On Fri, 2 Jun 2000, Andries Brouwer wrote:
> # echo "fake::0:0:fake:/:" >> /tmp/pw
> # ls -l /tmp/pw
> -rw-r--r-- 1 root root 2751 Apr 29 21:45 /tmp/pw
> Funny - the file is unchanged. Where did the echoed stuff go?
> Answer: it went to the underlying /tmp/pw, the unmounted one.
What? Oh, fsck... Bug in open_namei()
> # rm /tmp/pw
... and in vfs_unlink() <checking> ... and vfs_rename_other().
Thanks for spotting. OK, I've fixed that in dcache-ac7-J (will be in usual
place when I'll run some tests on it) and I'm making the thing (bind, that
is) root-only for the time being. I still think that making it safe for
regular users is the right thing, but yes, scenario with rm -rf on /tmp is
pretty convincing. End of the story.
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to email@example.com
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:16 EST