missing lock in percpu_rwlock? (Was: Re: New Defects reported by Coverity Scan for XenProject)

[Ian Campbell <ian.campbell@citrix.com>]

On Tue, 2016-02-02 at 20:23 -0800, scan-admin@coverity.com wrote:
> * CID 1351223:  Concurrent data access violations  (MISSING_LOCK)
> /xen/include/xen/spinlock.h: 362 in _percpu_write_unlock()

Coverity seems to think this is new in 41b0aa569adb..9937763265d,
presumably due to 

commit f9dd43dddc0a31a4343a58072935c1b5c0cbbee
Author: Malcolm Crossley <malcolm.crossley@citrix.com>
Date:   Fri Jan 22 16:04:41 2016 +0100

    rwlock: add per-cpu reader-writer lock infrastructure

> _________________________________________________________________________
> _______________________________
> *** CID 1351223:  Concurrent data access violations  (MISSING_LOCK)
> /xen/include/xen/spinlock.h: 362 in _percpu_write_unlock()
> 356                     percpu_rwlock_t *percpu_rwlock)
> 357     {
> 358         /* Validate the correct per_cpudata variable has been
> provided. */
> 359         _percpu_rwlock_owner_check(per_cpudata, percpu_rwlock);
> 360     
> 361         ASSERT(percpu_rwlock->writer_activating);
> >>>     CID 1351223:  Concurrent data access violations  (MISSING_LOCK)
> >>>     Accessing "percpu_rwlock->writer_activating" without holding lock
> "percpu_rwlock.rwlock". Elsewhere, "percpu_rwlock.writer_activating" is
> accessed with "percpu_rwlock.rwlock" held 1 out of 2 times (1 of these
> accesses strongly imply that it is necessary).
> 362         percpu_rwlock->writer_activating = 0;
> 363         write_unlock(&percpu_rwlock->rwlock);
> 364     }
> 365     
> 366     #define percpu_rw_is_write_locked(l)        
> _rw_is_write_locked(&((l)->rwlock))
> 367     

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel