Re: [PATCH v2] mm: Add kvfree_sensitive() for freeing sensitive data objects

From: Joe Perches <joe@perches.com>
To: Waiman Long <longman@redhat.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	David Howells <dhowells@redhat.com>,
	Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>,
	James Morris <jmorris@namei.org>,
	"Serge E. Hallyn" <serge@hallyn.com>
Cc: linux-mm@kvack.org, keyrings@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Matthew Wilcox <willy@infradead.org>,
	David Rientjes <rientjes@google.com>
Subject: Re: [PATCH v2] mm: Add kvfree_sensitive() for freeing sensitive data objects
Date: Tue, 07 Apr 2020 06:41:23 +0000	[thread overview]
Message-ID: <07e49a285eff9a22476c6b1c396485f6d5d39002.camel@perches.com> (raw)
In-Reply-To: <1e4a6174-04be-6c05-fd6e-b43fefd317fc@redhat.com>

On Mon, 2020-04-06 at 22:16 -0400, Waiman Long wrote:
> On 4/6/20 3:38 PM, Joe Perches wrote:
> > On Mon, 2020-04-06 at 14:58 -0400, Waiman Long wrote:
> > > For kvmalloc'ed data object that contains sensitive information like
> > > cryptographic key, we need to make sure that the buffer is always
> > > cleared before freeing it. Using memset() alone for buffer clearing may
> > > not provide certainty as the compiler may compile it away. To be sure,
> > > the special memzero_explicit() has to be used.
> > [] 
> > >  extern void kvfree(const void *addr);
> > > +extern void kvfree_sensitive(const void *addr, size_t len);
> > Question: why should this be const?
> > 
> > 2.1.44 changed kfree(void *) to kfree(const void *) but
> > I didn't find a particular reason why.
> 
> I am just following the function prototype used by kvfree(). Even
> kzfree(const void *) use const. I can remove "const" if others agree.

No worries.  Nevermind me...

Lots of warnings if allocated pointers are const, so const is necessary
in the definition and declaration.

struct foo {
	...
};

struct bar {
	const struct foo *baz;
	...
};

some_func(void)
{
	bar.baz = kvalloc(...);
}

kvfree can't free bar.baz if it's defined with void * without warning,
so it must be const void *.

Apologies for the noise.