Re: [PATCH v3 03/57] locking: Introduce __cleanup() based infrastructure

[Kees Cook]

On Tue, Jun 13, 2023 at 12:55:22PM +0200, Peter Zijlstra wrote:
> On Mon, Jun 12, 2023 at 11:07:16AM +0200, Peter Zijlstra wrote:
> 
> > --- /dev/null
> > +++ b/include/linux/cleanup.h
> > @@ -0,0 +1,167 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +#ifndef __LINUX_GUARDS_H
> > +#define __LINUX_GUARDS_H
> > +
> > +#include <linux/compiler.h>
> > +
> > +/*
> > + * DEFINE_FREE(name, type, free):
> > + *	simple helper macro that defines the required wrapper for a __free()
> > + *	based cleanup function. @free is an expression using '_T' to access
> > + *	the variable.
> > + *
> > + * __free(name):
> > + *	variable attribute to add a scoped based cleanup to the variable.
> > + *
> 
> 	no_free_ptr(var):
> 	  like a non-atomic xchg(var, NULL), such that the cleanup
> 	  function will be inhibited -- provided it sanely deals with a
> 	  NULL value.
> 
> > + * return_ptr(p):
> > + *	returns p while inhibiting the __free().
> > + *
> > + * Ex.
> > + *
> > + * DEFINE_FREE(kfree, void *, if (_T) kfree(_T))
> > + *
> > + *	struct obj *p = kmalloc(...);
> 
> That should obviously have been:
> 
> 	struct obj *p __free(kfree) = kmalloc(...);

Okay, good. I noticed that too. :)

I also didn't see any users of the kfree helpers (maybe I missed them?)
Maybe add at least one conversion just to have an example conversion for
it too?

-Kees

-- 
Kees Cook