diff options
author | David Gibson <david@gibson.dropbear.id.au> | 2024-07-05 20:44:03 +1000 |
---|---|---|
committer | Stefano Brivio <sbrivio@redhat.com> | 2024-07-05 15:26:33 +0200 |
commit | 9f9b15f949e40cc9bacd9c034637f3c2763ac1bb (patch) | |
tree | 3a7ab24fd7b950129683b7b7facb7d28b1b6d6e7 /doc/platform-requirements | |
parent | fbd78b6f3e95b1e6b712d8ad6283b588e8cfe9c4 (diff) | |
download | passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar.gz passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar.bz2 passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar.lz passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar.xz passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.tar.zst passt-9f9b15f949e40cc9bacd9c034637f3c2763ac1bb.zip |
udp: Unify udp[46]_mh_splice
We have separate mmsghdr arrays for splicing IPv4 and IPv6 packets, where
the only difference is that they point to different sockaddr buffers for
the destination address.
Unify these by having the common array point at a sockaddr_inany as the
address. This does mean slightly more work when we're about to splice,
because we need to write the whole socket address, rather than just the
port. However it removes 32 mmsghdr structures and we're going to need
more flexibility constructing that target address for the flow table.
Because future changes might mean that the address isn't always loopback,
change the name of the common address from *_localname to udp_splicename.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Stefano Brivio <sbrivio@redhat.com>
Diffstat (limited to 'doc/platform-requirements')
0 files changed, 0 insertions, 0 deletions