On Sun, Feb 16, 2020 at 03:14:55AM +, Mateusz Guzik wrote:
> Author: mjg
> Date: Sun Feb 16 03:14:55 2020
> New Revision: 357989
> URL: https://svnweb.freebsd.org/changeset/base/357989
>
> Log:
> refcount: add missing release fence to refcount_release_if_gt
I think you should update comment in refcount_release_last() to point
to if_gt() as well. And probably annotate this fence with a pointer
to acquire in release_last().
>
> The CPU succeeding in releasing the not last reference can still have
> pending
> stores to the object protected by the affected counter. This opens a time
> window where another CPU can release the last reference and free the object,
> resulting in use-after-free. On top of that this prevents the compiler from
> generating more accesses to the object regardless of how
> atomic_fcmpset_rel_int
> is implemented (of course as long as it provides the release semantic).
>
> Reviewed by:markj
>
> Modified:
> head/sys/sys/refcount.h
>
> Modified: head/sys/sys/refcount.h
> ==
> --- head/sys/sys/refcount.h Sun Feb 16 01:07:19 2020(r357988)
> +++ head/sys/sys/refcount.h Sun Feb 16 03:14:55 2020(r357989)
> @@ -198,7 +198,7 @@ refcount_release_if_gt(volatile u_int *count, u_int n)
> return (false);
> if (__predict_false(REFCOUNT_SATURATED(old)))
> return (true);
> - if (atomic_fcmpset_int(count, &old, old - 1))
> + if (atomic_fcmpset_rel_int(count, &old, old - 1))
> return (true);
> }
> }
___
svn-src-all@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"
