Re: [RFC PATCH] xfsrestore: fix rootdir due to xfsdump bulkstat misuse

[Gao Xiang <hsiangkao@redhat.com>]

Hi Eric,

On Fri, Nov 13, 2020 at 10:15:53PM +0800, Gao Xiang wrote:
> On Fri, Nov 13, 2020 at 08:10:39AM -0600, Eric Sandeen wrote:
> > On 11/13/20 6:51 AM, Gao Xiang wrote:
> 
> ...
> 
> > 
> > Thank you for looking into this - I think you now understand xfsdump &
> > xfsrestore better than anyone else on the planet.  ;)
> > 
> > One question - what happens if the wrong "root inode" is not a directory?
> > I think that it is possible from the old "get the first active inode" heuristic
> > to find any type of file and save it as the root inode.
> > 
> > I think that your approach still works in this case, but wanted to double check
> > and see what you think.
> 
> Yeah, good question. I also think it works too, but just in case let me
> do fault injection on a regular inode later (Donald's image is /var
> subdir...)
> 

Sorry for the previous wrong conclusion...

From the code itself, tree_begindir() only triggers for node_t == dir but all
dirents can be trigged by tree_addent(), so I update the patch and verified
with manual fault injection code as well...

RFC v2: https://lore.kernel.org/linux-xfs/20201116080723.1486270-1-hsiangkao@redhat.com/
fault injection: https://lore.kernel.org/linux-xfs/20201116081359.GA1486562@xiangao.remote.csb/

Thanks,
Gao Xiang

> Thanks,
> Gao Xiang
> 
> > 
> > Thanks,
> > -Eric
> >