linux-kernel - Re: [PATCH v2 04/10] ftrace: Store direct called addresses in their ops

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 for Android: free password hash cracker in your pocket

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

Message-ID: <20230316114544.5db09039@gandalf.local.home>
Date:   Thu, 16 Mar 2023 11:45:44 -0400
From:   Steven Rostedt <rostedt@...dmis.org>
To:     Florent Revest <revest@...omium.org>
Cc:     linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-trace-kernel@...r.kernel.org, bpf@...r.kernel.org,
        catalin.marinas@....com, will@...nel.org, mhiramat@...nel.org,
        mark.rutland@....com, ast@...nel.org, daniel@...earbox.net,
        andrii@...nel.org, kpsingh@...nel.org, jolsa@...nel.org,
        xukuohai@...weicloud.com, lihuafei1@...wei.com
Subject: Re: [PATCH v2 04/10] ftrace: Store direct called addresses in their
 ops

On Thu, 16 Mar 2023 16:40:48 +0100
Florent Revest <revest@...omium.org> wrote:

> > > @@ -5466,6 +5467,7 @@ __modify_ftrace_direct(struct ftrace_ops *ops, unsigned long addr)
> > >                       entry->direct = addr;
> > >               }
> > >       }
> > > +     WRITE_ONCE(ops->direct_call, addr);  
> >
> > I'm curious about the use of WRITE_ONCE(). It should not go outside the
> > mutex barrier.  
> 
> This WRITE_ONCE was originally suggested by Mark here:
> https://lore.kernel.org/all/Y9vW99htjOphDXqY@FVFF77S0Q05N.cambridge.arm.com/#t
> 
> My understanding is that it's not so much about avoiding re-ordering
> but rather about avoiding store tearing since a ftrace_caller
> trampoline could concurrently read ops->direct_call. Does that make
> sense ?

Yes, but a comment needs to be added:

     /* Prevent store tearing on some archs */
     WRITE_ONCE(ops->direct_call, addr);  

Or something to that affect. Otherwise I can see it confusing others in the
future. And probably me too, as I'll forget why it was a WRITE_ONCE() by
next month. ;-)

-- Steve