Re: [PATCH v6 net 1/8] xsk: tighten UMEM headroom validation to account for tailroom and min frame

[Maciej Fijalkowski <maciej.fijalkowski@intel.com>]

On Fri, Apr 03, 2026 at 05:01:26PM -0700, Jakub Kicinski wrote:
> On Thu,  2 Apr 2026 17:49:51 +0200 Maciej Fijalkowski wrote:
> > The current headroom validation in xdp_umem_reg() could leave us with
> > insufficient space dedicated to even receive minimum-sized ethernet
> > frame. Furthermore if multi-buffer would come to play then
> > skb_shared_info stored at the end of XSK frame would be corrupted.
> > 
> > HW typically works with 128-aligned sizes so let us provide this value
> > as bare minimum.
> > 
> > Multi-buffer setting is known later in the configuration process so
> > besides accounting for 128 bytes, let us also take care of tailroom space
> > upfront.
> 
> I guess it's a bit late to ask this, but am I the only one that has
> weird feelings about shinfo living in user-mapped memory?

This was the approach was picked to be consistent with XDP side and allow
to use mbuf related helpers. I can imagine you would be suggesting to pick
up pages from system page pool but that would be a major refactor (however
I have not put much not thought into it - i'm here without a coffee and in
easter mode).

Outside of this rewrite this set fixes existing memory corruption, so I
feel it is needed and we can think how should we address the concern you
are bringing up later?