linux-kernel - Re: Possible bug in ftrace_profile_enable

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: <c62985530910020513t45c32b00w30ac6204164e27ef@mail.gmail.com>
Date:	Fri, 2 Oct 2009 14:13:57 +0200
From:	Frédéric Weisbecker <fweisbec@...il.com>
To:	Paul Mackerras <paulus@...ba.org>
Cc:	Steven Rostedt <rostedt@...dmis.org>, Ingo Molnar <mingo@...e.hu>,
	Peter Zijlstra <peterz@...radead.org>,
	linux-kernel@...r.kernel.org
Subject: Re: Possible bug in ftrace_profile_enable_event

2009/10/1 Paul Mackerras <paulus@...ba.org>:
> I was looking through kernel/trace/trace_event_profile.c and I saw
> this code:
>
> static int ftrace_profile_enable_event(struct ftrace_event_call *event)
> {
>        char *buf;
>        int ret = -ENOMEM;
>
>        if (atomic_inc_return(&event->profile_count))
>                return 0;
>
>        if (!total_profile_count++) {
>                buf = (char *)alloc_percpu(profile_buf_t);
>                if (!buf)
>                        goto fail_buf;
>
>                rcu_assign_pointer(trace_profile_buf, buf);
>
>                buf = (char *)alloc_percpu(profile_buf_t);
>                if (!buf)
>                        goto fail_buf_nmi;
>
>                rcu_assign_pointer(trace_profile_buf_nmi, buf);
>        }
>
>        ret = event->profile_enable();
>        if (!ret)
>                return 0;
>
>        kfree(trace_profile_buf_nmi);
> fail_buf_nmi:
>        kfree(trace_profile_buf);
> fail_buf:
>        total_profile_count--;
>
> ...
>
> So we only allocate trace_profile_buf and trace_profile_buf_nmi if
> total_profile_count was zero on entry, but if we get an error returned
> from event->profile_enable(), we free them both unconditionally,
> regardless of the value of total_profile_count.  That seems wrong.  Is
> there a subtle reason why that is the right thing to do?


Oh right. Good catch.
I'll fix that soon.


>
> (Also, is kfree the appropriate counterpart to alloc_percpu?)

No indeed.
That said it looks harmless for now because percpu_free seem to just
roughly wrap kfree. But
its implementation may change later, so I'll fix that too.

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/