Messages in this thread | | | From | kaz@earth ... | Subject | Re: [PATCH] robust ext2fs against failure | Date | Fri, 27 Aug 1999 16:20:59 +0900 (JST) |
| |
"Theodore Y. Ts'o" <tytso@mit.edu> writes:
tytso> The situation which I mean is that one new directory entry is tytso> allocated to point a newly-created memory inode, but corresponding tytso> disk inode still contains GARBAGE then: tytso> tytso> dir-entry inode tytso> tytso> fs/ntfs/ ---------- Makefile tytso> : tytso> ......: tytso> : tytso> /tmp/ ------+--> tmp990825 tytso> tytso> ---> memory inode tytso> .... garbage on the disk (of cource, overwritten in future...) tytso> tytso> If system crashes before the memory INODE is written back over the tytso> garbage, it causes doubly-linked file. tytso> tytso> OK, this only happens where "Makefile" is a newly created file. (That
Hmm...? I do not think so.
1. 'Makefile' stably exists already. 2. /tmp/tmp99825 is newly created. 2a. Memory inode is allocated for /tmp/tmp99825. 2b. But its correspoding data is garbage and points Makefile (Perhaps it is rare case..., I agree with you) 3. Crash here. 4. Valid memory inode is lost, garbage disk-inode is left pointing Makefile.
tytso> system crash. Since metadata is synced to the disk every five seconds, tytso> and in order for this to happen you have to get fairly unlucky such that tytso> the ext2 file allocation algorithm picks the just-deleted-inode as the tytso> one to be used for the new inode, this is **not** a common case.
Yes, it is not so common case, but isn't rare case worth consideration?
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |