Re: [PATCH] [FOR 4.2] tmem: add matching unlock for an about-to-be-destroyed object

[Keir Fraser <keir.xen@gmail.com>]

On 31/08/2012 20:58, "Dan Magenheimer" <dan.magenheimer@oracle.com> wrote:

> (Guidance appreciated if the Xen patch submittal process
> has changed and I need to do more than send this email.)

This is fine, thanks!

 -- Keir

> A 4.2 changeset forces a preempt_disable/enable with
> every lock/unlock.
> 
> Tmem has dynamically allocated "objects" that contain a
> lock.  The lock is held when the object is destroyed.
> No reason to unlock something that's about to be destroyed!
> But with the preempt_enable/disable in the generic locking code,
> and the fact that do_softirq ASSERTs that preempt_count
> must be zero, a crash occurs soon after any object is
> destroyed.
> 
> So force lock to be released before destroying objects.
> 
> Signed-off-by: Dan Magenheimer <dan.magenheimer@oracle.com>
> 
> diff -r 1967c7c290eb xen/common/tmem.c
> --- a/xen/common/tmem.c Wed Feb 09 12:03:09 2011 +0000
> +++ b/xen/common/tmem.c Fri Aug 31 13:49:51 2012 -0600
> @@ -957,6 +957,7 @@
>      /* use no_rebalance only if all objects are being destroyed anyway */
>      if ( !no_rebalance )
>          rb_erase(&obj->rb_tree_node,&pool->obj_rb_root[oid_hash(&old_oid)]);
> +    tmem_spin_unlock(&obj->obj_spinlock);
>      tmem_free(obj,sizeof(obj_t),pool);
>  }
>