netdev - Re: [PATCH net v2] net: clear the dst when changing skb protocol

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <CANP3RGc+BjY-7LRFZmeXLqXVAbf3aYce=_H-Ru7B2sw8O+mbGQ@mail.gmail.com>
Date: Tue, 10 Jun 2025 00:02:09 +0200
From: Maciej Żenczykowski <maze@...gle.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: davem@...emloft.net, netdev@...r.kernel.org, edumazet@...gle.com, 
	pabeni@...hat.com, andrew+netdev@...n.ch, horms@...nel.org, 
	Daniel Borkmann <daniel@...earbox.net>, martin.lau@...ux.dev, john.fastabend@...il.com, 
	eddyz87@...il.com, sdf@...ichev.me, haoluo@...gle.com, willemb@...gle.com, 
	william.xuanziyang@...wei.com, alan.maguire@...cle.com, bpf@...r.kernel.org, 
	shuah@...nel.org, linux-kselftest@...r.kernel.org, yonghong.song@...ux.dev
Subject: Re: [PATCH net v2] net: clear the dst when changing skb protocol

On Mon, Jun 9, 2025 at 11:59 PM Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Sat, 7 Jun 2025 23:33:39 +0200 Maciej Żenczykowski wrote:
> > 1 meta question: as this is a fix and will thus be backported into
> > 5.4+ LTS, should this be split into two patches? Either making the
> > test a follow up, or even going with only the crash fix in patch 1 and
> > putting the 4-in-4 and 6-in-6 behavioural change in patch 2?  We'd end
> > up in the same state at tip of tree... but it would affect the LTS
> > backports.  Honestly I'm not even sure what's best.
>
> :) Did we go from wondering if we can strip dst unconditionally to
> wondering if stripping it on encap/decap may introduce regressions?

Yeah, well I have utterly enough regression chasing in my day job.

Just spent two days chasing this fun one.

enum bpf_cmd {
        BPF_MAP_CREATE,
...
        BPF_PROG_DETACH,
        BPF_GET_COMM_HASH,    <--- added
        BPF_PROG_TEST_RUN,
...
        BPF_OBJ_GET_INFO_BY_FD,
};

> I suppose it may be useful to split, just to make it clear which
> portion of the change is the crash fix and which one is just because
> we think it's more consistent.

Your call.
> --
> pw-bot: cr

--
Maciej Żenczykowski, Kernel Networking Developer @ Google