orphan_name, &orphan_insert, orphan_dir);if (status <0 ) {mlog_errno (status);
@@ -1909,7 +1909,7 @@ leave:static int ocfs2_orphan_add (struct ocfs2_super *osb,handle_t *handle,struct inode *inode, -struct ocfs2_dinode *fe, +struct buffer_head *fe_bh,char *name,struct ocfs2_dir_lookup_result *lookup,struct inode *orphan_dir_inode)
@@ -1917,6 +1917,7 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,struct buffer_head *orphan_dir_bh =NULL ;int status =0 ;struct ocfs2_dinode *orphan_fe; +struct ocfs2_dinode *fe = (struct ocfs2_dinode *) fe_bh->b_data;mlog_entry ("(inode->i_ino = %lu)\n" , inode->i_ino);
@@ -1957,6 +1958,21 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb, goto leave; } +/* +* We're going to journal the change of i_flags and i_orphaned_slot. +* It's safe anyway, though some callers may duplicate the journaling. +* Journaling within the func just make the logic look more +* straightforward. +*/ + status =ocfs2_journal_access_di (handle, +INODE_CACHE (inode), + fe_bh, +OCFS2_JOURNAL_ACCESS_WRITE ); +if (status <0 ) { +mlog_errno (status); + goto leave; + } +le32_add_cpu (&fe->i_flags,OCFS2_ORPHANED_FL );/* Record which orphan dir our inode now resides
@@ -1964,6 +1980,8 @@ static int ocfs2_orphan_add(struct ocfs2_super *osb,* dir to lock. */ fe->i_orphaned_slot =cpu_to_le16 (osb->slot_num); +ocfs2_journal_dirty (handle, fe_bh); +mlog (0 ,"Inode %llu orphaned in slot %d\n" , (unsigned long long )OCFS2_I (inode)->ip_blkno, osb->slot_num);
@@ -2125,7 +2143,7 @@ int ocfs2_create_inode_in_orphan(struct inode *dir, } di = (struct ocfs2_dinode *)new_di_bh->b_data; - status =ocfs2_orphan_add (osb, handle, inode, di, orphan_name, + status =ocfs2_orphan_add (osb, handle, inode, new_di_bh, orphan_name, &orphan_insert, orphan_dir);if (status <0 ) {mlog_errno (status);
Signature · 1 line
1.5.5
PATCH 2/2 Ocfs2: Handle deletion of reflinked oprhan inodes correctly.
Current rule of orphan dir is that all inodes in the orphan dir have ORPHANED_FL, otherwise we treated it as an ERROR. this rule works well except for some rare cases of reflink operation:
http://oss.oracle.com/bugzilla/show_bug.cgi?id=1215
The problem is introduced by the essense of how reflink and our orphan_scan thread were working:
* Orphan_scan scan the orphan dir into a queue first, and run queue in a later time, we only hold the orphan_dir's lock during scanning.
* Reflink create a oprhaned target in orphan_dir at the first step, and remove the targets and unset the flag at the third step, these two steps respectively hold the orphan_dir's lock themselves.
Based on above semantics, there is a possibility that a reflink inode can be moved out of the orphan dir and have its ORPHANED_FL cleared before the queue is run, which leads to a ERROR in ocfs2_query_wipde_inode().
This patch helps to judge if a orphan inode to be wiped off, which has NO ORPHANED_FL, is a legal alive reflinked target or not.
The patch also works for failed reflinked targets from a crash or other failures during the reflink operation, they can be wiped off as desired since these failed reflinked inodes always has ORPHANED_FL set ondisk.
--- fs/ocfs2/inode.c | 15 +++++++++++++++ 1 files changed, 15 insertions(+), 0 deletions(-)
fs/ocfs2/inode.c +15 -0
@@ -891,6 +891,21 @@ static int ocfs2_query_inode_wipe(struct inode *inode,/* Do some basic inode verification... */ di = (struct ocfs2_dinode *) di_bh->b_data;if (!(di->i_flags &cpu_to_le32 (OCFS2_ORPHANED_FL ))) { +/* +* Inodes in the orphan dir must have ORPHANED_FL. The only +* inodes that come back out of the orphan dir are reflink +* targets. A reflink target may be moved out of the orphan +* dir between the time we scan the directory and the time we +* process it. This would lead to HAS_REFCOUNT_FL being set but +* ORPHANED_FL not. +*/ +if (di->i_dyn_features &cpu_to_le16 (OCFS2_HAS_REFCOUNT_FL )) { +mlog (0 ,"Reflinked inode %llu is no longer orphaned. " +"it shouldn't be deleted\n" , + (unsigned long long )oi->ip_blkno); + goto bail; + } +/* for lack of a better error? */ status = -EEXIST ; mlog(ML_ERROR,
Signature · 1 line
1.5.5
PATCH 2/2 Ocfs2: Handle deletion of reflinked oprhan inodes correctly.
Quoted 23 lines · Tristan Ye
Current rule of orphan dir is that all inodes in the orphan dir have
ORPHANED_FL, otherwise we treated it as an ERROR. this rule works well except for some rare cases of reflink operation:http://oss.oracle.com/bugzilla/show_bug.cgi?id=1215
The problem is introduced by the essense of how reflink and our
orphan_scanthread were working:*
Orphan_scanscan the orphan dir into a queue first, and run queue in a later time, we only hold the orphan_dir's lock during scanning.* Reflink create a oprhaned target in
orphan_dirat the first step, and remove the targets and unset the flag at the third step, these two steps respectively hold the orphan_dir's lock themselves.Based on above semantics, there is a possibility that a reflink inode can be moved out of the orphan dir and have its
ORPHANED_FLcleared before the queue is run, which leads to a ERROR inocfs2_query_wipde_inode().
Quoted 7 lines
This patch helps to judge if a orphan inode to be wiped off, which has NO
ORPHANED_FL, is a legal alive reflinked target or not.The patch also works for failed reflinked targets from a crash or other failures during the reflink operation, they can be wiped off as desired since these failed reflinked inodes always has
ORPHANED_FLset ondisk.
How is this? Wouldn't the failed reflink still have OCFS2_HAS_REFCOUNT_FL set as well as OCFS2_ORPHANED_FL? In the code below, we (correctly) skip those for delete.
Btw, other than my question above, the patch looks correct for the case of not wiping a reflinked inode.
--Mark
--
Mark Fasheh
PATCH 2/2 Ocfs2: Handle deletion of reflinked oprhan inodes correctly.
Quoted 27 lines · Mark Fasheh
Quoted 26 lines · Tristan Ye
Current rule of orphan dir is that all inodes in the orphan dir have
ORPHANED_FL, otherwise we treated it as an ERROR. this rule works well except for some rare cases of reflink operation:http://oss.oracle.com/bugzilla/show_bug.cgi?id=1215
The problem is introduced by the essense of how reflink and our
orphan_scanthread were working:*
Orphan_scanscan the orphan dir into a queue first, and run queue in a later time, we only hold the orphan_dir's lock during scanning.* Reflink create a oprhaned target in
orphan_dirat the first step, and remove the targets and unset the flag at the third step, these two steps respectively hold the orphan_dir's lock themselves.Based on above semantics, there is a possibility that a reflink inode can be moved out of the orphan dir and have its
ORPHANED_FLcleared before the queue is run, which leads to a ERROR inocfs2_query_wipde_inode().This patch helps to judge if a orphan inode to be wiped off, which has NO
ORPHANED_FL, is a legal alive reflinked target or not.
Quoted 8 lines
Quoted 4 lines
The patch also works for failed reflinked targets from a crash or other failures during the reflink operation, they can be wiped off as desired since these failed reflinked inodes always has
ORPHANED_FLset ondisk.How is this? Wouldn't the failed reflink still have
OCFS2_HAS_REFCOUNT_FLset as well asOCFS2_ORPHANED_FL? In the code below, we (correctly) skip those for delete.
A failed reflink indeed has ORPHANED_FL. Thus the if block is
skipped and we continue to the rest of the query_wipe checks.
Joel
Signature · 9 lines
"Under capitalism, man exploits man. Under Communism, it's just the opposite." - John Kenneth Galbraith Joel Becker Principal Software Developer Oracle E-mail: joel.becker at oracle.com Phone: (650) 506-8127
PATCH 2/2 Ocfs2: Handle deletion of reflinked oprhan inodes correctly.
In reply to Mark FashehQuoted 34 lines · Mark Fasheh
Quoted 23 lines · Tristan Ye
Current rule of orphan dir is that all inodes in the orphan dir have
ORPHANED_FL, otherwise we treated it as an ERROR. this rule works well except for some rare cases of reflink operation:http://oss.oracle.com/bugzilla/show_bug.cgi?id=1215
The problem is introduced by the essense of how reflink and our
orphan_scanthread were working:*
Orphan_scanscan the orphan dir into a queue first, and run queue in a later time, we only hold the orphan_dir's lock during scanning.* Reflink create a oprhaned target in
orphan_dirat the first step, and remove the targets and unset the flag at the third step, these two steps respectively hold the orphan_dir's lock themselves.Based on above semantics, there is a possibility that a reflink inode can be moved out of the orphan dir and have its
ORPHANED_FLcleared before the queue is run, which leads to a ERROR inocfs2_query_wipde_inode().Quoted 8 lines
This patch helps to judge if a orphan inode to be wiped off, which has NO
ORPHANED_FL, is a legal alive reflinked target or not.The patch also works for failed reflinked targets from a crash or other failures during the reflink operation, they can be wiped off as desired since these failed reflinked inodes always has
ORPHANED_FLset ondisk.
Quoted 4 lines
How is this? Wouldn't the failed reflink still have
OCFS2_HAS_REFCOUNT_FLset as well asOCFS2_ORPHANED_FL? In the code below, we (correctly) skip those for delete.
Mark,
We firstly check the flag ORPHANED_FL, if ORPHANED_FL were already there, no need to check REFCOUNT_FL, it will be definitely wiped off since we know for sure it's a deleted inode or a failed reflink inode. if ORPHANED_FL were not set. there may be 2 cases, one is a unknown error somehow which caused a deleted inode in orphan_dir having its ORPHAN_FL unset(that way, we need to bailout with a error). and the other case is all what our patch described: it's a legal reflink inode(in a incomplete state),which should not be treated as an error,and also can not be wiped off as well.
Then you may ask, how about the case when the reflink inode has ORPHANED_FL, but without REFCOUNT_FT been set yet. will our ocfs2_delete() codes incorrectly delete them? the answer is definitely no, since reflink operation hold the inode lock all its operating time except a failure.
Regards,
Tristan.
Quoted 8 lines
Btw, other than my question above, the patch looks correct for the case of not wiping a reflinked inode.
--Mark--
Mark Fasheh
PATCH 2/2 Ocfs2: Handle deletion of reflinked oprhan inodes correctly.
In reply to Tristan YeQuoted 26 lines · Tristan Ye
Current rule of orphan dir is that all inodes in the orphan dir have
ORPHANED_FL, otherwise we treated it as an ERROR. this rule works well except for some rare cases of reflink operation:http://oss.oracle.com/bugzilla/show_bug.cgi?id=1215
The problem is introduced by the essense of how reflink and our
orphan_scanthread were working:*
Orphan_scanscan the orphan dir into a queue first, and run queue in a later time, we only hold the orphan_dir's lock during scanning.* Reflink create a oprhaned target in
orphan_dirat the first step, and remove the targets and unset the flag at the third step, these two steps respectively hold the orphan_dir's lock themselves.Based on above semantics, there is a possibility that a reflink inode can be moved out of the orphan dir and have its
ORPHANED_FLcleared before the queue is run, which leads to a ERROR inocfs2_query_wipde_inode().This patch helps to judge if a orphan inode to be wiped off, which has NO
ORPHANED_FL, is a legal alive reflinked target or not.
Quoted 6 lines
The patch also works for failed reflinked targets from a crash or other failures during the reflink operation, they can be wiped off as desired since these failed reflinked inodes always has
ORPHANED_FLset ondisk.Signed-off-by Tristan Ye <[email protected]>
This patch has been added to the 'fixes' branch of ocfs2.git.
Joel
Signature · 9 lines
Life's Little Instruction Book #407 "Every once in a while, take the scenic route." Joel Becker Principal Software Developer Oracle E-mail: joel.becker at oracle.com Phone: (650) 506-8127
PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
In reply to Tristan YeQuoted 8 lines · Tristan Ye
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Now we're going to journal such modifications within the
ocfs2_orphan_add()itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.Signed-off-by Tristan Ye <[email protected]>
I would have liked to see if any of the callers could be cleaned up. Otherwise though, this patch looks good and it fixes a bug :)
--Mark
--
Mark Fasheh
PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
Quoted 4 lines · Mark Fasheh
Quoted 3 lines · Tristan Ye
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Quoted 8 lines
Quoted 5 lines
Now we're going to journal such modifications within the
ocfs2_orphan_add()itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.Signed-off-by Tristan Ye <[email protected]>I would have liked to see if any of the callers could be cleaned up. Otherwise though, this patch looks good and it fixes a bug :)
Let's call this the fix and expect the followup.
Joel
Signature · 8 lines
"Anything that is too stupid to be spoken is sung."
- Voltaire
Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
In reply to Tristan YeQuoted 8 lines · Tristan Ye
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Now we're going to journal such modifications within the
ocfs2_orphan_add()itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.Signed-off-by Tristan Ye <[email protected]>
This patch has been added to the 'fixes' branch of ocfs2.git.
Can you go over the callers of ocfs2_orphan_add() to see if there is any cleanup you can do? Redundant dirty calls where nothing else is dirtied, etc. That cleanup will be a separate patch that I push to the merge window when done.
Joel
Signature · 9 lines
Joel's First Law: Nature abhors a GUI. Joel Becker Principal Software Developer Oracle E-mail: joel.becker at oracle.com Phone: (650) 506-8127
PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
Quoted 9 lines · Joel Becker
Quoted 7 lines · Tristan Ye
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Now we're going to journal such modifications within the
ocfs2_orphan_add()itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.
Quoted 9 lines
Quoted 2 lines
Signed-off-by Tristan Ye <[email protected]>This patch has been added to the 'fixes' branch of ocfs2.git.
Can you go over the callers ofocfs2_orphan_add()to see if there is any cleanup you can do? Redundant dirty calls where nothing else is dirtied, etc. That cleanup will be a separate patch that I push to the merge window when done.
Sure, you're right, redundant dirty calls doesn't make any sense, and even makes the logic confused.
I'll handle this.
Thanks,
Tristan.
Quoted 3 lines
Joel
PATCH 1/2 Ocfs2: Journaling i_flags and i_orphaned_slot when adding inode to orphan dir.
In reply to Joel BeckerQuoted 8 lines · Joel Becker
Quoted 7 lines · Tristan Ye
Currently, some callers were missing to journal the dirty inode after adding it to orphan dir.
Now we're going to journal such modifications within the
ocfs2_orphan_add()itself, It's safe to do so, though some existing caller may duplicate this, and it makes the logic look more straightforward anyway.
Quoted 7 lines
Quoted 1 line
Signed-off-by Tristan Ye <[email protected]>This patch has been added to the 'fixes' branch of ocfs2.git.
Can you go over the callers ofocfs2_orphan_add()to see if there is any cleanup you can do? Redundant dirty calls where nothing else is dirtied, etc. That cleanup will be a separate patch that I push to the merge window when done.
3 callers are calling ocfs2_orphan_add() for now:
1. ocfs2_unlink()
2. ocfs2_rename()
3. ocfs2_create_inode_in_orphan()
caller #1 and #2 are also doing dirty journals, they however are not redundant, just for i_links_count change on-disk.
while #3 caller is doing nothing dirty journal outside the ocfs2_oprhan_add(), it's fine.
As in, we're not going to do any cleanup in the callers.
Regards,
Tristan.
Quoted 3 lines
Joel