Re: [PATCH 1/3] compiler.h: add ASSERT_STATIC_STORAGE()

[Yury Norov <ynorov@nvidia.com>]

On Fri, Sep 11, 2026 at 10:57:49PM -0700, Andrew Morton wrote:
> On Fri, 11 Sep 2026 18:14:41 -0400 Yury Norov <yury.norov@gmail.com> wrote:
> 
> > Static lock initializers rely on a persistent object address when lockdep
> > assigns a lock-class key. Using such an initializer for an automatic local
> > object can compile successfully but disable lockdep on the first lock
> > acquisition.
> > 
> > Add ASSERT_STATIC_STORAGE() for declaration macros that require
> > static storage duration. It declares an unused static pointer initialized
> > with the object's address. An automatic local object's address is not a
> > valid static initializer, so the compiler rejects it.
> > 
> > Mirror the helper in tools/include/linux/compiler.h because the userspace
> > radix-tree tests include the kernel IDA and Maple Tree headers with the
> > tools compiler definitions.
> > 
> > For example:
> > 
> >   void example(void)
> >   {
> >           int object;
> >           ASSERT_STATIC_STORAGE(object);
> >   }
> > 
> > GCC reports:
> > 
> >   error: initializer element is not constant
> >     name##_storage_check = &(name)
> >                            ^
> >   note: in expansion of macro 'ASSERT_STATIC_STORAGE'
> >     ASSERT_STATIC_STORAGE(object);
> > 
> > File-scope objects and static local objects remain valid. The helper takes
> > an object identifier and must be used as a declaration after that object
> > has been declared.
> > 
> 
> Thanks, I'll queue it.
> 
> Sashiko wants the same treatment for a few other macros:
> 	https://sashiko.dev/#/patchset/20260911221444.1523311-1-ynorov@nvidia.com

Below is the list of candidates that my AI finds relevant:

 /* Primitive locks */
  DEFINE_RAW_SPINLOCK()
  DEFINE_SPINLOCK()                /* both non-RT and RT definitions */
  DEFINE_RWLOCK()                  /* both non-RT and RT definitions */
  DEFINE_MUTEX()                   /* both non-RT and RT definitions */
  DEFINE_RT_MUTEX()
  DECLARE_RWSEM()                  /* both non-RT and RT definitions */
  DEFINE_SEMAPHORE()
  DEFINE_SEQLOCK()

  /* Wait queues and completions — preserve _ONSTACK aliases */
  DECLARE_WAIT_QUEUE_HEAD()
  DECLARE_SWAIT_QUEUE_HEAD()
  DECLARE_COMPLETION()

  /* Containers */
  DEFINE_XARRAY_FLAGS()            /* covers DEFINE_XARRAY(),
                                   * DEFINE_XARRAY_ALLOC(),
                                   * DEFINE_XARRAY_ALLOC1() */
  DEFINE_IDA()                     /* already enforced */
  DEFINE_MTREE()                   /* already enforced */
  DEFINE_KLIST()

  /* Work and timers */
  DECLARE_WORK()
  DECLARE_DELAYED_WORK()
  DECLARE_DEFERRABLE_WORK()
  DEFINE_TIMER()
  DEFINE_KTHREAD_DELAYED_WORK()

  /* Other objects containing locks */
  DEFINE_RATELIMIT_STATE()
  ATOMIC_NOTIFIER_HEAD()
  BLOCKING_NOTIFIER_HEAD()
  DEFINE_RCU_SYNC()
  DEFINE_CTL_TABLE_POLL()
  DEFINE_BPF_STORAGE_CACHE()

  Lower-priority or configuration-dependent candidates:

  DEFINE_IDR()                    /* normal operations don't use embedded lock */
  RADIX_TREE()                    /* normal operations don't use embedded lock */

  __DEFINE_PERCPU_RWSEM()          /* covers DEFINE_PERCPU_RWSEM(),
                                   * DEFINE_STATIC_PERCPU_RWSEM() */

  __DEFINE_SRCU()                  /* Tree SRCU: covers all six wrappers */
  DEFINE_SRCU()                    /* Tiny SRCU: shared non-static declaration */

  _SRCU_NOTIFIER_HEAD()            /* covers SRCU_NOTIFIER_HEAD(),
                                   * SRCU_NOTIFIER_HEAD_STATIC() */

I planned to submit this hardening for them per the relevant subsystems
after at least landing the ASSERT_STATIC_STORAGE() in -next.

If you're OK to move that material through your tree all in one,
I can prepare it as a single series.

Thanks,
Yury