Skip to contentSkip to conversation

ocfs2: use retval instead of status for checking error

PATCH In review Daeseok Youn 4 participants 12 messages Feb 27, 2015 – Mar 24, 2015 OCFS2

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/4 ocfs2: use retval instead of status for checking error

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
fs/ocfs2/dir.c +6 -6
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
 	struct ocfs2_dir_entry *de, *de1;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
 	struct super_block *sb = dir->i_sb;
-	int retval, status;
+	int retval;
 	unsigned int size = sb->s_blocksize;
 	struct buffer_head *insert_bh = lookup->dl_leaf_bh;
 	char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
 			}
 
 			if (insert_bh == parent_fe_bh)
-				status = ocfs2_journal_access_di(handle,
+				retval = ocfs2_journal_access_di(handle,
 								 INODE_CACHE(dir),
 								 insert_bh,
 								 OCFS2_JOURNAL_ACCESS_WRITE);
 			else {
-				status = ocfs2_journal_access_db(handle,
+				retval = ocfs2_journal_access_db(handle,
 								 INODE_CACHE(dir),
 								 insert_bh,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 
 				if (ocfs2_dir_indexed(dir)) {
-					status = ocfs2_dx_dir_insert(dir,
+					retval = ocfs2_dx_dir_insert(dir,
 								handle,
 								lookup);
-					if (status) {
-						mlog_errno(status);
+					if (retval) {
+						mlog_errno(retval);
 						goto bail;
 					}
 				}
Signature · 1 line
1.7.1
#1 lore raw

Am 28.02.2015 um 00:48 schrieb Daeseok Youn:

Quoted 9 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

As this patch is untested and the issue is theoretical I'm nervous. But the final decision is up to ocfs2 maintainers.

Thanks,
//richard

#2 lore raw

Hi, all

please, review this patch.

thanks.

regards,
Daeseok Youn.

2015-03-02 18:04 GMT+09:00 Richard Weinberger <[email protected]>:

Quoted 16 lines

Am 28.02.2015 um 00:48 schrieb Daeseok Youn:

Quoted 9 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

As this patch is untested and the issue is theoretical I'm nervous. But the final decision is up to ocfs2 maintainers.

Thanks,
//richard

#3 lore raw

Hi,

How is this patch going on?

please check for me.

Thanks.
Daeseok Youn.

2015-03-03 10:38 GMT+09:00 DaeSeok Youn <[email protected]>:

Quoted 26 lines

Hi, all

please, review this patch.

thanks.

regards,
Daeseok Youn.

2015-03-02 18:04 GMT+09:00 Richard Weinberger <[email protected]>:

Quoted 16 lines

Am 28.02.2015 um 00:48 schrieb Daeseok Youn:

Quoted 9 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

As this patch is untested and the issue is theoretical I'm nervous. But the final decision is up to ocfs2 maintainers.

Thanks,
//richard

#4 lore raw
In reply to Daeseok Youn

Hi,

This patch have been pending for 2 weeks.
Do I need to check other things?

please, check for me.

Thanks!

regards,
Daeseok Youn

2015-02-28 8:48 GMT+09:00 Daeseok Youn <[email protected]>:

Quoted 61 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
fs/ocfs2/dir.c +6 -7
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
        struct ocfs2_dir_entry *de, *de1;
        struct ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
        struct super_block *sb = dir->i_sb;
-       int retval, status;
+       int retval;
        unsigned int size = sb->s_blocksize;
        struct buffer_head *insert_bh = lookup->dl_leaf_bh;
        char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
                        }
 
                        if (insert_bh == parent_fe_bh)
-                               status = ocfs2_journal_access_di(handle,
+                               retval = ocfs2_journal_access_di(handle,
                                                                 INODE_CACHE(dir),
                                                                 insert_bh,
                                                                 OCFS2_JOURNAL_ACCESS_WRITE);
                        else {
-                               status = ocfs2_journal_access_db(handle,
+                               retval = ocfs2_journal_access_db(handle,
                                                                 INODE_CACHE(dir),
                                                                 insert_bh,
                                              OCFS2_JOURNAL_ACCESS_WRITE);
 
                                if (ocfs2_dir_indexed(dir)) {
-                                       status = ocfs2_dx_dir_insert(dir,
+                                       retval = ocfs2_dx_dir_insert(dir,
                                                                handle,
                                                                lookup);
-                                       if (status) {
-                                               mlog_errno(status);
+                                       if (retval) {
+                                               mlog_errno(retval);
                                                goto bail;
                                        }
                                }
--

1.7.1

#5 lore raw

Am 13.03.2015 um 04:59 schrieb DaeSeok Youn:

Quoted 6 lines

Hi,

This patch have been pending for 2 weeks.
Do I need to check other things?

please, check for me.

I cannot speak for OCFS2 folks. But maybe you can give them a better feeling if you proof that your patch is tested.

Thanks,
//richard

Quoted 67 lines

Thanks!

regards,
Daeseok Youn

2015-02-28 8:48 GMT+09:00 Daeseok Youn <[email protected]>:

Quoted 61 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
fs/ocfs2/dir.c +6 -7
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
        struct ocfs2_dir_entry *de, *de1;
        struct ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
        struct super_block *sb = dir->i_sb;
-       int retval, status;
+       int retval;
        unsigned int size = sb->s_blocksize;
        struct buffer_head *insert_bh = lookup->dl_leaf_bh;
        char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
                        }
 
                        if (insert_bh == parent_fe_bh)
-                               status = ocfs2_journal_access_di(handle,
+                               retval = ocfs2_journal_access_di(handle,
                                                                 INODE_CACHE(dir),
                                                                 insert_bh,
                                                                 OCFS2_JOURNAL_ACCESS_WRITE);
                        else {
-                               status = ocfs2_journal_access_db(handle,
+                               retval = ocfs2_journal_access_db(handle,
                                                                 INODE_CACHE(dir),
                                                                 insert_bh,
                                              OCFS2_JOURNAL_ACCESS_WRITE);
 
                                if (ocfs2_dir_indexed(dir)) {
-                                       status = ocfs2_dx_dir_insert(dir,
+                                       retval = ocfs2_dx_dir_insert(dir,
                                                                handle,
                                                                lookup);
-                                       if (status) {
-                                               mlog_errno(status);
+                                       if (retval) {
+                                               mlog_errno(retval);
                                                goto bail;
                                        }
                                }
--

1.7.1

#6 lore raw

PATCH v2 1/4 ocfs2: use retval instead of status for checking error

Hi, Richard.

Thanks for reply.
Ok. I will keep trying to set the environment for verifying my paches. But, before I said, it takes some time... Setting this filesystem in my machine with qemu is too hard for me. :(

Regards,
Daeseok Youn.

2015. 3. 13. ?? 5:16? "Richard Weinberger" <[email protected]>?? ??:

Quoted 38 lines

Am 13.03.2015 um 04:59 schrieb DaeSeok Youn:

Quoted 6 lines

Hi,

This patch have been pending for 2 weeks.
Do I need to check other things?

please, check for me.

I cannot speak for OCFS2 folks. But maybe you can give them a better feeling if you proof that your patch is tested.

Thanks,
//richard

Quoted 23 lines

Thanks!

regards,
Daeseok Youn

2015-02-28 8:48 GMT+09:00 Daeseok Youn <[email protected]>:

Quoted 17 lines

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
Quoted 7 lines
Quoted 7 lines

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1618

Quoted 7 lines
fs/ocfs2/dir.c +0 -0
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
        struct ocfs2_dir_entry *de, *de1;
        struct ocfs2_dinode *di = (struct ocfs2_dinode

*)parent_fe_bh->b_data;

Quoted 11 lines
Quoted 11 lines

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1698

Quoted 11 lines
fs/ocfs2/dir.c +1 -2
        struct super_block *sb = dir->i_sb;
-       int retval, status;
+       int retval;
        unsigned int size = sb->s_blocksize;
        struct buffer_head *insert_bh = lookup->dl_leaf_bh;
        char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
                        }
 
                        if (insert_bh == parent_fe_bh)
-                               status =

ocfs2_journal_access_di(handle,

Quoted 1 line
Quoted 1 line

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1698

Quoted 1 line
fs/ocfs2/dir.c +1 -0
+                               retval =

ocfs2_journal_access_di(handle,

Quoted 1 line
Quoted 1 line
Quoted 1 line

INODE_CACHE(dir),

Quoted 1 line
Quoted 1 line
Quoted 1 line

insert_bh,

Quoted 1 line
Quoted 1 line
Quoted 1 line

OCFS2_JOURNAL_ACCESS_WRITE);

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
                        else {

- status =

ocfs2_journal_access_db(handle,

Quoted 1 line
Quoted 1 line
Quoted 1 line

+ retval =

ocfs2_journal_access_db(handle,

Quoted 1 line
Quoted 1 line
Quoted 1 line

INODE_CACHE(dir),

Quoted 1 line
Quoted 1 line
Quoted 1 line

insert_bh,

Quoted 1 line
Quoted 1 line
Quoted 1 line

OCFS2_JOURNAL_ACCESS_WRITE);

Quoted 3 lines
Quoted 3 lines
Quoted 3 lines
                                if (ocfs2_dir_indexed(dir)) {

- status =

ocfs2_dx_dir_insert(dir,

Quoted 1 line
Quoted 1 line
Quoted 1 line

+ retval =

ocfs2_dx_dir_insert(dir,

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

handle,

lookup);

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
fs/ocfs2/dir.c +2 -3
-                                       if (status) {
-                                               mlog_errno(status);
+                                       if (retval) {
+                                               mlog_errno(retval);
                                                goto bail;
                                        }
                                }
--

1.7.1

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://oss.oracle.com/pipermail/ocfs2-devel/attachments/20150313/7c2e5ba8/attachment.html

Reviewed-by on v2 1/4
#7 lore raw
In reply to Daeseok Youn

Looks good to me.

Quoted 11 lines · Daeseok Youn

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

Quoted 48 lines

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
fs/ocfs2/dir.c +6 -6
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
 	struct ocfs2_dir_entry *de, *de1;
 	struct ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
 	struct super_block *sb = dir->i_sb;
-	int retval, status;
+	int retval;
 	unsigned int size = sb->s_blocksize;
 	struct buffer_head *insert_bh = lookup->dl_leaf_bh;
 	char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
 			}
 
 			if (insert_bh == parent_fe_bh)
-				status = ocfs2_journal_access_di(handle,
+				retval = ocfs2_journal_access_di(handle,
 								 INODE_CACHE(dir),
 								 insert_bh,
 								 OCFS2_JOURNAL_ACCESS_WRITE);
 			else {
-				status = ocfs2_journal_access_db(handle,
+				retval = ocfs2_journal_access_db(handle,
 								 INODE_CACHE(dir),
 								 insert_bh,
 					      OCFS2_JOURNAL_ACCESS_WRITE);
 
 				if (ocfs2_dir_indexed(dir)) {
-					status = ocfs2_dx_dir_insert(dir,
+					retval = ocfs2_dx_dir_insert(dir,
 								handle,
 								lookup);
-					if (status) {
-						mlog_errno(status);
+					if (retval) {
+						mlog_errno(retval);
 						goto bail;
 					}
 				}
#8 lore raw

2015-03-19 15:24 GMT+09:00 Joseph Qi <[email protected]>:

Quoted 11 lines

Looks good to me.

Quoted 8 lines · Daeseok Youn

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

Quoted 5 lines
Quoted 3 lines

So replace 'status' with 'retval'.

Hi, Joseph.

Thanks for review this patch. I have actually tried to set this file system on my machine with qemu but I didn't success, yet.

Regards,
Daeseok Youn.

Quoted 50 lines
Quoted 48 lines

---
V2 : update changelog

 fs/ocfs2/dir.c |   12 ++++++------
 1 files changed, 6 insertions(+), 6 deletions(-)
fs/ocfs2/dir.c +6 -6
index b08050b..1478a50 100644
@@ -1617,7 +1617,7 @@ int __ocfs2_add_entry(handle_t *handle,
      struct ocfs2_dir_entry *de, *de1;
      struct ocfs2_dinode *di = (struct ocfs2_dinode *)parent_fe_bh->b_data;
      struct super_block *sb = dir->i_sb;
-     int retval, status;
+     int retval;
      unsigned int size = sb->s_blocksize;
      struct buffer_head *insert_bh = lookup->dl_leaf_bh;
      char *data_start = insert_bh->b_data;
@@ -1695,22 +1695,22 @@ int __ocfs2_add_entry(handle_t *handle,
                      }
 
                      if (insert_bh == parent_fe_bh)
-                             status = ocfs2_journal_access_di(handle,
+                             retval = ocfs2_journal_access_di(handle,
                                                               INODE_CACHE(dir),
                                                               insert_bh,
                                                               OCFS2_JOURNAL_ACCESS_WRITE);
                      else {
-                             status = ocfs2_journal_access_db(handle,
+                             retval = ocfs2_journal_access_db(handle,
                                                               INODE_CACHE(dir),
                                                               insert_bh,
                                            OCFS2_JOURNAL_ACCESS_WRITE);
 
                              if (ocfs2_dir_indexed(dir)) {
-                                     status = ocfs2_dx_dir_insert(dir,
+                                     retval = ocfs2_dx_dir_insert(dir,
                                                              handle,
                                                              lookup);
-                                     if (status) {
-                                             mlog_errno(status);
+                                     if (retval) {
+                                             mlog_errno(retval);
                                              goto bail;
                                      }
                              }
#9 lore raw
In reply to Daeseok Youn
Quoted 10 lines · On Sat, 28 Feb 2015 08:48:40 +0900 Daeseok Youn <[email protected]> wrote:

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1714

Quoted 4 lines
fs/ocfs2/dir.c +2 -1
-						mlog_errno(status);
+					if (retval) {
+						mlog_errno(retval);
 						goto bail;

and

bail:

	if (retval)
		mlog_errno(retval);

	return retval;
}

so we'll clearly log the same error twice.

#10 lore raw

Hi Andrew,

Quoted 19 lines · Andrew Morton
Quoted 10 lines · On Sat, 28 Feb 2015 08:48:40 +0900 Daeseok Youn <[email protected]> wrote:

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1714

Quoted 4 lines
fs/ocfs2/dir.c +2 -1
-						mlog_errno(status);
+					if (retval) {
+						mlog_errno(retval);
 						goto bail;

and

Quoted 9 lines

bail:

	if (retval)
		mlog_errno(retval);

	return retval;
}

so we'll clearly log the same error twice.

IMO, if we only depends on the bail error log, we still don't know where the error occurs.
So if we want to do the cleanup, the bail error log should be cleaned.

Quoted 6 lines

_______________________________________________
Ocfs2-devel mailing list
Ocfs2-devel at oss.oracle.com
https://oss.oracle.com/mailman/listinfo/ocfs2-devel

#11 lore raw

Hi, Andrew.

2015-03-20 10:17 GMT+09:00 Joseph Qi <[email protected]>:

Quoted 30 lines

Hi Andrew,

Quoted 27 lines · Andrew Morton
Quoted 10 lines · On Sat, 28 Feb 2015 08:48:40 +0900 Daeseok Youn <[email protected]> wrote:

The use of 'status' in __ocfs2_add_entry() can return wrong status when some functions are failed.

If ocfs2_journal_access_db() in __ocfs2_add_entry() is failed, that status is saved to 'status' but return variable is 'retval' which is saved 'success' status. In case of this, __ocfs2_add_entry() is failed but can be returned as 'success'.

So replace 'status' with 'retval'.

fs/ocfs2/dir.c int __ocfs2_add_entry(handle_t *handle, · line 1714

Quoted 4 lines
fs/ocfs2/dir.c +2 -1
-                                            mlog_errno(status);
+                                    if (retval) {
+                                            mlog_errno(retval);
                                             goto bail;

and

bail:

      if (retval)
              mlog_errno(retval);

      return retval;
}

so we'll clearly log the same error twice.

As joseph comment, it doesn't need to cleanup mlog_errno() call in this patch. And he reviewed this patch.

Check, please.

Thanks.
Daeseok Youn

Quoted 13 lines
Quoted 1 line

IMO, if we only depends on the bail error log, we still don't know where the error occurs.
So if we want to do the cleanup, the bail error log should be cleaned.

Quoted 6 lines

_______________________________________________
Ocfs2-devel mailing list
Ocfs2-devel at oss.oracle.com
https://oss.oracle.com/mailman/listinfo/ocfs2-devel