[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CANP3RGcb9cpOgSOFaX2=eKoPicawsWtdEQU4EeCt3+g0V8Pw2A@mail.gmail.com>
Date: Thu, 17 Oct 2024 15:21:22 +0200
From: Maciej Żenczykowski <maze@...gle.com>
To: Florian Westphal <fw@...len.de>
Cc: netdev@...r.kernel.org, Nathan Harold <nharold@...gle.com>,
Steffen Klassert <steffen.klassert@...unet.com>, Yan Yan <evitayan@...gle.com>
Subject: Re: [PATCH ipsec] xfrm: migrate: work around 0 if_id on migrate
On Thu, Oct 17, 2024 at 1:33 PM Florian Westphal <fw@...len.de> wrote:
>
> Maciej Żenczykowski <maze@...gle.com> wrote:
> > > If you think a Kconfig knob makes sense for Android sake I can respin
> > > with such a knob, but I think I'd have to make it default-y.
> >
> > I'll trust your judgment. I thought we could maybe eventually
> > deprecate and delete this, but it sounds like that isn't going to be
> > the case...
>
> We could also say 'android fixed it in userspace' and not fix it
> in kernel for now until someone else complains.
The Android patch is here:
https://android-review.googlesource.com/c/platform/system/netd/+/3303667
But I'm not certain if this is enough (though I would assume so).
It's so much simpler and more correct than the kernel fix...
> Or modify the pr_warn to also say something like
> "deprecated and scheduled to be removed in 2026"?
>
> I'm not thrilled with this patch :-)
Let's see what Yan Yan thinks is the right approach.
XFRM migration seems like one of those things with very very few
users, so it might be easy to just fix them.
Powered by blists - more mailing lists