On Mon, Apr 15, 2019 at 9:17 AM Josh Poimboeuf <jpoimboe@xxxxxxxxxx> wrote:
>
> On Mon, Apr 15, 2019 at 06:07:44PM +0200, Thomas Gleixner wrote:
> > On Mon, 15 Apr 2019, Josh Poimboeuf wrote:
> > > On Mon, Apr 15, 2019 at 11:02:58AM +0200, Thomas Gleixner wrote:
> > > > addr = (unsigned long *)&((char *)addr)[obj_offset(cachep)];
> > > >
> > > > - if (size < 5 * sizeof(unsigned long))
> > > > + if (size < 5)
> > > > return;
> > > >
> > > > *addr++ = 0x12345678;
> > > > *addr++ = caller;
> > > > *addr++ = smp_processor_id();
> > > > - size -= 3 * sizeof(unsigned long);
> > > > + size -= 3;
> > > > +#ifdef CONFIG_STACKTRACE
> > > > {
> > > > - unsigned long *sptr = &caller;
> > > > - unsigned long svalue;
> > > > -
> > > > - while (!kstack_end(sptr)) {
> > > > - svalue = *sptr++;
> > > > - if (kernel_text_address(svalue)) {
> > > > - *addr++ = svalue;
> > > > - size -= sizeof(unsigned long);
> > > > - if (size <= sizeof(unsigned long))
> > > > - break;
> > > > - }
> > > > - }
> > > > + struct stack_trace trace = {
> > > > + /* Leave one for the end marker below */
> > > > + .max_entries = size - 1,
> > > > + .entries = addr,
> > > > + .skip = 3,
> > > > + };
> > > >
> > > > + save_stack_trace(&trace);
> > > > + addr += trace.nr_entries;
> > > > }
> > > > - *addr++ = 0x87654321;
> > > > +#endif
> > > > + *addr = 0x87654321;
> > >
> > > Looks like stack_trace.nr_entries isn't initialized? (though this code
> > > gets eventually replaced by a later patch)
> >
> > struct initializer initialized the non mentioned fields to 0, if I'm not
> > totally mistaken.
>
> Hm, it seems you are correct. And I thought I knew C.
>
> > > Who actually reads this stack trace? I couldn't find a consumer.
> >
> > It's stored directly in the memory pointed to by @addr and that's the freed
> > cache memory. If that is used later (UAF) then the stack trace can be
> > printed to see where it was freed.
>
> Right... but who reads it?
That seems like a reasonable question. After some grepping and some
git searching, it looks like there might not be any users. I found
SLAB_STORE_USER, but that seems to be independent.
So maybe the whole mess should just be deleted. If anyone ever
notices, they can re-add it better.
--Andy
