Messages in this thread | | | Subject | Re: [PATCH RFC] tun: add ioctl to modify vnet header size | From | David Stevens <> | Date | Wed, 17 Mar 2010 15:02:44 -0700 |
| |
netdev-owner@vger.kernel.org wrote on 03/17/2010 02:35:04 PM:
> On Wed, Mar 17, 2010 at 02:10:11PM -0700, David Stevens wrote: > > Shouldn't we enforce a maximum too? Esp. if overflow/underflow > > will break any of the checks when it's used. > > > > +-DLS > > So the maximum is MAX_INT :) > I don't think it can break any checks that aren't > already broken - what do you have in mind?
I was thinking more like a page. At least, it'd be better to fail when trying to set it large than failing allocations later. As a header, it really ought to be small. But if it works, or fails gracefully, at 2^31-1 on 32-bit machines, negative values, etc, then it's ok. Just a suggestion.
+-DLS
| |