Re: [Bug 216073] New: [s390x] kernel BUG at mm/usercopy.c:101! usercopy: Kernel memory exposure attempt detected from vmalloc 'n o area' (offset 0, size 1)!

[Matthew Wilcox <willy@infradead.org>]

On Sun, Jun 12, 2022 at 03:03:20PM +0200, Uladzislau Rezki wrote:
> > @@ -181,8 +181,9 @@ static inline void check_heap_object(const void *ptr, unsigned long n,
> >  			return;
> >  		}
> >  
> > -		offset = ptr - area->addr;
> > -		if (offset + n > get_vm_area_size(area))
> > +		/* XXX: We should also abort for free vmap_areas */
> > +		offset = (unsigned long)ptr - area->va_start;
> >
> I was a bit confused about "offset" and why it is needed here. It is always zero. 
> So we can get rid of it to make it less confused. From the other hand a zero offset
> contributes to nothing.

I don't think offset is necessarily zero.  'ptr' is a pointer somewhere
in the object, not necessarily the start of the object.

> >
> > +		if (offset + n >= area->va_end)
> >
> I think it is a bit wrong. As i see it, "n" is a size and what we would like to do
> here is boundary check:
> 
> <snip>
> if (n > va_size(area))
>     usercopy_abort("vmalloc", NULL, to_user, 0, n);
> <snip>

Hmm ... we should probably be more careful about wrapping.

                if (n > area->va_end - addr)
                        usercopy_abort("vmalloc", NULL, to_user, offset, n);

... and that goes for the whole function actually.  I'll split that into
a separate change.