Messages in this thread |  | | | From | Linus Torvalds <> | | Date | Sun, 3 Jul 2011 12:10:29 -0700 | | Subject | Re: [RFC v1] implement SL*B and stack usercopy runtime checks |
| |
On Sun, Jul 3, 2011 at 11:57 AM, Vasiliy Kulikov <segoon@openwall.com> wrote: >> If you seriously clean it up (that at a minimum includes things like >> making it configurable using some pretty helper function that just >> compiles away for all the normal cases, > > Hm, it is not as simple as it looks at the first glance - even if the > object size is known at the compile time (__compiletime_object_size), it > might be a field of a structure, which crosses the slab object > boundaries because of an overflow.
No, I was more talking about having something like
#ifdef CONFIG_EXPENSIVE_CHECK_USERCOPY extern int check_user_copy(const void *kptr, unsigned long size); #else static inline int check_user_copy(const void *kptr, unsigned long size) { return 0; } #endif so that the actual user-copy routines end up being clean and not have #ifdefs in them or any implementation details like what you check (stack, slab, page cache - whatever)
If you can also make it automatically not generate any code for cases that are somehow obviously safe, then that's an added bonus.
But my concern is that performance is a real issue, and the strict user-copy checking sounds like mostly a "let's enable this for testing kernels when chasing some particular issue" feature, the way DEBUG_PAGEALLOC is. And at the same time, code cleanliness and maintainability is a big deal, so the usercopy code itself should have minimal impact and look nice regardless (which is why I strongly object to that kind of "(!slab_access_ok(to, n) || !stack_access_ok(to, n))" crud - the internal details of what you check are *totally* irrelevant to the usercopy code. Linus
|  |