[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220916143740.831881-1-Jason@zx2c4.com>
Date: Fri, 16 Sep 2022 15:37:37 +0100
From: "Jason A. Donenfeld" <Jason@...c4.com>
To: kuba@...nel.org, pablo@...filter.org, davem@...emloft.net,
netdev@...r.kernel.org
Cc: "Jason A. Donenfeld" <Jason@...c4.com>
Subject: [PATCH net 0/3] wireguard patches for 6.0-rc6
Hey guys,
Sorry we didn't get a chance to talk at Plumbers. I saw some of you very
briefly in passing but never had the opportunity to chat. Next time, I
hope.
Please pull the following fixes:
1) The ratelimiter timing test doesn't help outside of development, yet
it is currently preventing the module from being inserted on some
kernels when it flakes at insertion time. So we disable it.
2) A fix for a build error on UML, caused by a recent change in a
different tree.
3) A WARN_ON() is triggered by Kees' new fortified memcpy() patch, due
to memcpy()ing over a sockaddr pointer with the size of a
sockaddr_in[6]. The type safe fix is pretty simple. Given how classic
of a thing sockaddr punning is, I suspect this may be the first in a
few patches like this throughout the net tree, once Kees' fortify
series is more widely deployed (current it's just in next).
Thanks,
Jason
Jason A. Donenfeld (3):
wireguard: ratelimiter: disable timings test by default
wireguard: selftests: do not install headers on UML
wireguard: netlink: avoid variable-sized memcpy on sockaddr
drivers/net/wireguard/netlink.c | 13 +++++-----
drivers/net/wireguard/selftest/ratelimiter.c | 25 ++++++++-----------
.../testing/selftests/wireguard/qemu/Makefile | 2 ++
3 files changed, 18 insertions(+), 22 deletions(-)
--
2.37.3
Powered by blists - more mailing lists