Re: [PATCH] proc: do not allow negative offsets on /proc/<pid>/environ

[Djalal Harouni <tixxdz@opendz.org>]

On Mon, Jul 23, 2012 at 05:49:27PM +0200, Oleg Nesterov wrote:
> Hi Djalal,
> 
> On 07/23, Djalal Harouni wrote:
> >
> > Hi Oleg,
> >
> > On Sun, Jul 22, 2012 at 10:00:49PM +0200, Oleg Nesterov wrote:
> > >
> > > Probablt the patch makes sense, but I can't understand the changelog...
> > >
> > > > Allowing negative offsets on /proc/<pid>/environ can turn it to act like
> > > > /proc/<pid>/mem. A negative offset will pass the
> > > > fs/read_write.c:lseek_execute() and the environ_read() checks and will
> > > > point to another VMA.
> > >
> > > which VMA?
> > It depends on the offset. Please see below.
> >
> > > environ_read() can only read the memory from [env_start, env_end], and
> > > it should check *ppos anyway to ensure it doesn't read something else.
> > Yes I agree, but currently that's not the case, there are no checks on *ppos.
>                                                   ^^^^^^^^^^^^^^^^^^^
> There is, unless I missed something, just it is buggy, no?
Right.

> > So if you pass a negative offset you will be able to read from an arbitrary
> > address.
> >
> > [...snip...]
> >
> >   inside environ_read() there is only a one check:
> >
> >   int this_len = mm->env_end - (mm->env_start + src);
> >
> >   if (this_len <= 0)
> >     break;
> >
> >
> >   Here 'src' is 'src = *ppos' the negative offset converted to unsigned long
> >   and (mm->env_start + src) can overflow and point to another VMA.
> >
> >   int this_len = mm->env_end - (mm->env_start + src)
> >
> >   'this_len' will be positive and we pass that check.
> 
> OK, thanks, but doesn't this mean that this check should be fixed
> to avoid the overflow, no matter what *ppos is?
Yes, we must also fix it and check if we are in the [env_start, env_end]
range.

> With or without FMODE_UNSIGNED_OFFSET change. And perhaps it is
> possible to trigger the overflow even with the positive *ppos,
> because:
> 
> > I also don't like the truncation of the result to 'int this_len'
> 
> Yes.
> 
> > BTW should I resend the patch with a better changelog entry ?
> 
> Up to you, but I think this makes sense ;)
> 
> > I'll also add another patch to check the offsets inside environ_read().
> 
> Yes, agreed, but please see above.
> 
> Please correct me, but afaics this patch should come 1st and fix the bug.
> FMODE_UNSIGNED_OFFSET change can be considered as a cleanup after that.
> 
> What do you think?
I agree, that makes more sense and we do not hide the *ppos bug.

I'll resend the two patches soon (fix and check *ppos + this patch with an
updated changelog).

Thanks.

-- 
tixxdz
http://opendz.org