public/io/netif.h: document the reality of netif_rx_request/reponse
authorPaul Durrant <paul.durrant@citrix.com>
Wed, 25 Nov 2015 10:12:26 +0000 (11:12 +0100)
committerJan Beulich <jbeulich@suse.com>
Wed, 25 Nov 2015 10:12:26 +0000 (11:12 +0100)
commit19167b11606df3c9598566a870ccd312b10d6d4b
tree30cea8e6f12c45c839cc780e45944d3fb25aea63
parent0c3f24645b07b875bc1294fb4627f01e030690fe
public/io/netif.h: document the reality of netif_rx_request/reponse

Because GSO metadata is passed from backend to frontend using
netif_extra_info segments, which do not carry information stating which
netif_rx_request_t was consumed to free up their slot, frontends must
assume some form of identity relation between ring slot and request.
Hence, so that it is able to use GSO metadata, Linux netfront simply
assumes rx responses appear in the same ring slot as their corresponding
request.

This patch documents the assumption made by Linux netfront and the
necessity of the assumption (to support GSO) so that backends are coded
to be compatible.

Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
xen/include/public/io/netif.h