Re: 3.9-rc6 ext4: free_rb_tree_fname oops

[Zheng Liu <gnehzuil.liu@gmail.com>]

Hi Daniel,

On Mon, Jun 24, 2013 at 02:34:00PM +0800, Daniel J Blueman wrote:
> On 16 April 2013 15:37, Daniel J Blueman <daniel@quora.org> wrote:
> > When using e4defrag on a ext4 filesystem created a month ago, I ran
> > into this fatal page fault [1]
> >  while running e4defrag on 3.9-rc6 (Ubuntu mainline).
> >
> > e2fsdump output is at http://quora.org/2012/e2fsdump.txt ; let me know
> > if you need any more info.
> 
> With 3.9.6 mainline, I got the exact same protection fault at
> free_rb_tree_fname() from ext4_htree_free_dir_info() [1]. This
> suggests use-after-free, as there's no pagetable mapping.
> 
> There is nothing special with my setups, so there is fair chance it's
> reproducible there with e4defrag on a few month old filesystem and
> recent kernels.

These days I try to reproduce this bug, but unfortunately I couldn't
hit it.  I create/read/write/delete some files in a SSD disk to simulate
a file system that has been used for a while.  Then I use e4defrag to
defrag this file system.  But I couldn't trigger the bug.  The kernel
version is the latest ext4/dev branch, and the e2fsprgs version is the
1.42.7.  Do you have a method to easily reproduce this bug?

Thanks,
                                                - Zheng