Re: Max binds

[Herbert Poetzl]

On Tue, Jul 08, 2003 at 12:00:49PM +0300, Jarmo Järvenpää wrote:
> Ville Herva wrote:
> > 
> > See
> > >http://groups.google.com/groups?dq=&hl=en&lr=&ie=UTF-8&threadm=linux.kernel.3AE307AD.821AB47C%40linuxjedi.org&rn
> 
> >um=1&prev=/groups%3Fq%3Dg:thl1557665770d%26dq%3D%26hl%3Den%26lr%3D%26ie%3DUTF-8%26selm%3Dlinux.kernel.3AE307AD.8>21AB47C%2540linuxjedi.org
> 
> Hmm, I read the thread and it contained mostly stats on memory usage on
> different scenarios. Also, bit outdated discussion (kernel v 2.4.4). No
> direct answer though. 

I'm no expert on the mount --bind stuff, but as far
as I understand it, this is simple (I guess Al Viro 
would agree *g*) VFS magic, so what happens seems to
be the following:

 - a new vfsmount entry is created by either
   copying the old tree or cloning the original
   mount (if it is a mountpoint)
 - the required dentries are added to the dcache

so the number of --bind mounts from another mount
point will be roughly limited by the number of
vfsmount structures possible, and the space available
to cache/store dentries ...

I guess there will be some tradeoff when the number
of --bind mounts becomes very high, but I guess it
is neglectible for normal use ...

I might be totaly wrong, so do not rely on it 8-)

HTH,
Herbert

> Jarmo
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [e-mail]
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/