usercopy: Make usercopy resilient against ridiculously large copies
authorMatthew Wilcox (Oracle) <willy@infradead.org>
Sun, 12 Jun 2022 21:32:27 +0000 (22:32 +0100)
committerKees Cook <keescook@chromium.org>
Mon, 13 Jun 2022 16:54:52 +0000 (09:54 -0700)
If 'n' is so large that it's negative, we might wrap around and mistakenly
think that the copy is OK when it's not.  Such a copy would probably
crash, but just doing the arithmetic in a more simple way lets us detect
and refuse this case.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Tested-by: Zorro Lang <zlang@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20220612213227.3881769-4-willy@infradead.org
mm/usercopy.c

index 30a4db3..4e1da70 100644 (file)
@@ -162,27 +162,26 @@ static inline void check_heap_object(const void *ptr, unsigned long n,
                                     bool to_user)
 {
        uintptr_t addr = (uintptr_t)ptr;
+       unsigned long offset;
        struct folio *folio;
 
        if (is_kmap_addr(ptr)) {
-               unsigned long page_end = addr | (PAGE_SIZE - 1);
-
-               if (addr + n - 1 > page_end)
-                       usercopy_abort("kmap", NULL, to_user,
-                                       offset_in_page(ptr), n);
+               offset = offset_in_page(ptr);
+               if (n > PAGE_SIZE - offset)
+                       usercopy_abort("kmap", NULL, to_user, offset, n);
                return;
        }
 
        if (is_vmalloc_addr(ptr)) {
                struct vmap_area *area = find_vmap_area(addr);
-               unsigned long offset;
 
                if (!area)
                        usercopy_abort("vmalloc", "no area", to_user, 0, n);
 
-               offset = addr - area->va_start;
-               if (addr + n > area->va_end)
+               if (n > area->va_end - addr) {
+                       offset = addr - area->va_start;
                        usercopy_abort("vmalloc", NULL, to_user, offset, n);
+               }
                return;
        }
 
@@ -195,8 +194,8 @@ static inline void check_heap_object(const void *ptr, unsigned long n,
                /* Check slab allocator for flags and size. */
                __check_heap_object(ptr, n, folio_slab(folio), to_user);
        } else if (folio_test_large(folio)) {
-               unsigned long offset = ptr - folio_address(folio);
-               if (offset + n > folio_size(folio))
+               offset = ptr - folio_address(folio);
+               if (n > folio_size(folio) - offset)
                        usercopy_abort("page alloc", NULL, to_user, offset, n);
        }
 }