linux-kernel - Re: [PATCH v4a 00/38] timers: Use timer

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: <b52c207c-e7bf-af2e-61d9-116c9e76af86@inria.fr>
Date:   Sun, 6 Nov 2022 11:05:14 +0800 (+08)
From:   Julia Lawall <julia.lawall@...ia.fr>
To:     Steven Rostedt <rostedt@...dmis.org>
cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        linux-kernel@...r.kernel.org, Thomas Gleixner <tglx@...utronix.de>,
        Stephen Boyd <sboyd@...nel.org>,
        Guenter Roeck <linux@...ck-us.net>,
        Anna-Maria Gleixner <anna-maria@...utronix.de>,
        Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [PATCH v4a 00/38] timers: Use timer_shutdown*() before freeing
 timers



On Sat, 5 Nov 2022, Steven Rostedt wrote:

> On Sun, 6 Nov 2022 07:08:48 +0800 (+08)
> Julia Lawall <julia.lawall@...ia.fr> wrote:
>
> > Various suggestions:
> >
> > 1.  On your ... put when strict and then on a separate line put when !=
> > ptr.  The when strict will get rid of the goto problem (usually a
> > desirable feature, but not here) and the when != ptr will be sure that ptr
> > is not used before the free.
>
> You mean ptr->timer.function? because it's allowed to be touched. Only
> this case is weird (and I believe I covered it).
>
> Not sure what you mean by "put when strict" I added:
>
>  ... when strict
>
> Thinking that's what you meant (examples would be easier to understand,
> than descriptions). And it didn't cover the return case. Does it only
> cover gotos?
>
>   See drivers/net/wireless/intel/iwlwifi/mvm/sta.c for the false positive case:
>
> 	del_timer_sync(&baid_data->session_timer);
> [..]

But there is a kfree_rcu(baid_data, rcu_head); right below.  So it looks
fine?

julia

>         return 0;
>
> out_free:
>         kfree(baid_data);
>         return ret;
> }
>
> That "return 0" should make the match fail.
>
> >
> > 2.  If you want to handle the initialization of the function field, then
> > you can duplicate the rule and add the removal of that assignment in the
> > first one.  This only seems worth it if it is a very common case.
> > Otherwise, I would agree with Linus and just take care of it by hand
> > later.
>
> No, Linus wants the script to not touch the initialization case. That
> is, currently, the script does the conversion (which also initializes
> it), and the timer.function = NULL is just redundant.
>
> What Linus wanted, was my script to do nothing in this case. But I
> figured this part out.
>
> >
> > 3. Running the rule three times seems to me like a reasonable choice.  Or
> > you could duplicate the rule three times.  But that would be more script
> > to read through.  If this is not a common case, though, you could probably
> > also fix the one up later by hand.
>
> Yeah, that's fine.
>
> I'm just looking for how to avoid the goto / return case.
>
> -- Steve
>