[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20240916101942.ZJP2h0NM@linutronix.de>
Date: Mon, 16 Sep 2024 12:19:42 +0200
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: Daniel Borkmann <daniel@...earbox.net>
Cc: Breno Leitao <leitao@...ian.org>, Jakub Kicinski <kuba@...nel.org>,
razor@...ckwall.org, andrii@...nel.org, ast@...nel.org,
syzbot <syzbot+08811615f0e17bc6708b@...kaller.appspotmail.com>,
bpf@...r.kernel.org, davem@...emloft.net, eddyz87@...il.com,
haoluo@...gle.com, hawk@...nel.org, john.fastabend@...il.com,
jolsa@...nel.org, kpsingh@...nel.org, linux-kernel@...r.kernel.org,
martin.lau@...ux.dev, netdev@...r.kernel.org, sdf@...ichev.me,
song@...nel.org, syzkaller-bugs@...glegroups.com,
yonghong.song@...ux.dev
Subject: Re: [PATCH net-net] tun: Assign missing bpf_net_context.
On 2024-09-12 17:03:15 [+0200], Daniel Borkmann wrote:
>
> Oh well, quite annoying that we need this context now everywhere also outside of XDP :(
> Sebastian, do you see any way where this could be noop for !PREEMPT_RT?
This isn't related to XDP but to the redirect part of BPF which is (or
was) using per-CPU variables.
I don't know how much pain it causes here for you and how much of this
is actually helping and not making anything worse:
- If netkit::active is likely to be NULL you could limit assigning the
context only if it != NULL
- If you can ensure (via verifier) that netkit_run() won't access the
redirect helper (such as bpf_redirect()) and won't return
NETKIT_REDIRECT (as a consequence) then the assignment could be
avoided in this case.
Sebastian
Powered by blists - more mailing lists