On Wed, May 20, 2026 at 11:07:50AM +1000, David Gibson wrote:
> On Tue, May 19, 2026 at 10:03:39AM +0200, Laurent Vivier wrote:
> > On 5/19/26 07:08, David Gibson wrote:
> > > On Mon, May 18, 2026 at 02:26:25PM +0200, Laurent Vivier wrote:
> > > > Change udp_update_hdr4() and udp_update_hdr6() to take an iov_tail
> > > > pointing at the UDP frame instead of a contiguous udp_payload_t buffer
> > > > and explicit data length.  This lets vhost-user pass scatter-gather
> > > > virtqueue buffers directly without an intermediate copy.
> > > > 
> > > > The UDP header is built into a local struct udphdr and written back with
> > > > IOV_PUSH_HEADER().  On the tap side, udp_tap_prepare() wraps the
> > > > existing udp_payload_t in a two-element iov to match the new interface.
> > > > 
> > > > Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> > > 
> > > Alas, this still has a potentially aliased memcpy(), see below.
> > 
> > Did you see I updated patch 2 to avoid the memcpy() if the header is
> > already in place?
> > (I think I should have removed your R-b...)
> 
> I saw the note in the changelog, but I didn't spot the place in the
> code specifically - I think I skipped patch 2 because of the R-b and
> assumed the change was in patch 3.  Unfortunately the one in this
> patch is still there.

Duh, sorry, I misunderstood how you were fixing the problem.  I see
you changed PUSH_HEADER to be safe in this circumstance, which does
also fix the case in patch 3.  Will have that in mind when I review
v10.

-- 
David Gibson (he or they)	| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you, not the other way
				| around.
http://www.ozlabs.org/~dgibson