netdev - Re: [PATCH] bpf: don't fail kmalloc while releasing raw

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

Message-ID: <20201118004242.rygrwivqcdgeowi7@hydra.tuxags.com>
Date:   Tue, 17 Nov 2020 16:42:44 -0800
From:   Matt Mullins <mmullins@...x.us>
To:     Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
Cc:     rostedt <rostedt@...dmis.org>, paulmck <paulmck@...nel.org>,
        Matt Mullins <mmullins@...x.us>,
        Ingo Molnar <mingo@...hat.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Dmitry Vyukov <dvyukov@...gle.com>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        Andrii Nakryiko <andriin@...com>,
        John Fastabend <john.fastabend@...il.com>,
        KP Singh <kpsingh@...omium.org>,
        linux-kernel <linux-kernel@...r.kernel.org>,
        netdev <netdev@...r.kernel.org>, bpf <bpf@...r.kernel.org>
Subject: Re: [PATCH] bpf: don't fail kmalloc while releasing raw_tp

On Tue, Nov 17, 2020 at 06:05:51PM -0500, Mathieu Desnoyers wrote:
> ----- On Nov 16, 2020, at 5:10 PM, rostedt rostedt@...dmis.org wrote:
> 
> > On Mon, 16 Nov 2020 16:34:41 -0500 (EST)
> > Mathieu Desnoyers <mathieu.desnoyers@...icios.com> wrote:
> 
> [...]
> 
> >> I think you'll want a WRITE_ONCE(old[i].func, tp_stub_func) here, matched
> >> with a READ_ONCE() in __DO_TRACE. This introduces a new situation where the
> >> func pointer can be updated and loaded concurrently.
> > 
> > I thought about this a little, and then only thing we really should worry
> > about is synchronizing with those that unregister. Because when we make
> > this update, there are now two states. the __DO_TRACE either reads the
> > original func or the stub. And either should be OK to call.
> > 
> > Only the func gets updated and not the data. So what exactly are we worried
> > about here?
> 
> Indeed with a stub function, I don't see any need for READ_ONCE/WRITE_ONCE.

I'm not sure if this is a practical issue, but without WRITE_ONCE, can't
the write be torn?  A racing __traceiter_ could potentially see a
half-modified function pointer, which wouldn't work out too well.

This was actually my gut instinct before I wrote the __GFP_NOFAIL
instead -- currently that whole array's memory ordering is provided by
RCU and I didn't dive deep enough to evaluate getting too clever with
atomic modifications to it.

> 
> However, if we want to compare the function pointer to some other value and
> conditionally do (or skip) the call, I think you'll need the READ_ONCE/WRITE_ONCE
> to make sure the pointer is not re-fetched between comparison and call.
> 
> Thanks,
> 
> Mathieu
> 
> -- 
> Mathieu Desnoyers
> EfficiOS Inc.
> http://www.efficios.com