Re: [PATCH RFC 8/8] kasan: add and integrate kasan_mode boot param
From: Andrey Konovalov
Date: Fri Oct 16 2020 - 09:10:59 EST
On Thu, Oct 15, 2020 at 3:56 PM Marco Elver <elver@xxxxxxxxxx> wrote:
>
> On Wed, 14 Oct 2020 at 22:45, Andrey Konovalov <andreyknvl@xxxxxxxxxx> wrote:
> >
[...]
> > @@ -180,6 +182,7 @@ size_t kasan_metadata_size(struct kmem_cache *cache)
> > struct kasan_alloc_meta *kasan_get_alloc_meta(struct kmem_cache *cache,
> > const void *object)
> > {
> > + WARN_ON(!static_branch_unlikely(&kasan_debug));
>
> The WARN_ON condition itself should be unlikely, so that would imply
> that the static branch here should be likely since you're negating it.
Here I was thinking that we should optimize for the production use
case, which shouldn't have kasan_debug enabled, hence the unlikely.
But technically this function shouldn't be called in production
anyway, so likely will do fine too.
> And AFAIK, this function should only be called if kasan_debug is true.
Yes, this WARN_ON is to make sure this doesn't happen.
[...]
> > +/* Whether to use syncronous or asynchronous tag checking. */
> > +static bool kasan_sync __ro_after_init;
>
> s/syncronous/synchronous/
Ack.
>
> > +static int __init early_kasan_mode(char *arg)
> > +{
> > + if (!arg)
> > + return -EINVAL;
> > +
> > + if (strcmp(arg, "on") == 0)
> > + kasan_mode = KASAN_MODE_ON;
> > + else if (strcmp(arg, "debug") == 0)
>
> s/strcmp(..) == 0/!strcmp(..)/ ?
Sounds good.
[...]
> > @@ -60,6 +111,7 @@ void kasan_set_free_info(struct kmem_cache *cache,
> > {
> > struct kasan_alloc_meta *alloc_meta;
> >
> > + WARN_ON(!static_branch_unlikely(&kasan_debug));
>
> What actually happens if any of these are called with !kasan_debug and
> the warning triggers? Is it still valid to execute the below, or
> should it bail out? Or possibly even disable KASAN entirely?
It shouldn't happen, but if it happens maybe it indeed makes sense to
disable KASAN here is a failsafe. It might be tricky to disable MTE
though, but I'll see what we can do here.
Thank you!