ReiserFS patch for updating ctimes of renamed files

Hans Reiser:

Many years ago we had the same problem with ext2: it did not change ctime on rename, so GNU tar did not pick up the renamed files on incremental backup. Fortunately a few kernel versions later ext2 changed to the current behaviour (unfortunately I don't remember the kernel version).

IIRC our sysadmin submitted a bug report for GNU tar at the time, and got the answer that GNU tar would not change.

- anton

M. Anton Ertl                    Some things have to be seen to be believed
[email protected] Most things have to be believed to be seen
http://www.complang.tuwien.ac.at/anton/home.html

I looked again at the definition of the difference between ctime and mtime on the stat man page, and I think that updating ctime in response to rename is as reasonable as updating it in response to changing the number of links.

Ok, we will conform, and I will accept the kindly donated patch, along with Andrew's optimization of our evaluation of CURRENT_TIME. vs, please add Andrew's suggested optimization and sent the result through QA. Thanks to all for your good advice.

Hans

Great. Thanks a lot!

Where would I be able to download the patch that you guys will be producing once it's through QA? At ftp://ftp.namesys.com/pub/reiserfs-for-2.4/ or is there some other place I can get it (e.g. BK/CVS/HTTP/FTP) before it's posted at that link?

Thanks,

Alex

-----Original Message-----
From: [email protected] [mailto:[email protected]]On Behalf Of Hans Reiser
Sent: Tuesday, October 14, 2003 3:41 AM
To: Anton Ertl
Cc: [email protected]; Andrew Morton; [email protected]; jw schultz; Alex Adriaanse
Subject: Re: ReiserFS patch for updating ctimes of renamed files

I looked again at the definition of the difference between ctime and mtime on the stat man page, and I think that updating ctime in response to rename is as reasonable as updating it in response to changing the number of links.

Ok, we will conform, and I will accept the kindly donated patch, along with Andrew's optimization of our evaluation of CURRENT_TIME. vs, please add Andrew's suggested optimization and sent the result through QA. Thanks to all for your good advice.

--
Hans

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

Hi Hans,

I updated my patch to include Andrew's suggestion of eliminating extra calls to CURRENT_TIME. I also finally got a chance to test it out, and it seems to work. After applying this patch, ctime gets updated after a rename, and GNU tar now backs things up properly. I also could not detect any filesystem corruption after doing some renames.

Alex

@@ -1205,8 +1205,11 @@
 
     mark_de_hidden (old_de.de_deh + old_de.de_entry_num);
     journal_mark_dirty (&th, old_dir->i_sb, old_de.de_bh);
-    old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME;
-    new_dir->i_ctime = new_dir->i_mtime = CURRENT_TIME;
+    ctime = CURRENT_TIME;
+    old_dir->i_ctime = old_dir->i_mtime = ctime;
+    new_dir->i_ctime = new_dir->i_mtime = ctime;
+    old_inode->i_ctime = ctime;
+    reiserfs_update_sd (&th, old_inode);
 
     if (new_dentry_inode) {
        // adjust link number of the victim

@@ -1215,7 +1218,6 @@
        } else {
            new_dentry_inode->i_nlink--;
        }
-       ctime = CURRENT_TIME;
        new_dentry_inode->i_ctime = ctime;
        savelink = new_dentry_inode->i_nlink;
     }