On 04/08/15 at 04:28pm, Baoquan He wrote:
On 04/08/15 at 04:23pm, Minfei Huang wrote:
> On 04/08/15 at 04:03pm, Baoquan He wrote:
> > > We can find that /mnt/bind-dir is a bind mounted directory which points
> > > to the /dev/vda[/var], meanwhile, the device /dev/vda is mounted on
> > > directory /mnt/ext4. So the bind mounted directory /mnt/bind-dir points
> > > to the /mnt/ext4/var.
> >
> > Then what's the mount point you get? You tell a series of analysis,
> > what's the result?
> >
> > Could you please assume a reviewer who didn't follow this issue want to
> > help review this patchset and get into it quickly? The reason for this
> > is we can recall these later by checking the commit log. I often forget
> > the details several months later.
> >
>
> Hi, bao.
>
> I just show the view what bind mounted is. Where the real location is, if
> the path is bind mounted.
OK. I think in patch log you should present:
1) what's wrong when bind mount in current code, you can give a example
command.
2)The reason is the mount point is not correct and that cause storing in
wrong place in 2nd kernel.
3) how do you fix it with code in this patch. Namely how to get the
mount point correctly.
With the help of command example it should be easy to explain
what/why/how. So if you just show the view what bind mounted is, where
is the real location is if bind mount, I still need try it by myself to
get what's the correct mount point, but not get it directly from you
patch log.
Thanks
Baoquan
Thanks for your suggestion. Will modify the commit log as what you said.
Thanks
Minfei