Re: [PATCH v3 1/2] kasan: catch invalid free before SLUB reinitializes the object

[Andrey Konovalov]

On Fri, Jul 26, 2024 at 3:52 PM Jann Horn <jannh@xxxxxxxxxx> wrote:
>
> > Do we still need this patch?
>
> I just tried removing this patch from the series; without it, the
> kmem_cache_invalid_free kunit test fails because the kmem_cache_free()
> no longer synchronously notices that the pointer is misaligned. I
> guess I could change the testcase like this to make the tests pass
> without this patch, but I'd like to hear from you or another KASAN
> person whether you think that's a reasonable change:

Ah, I see. I think detecting a bug earlier if we can is better. So I
don't mind keeping this patch, was just confused by the commit
message.

Adding on top of my comments from before: I think if you move
check_slab_free() out of poison_slab_object() (but add to
__kasan_mempool_poison_object()), and move is_kfence_address() and
kasan_arch_is_ready() to poison_slab_object()'s callers, you won't
even need the free_validation_result enum, so the patch should become
simpler.

You can also rename check_slab_free() to check_slab_allocation() to
make it be named similarly to the already existing
check_page_allocation(). (I think we should also later move
kasan_arch_is_ready() out of check_page_allocation() into the
high-level hooks for consistency; it also seems cleaner to have all of
these ignore checks in the high-level functions instead of lower-level
inlined ones.)

Thanks!