linux-kernel - Re: [PATCH] tracing: Fix an off by one in _

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] [day] [month] [year] [list]

Date:   Tue, 27 Nov 2018 23:04:08 +0300
From:   Dan Carpenter <dan.carpenter@...cle.com>
To:     Steven Rostedt <rostedt@...dmis.org>
Cc:     Ingo Molnar <mingo@...hat.com>, linux-kernel@...r.kernel.org,
        kernel-janitors@...r.kernel.org
Subject: Re: [PATCH] tracing: Fix an off by one in __next()

On Tue, Nov 27, 2018 at 01:44:12PM -0500, Steven Rostedt wrote:
> 
> Doing the sweep of my INBOX, I came across this patch (it was sent
> while I was in the Alps :-)
> 
> 
> On Wed, 20 Jun 2018 14:08:00 +0300
> Dan Carpenter <dan.carpenter@...cle.com> wrote:
> 
> > The > should be >= to prevent an off by one bug.
> 
> Well, not really.
> 
> > 
> > >From reviewing the code, it seems possible for  
> > stack_trace_max.nr_entries to be set to .max_entries and in that case we
> > would be reading one element beyond the end of the stack_dump_trace[]
> > array.  If it's not set to .max_entries then the bug doesn't affect
> > runtime.
> > 
> > Signed-off-by: Dan Carpenter <dan.carpenter@...cle.com>
> > 
> > diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c
> > index 4237eba4ef20..6e3edd745c68 100644
> > --- a/kernel/trace/trace_stack.c
> > +++ b/kernel/trace/trace_stack.c
> > @@ -286,7 +286,7 @@ __next(struct seq_file *m, loff_t *pos)
> >  {
> >  	long n = *pos - 1;
> >  
> > -	if (n > stack_trace_max.nr_entries || stack_dump_trace[n] == ULONG_MAX)
> > +	if (n >= stack_trace_max.nr_entries || stack_dump_trace[n] == ULONG_MAX)
> 
> We have:
> 
> static unsigned long stack_dump_trace[STACK_TRACE_ENTRIES+1] =
> 	 { [0 ... (STACK_TRACE_ENTRIES)] = ULONG_MAX };
> 
>  And
> 
> struct stack_trace stack_trace_max = {
> 	.max_entries		= STACK_TRACE_ENTRIES - 1,
> 	.entries		= &stack_dump_trace[0],
> };
> 
> 
> And nr_entries is set as this, and we have after that this:
> 
> 	stack_trace_max.nr_entries = x;
> 	for (; x < i; x++)
> 		stack_dump_trace[x] = ULONG_MAX;
> 
> Where we set stack_dump_trace[nr_entries] to ULONG_MAX.
> 
> Thus, nr_entries will not go pass the size of stack_dump_trace.
> 
> That said, if n == nr_entries, the second part of that if will always
> be true. And this is a bit subtle, so I will apply the patch. But it is
> not an off by one bug ;-)

Ah, yes.  I follow that now.  Thanks for taking the time to review this
patch.

I am optimistic that eventually I will fix how Smatch handles loops so
it maybe will be able to figure out that "x <= STACK_TRACE_ENTRIES - 1"
but that's probably some time off.

regards,
dan carpenter