Re: Quota on tmpfs

[Francois Tigeot <ftigeot%wolfpond.org@localhost>]

On Thu, Jul 12, 2012 at 07:20:37PM +0000, David Holland wrote:
> On Thu, Jul 12, 2012 at 07:14:07PM +0000, David Holland wrote:
> 
> ...and to actually answer the question, the fs-independent quota
> system is a data model, a client interface, and transport/request
> routing code under the covers. Plus the userlevel tools, which use the
> client interface.
> 
> The fs-dependent quota system takes care of storing the quota
> information, storing/maintaining the usage information, enforcement
> based on the usage information, and also recovering/restoring
> consistency of the usage information after crashing.

OK, more like a framework for managing filesystem-specific quotas then.
Thanks for the clarification.

> It is possible that some of the enforcement logic could be made
> fs-independent, but it's probably difficult without knowing how the
> filesystem allocates resources. E.g. if you write() to fill in holes
> in a file, outside the filesystem it's very difficult to know that
> you're increasing your block usage, or for that matter by how much
> when indirect blocks are (might be) involved.

This is not really an issue for me; I don't care about low-level storage
and only manage visible file sizes.

-- 
Francois Tigeot