Messages in this thread | | | From | "J. R. Okajima" <> | Subject | Re: Linux 3.15 .. and continuation of merge window | Date | Thu, 12 Jun 2014 02:56:05 +0900 |
| |
Al Viro: > The former guarantees that the address we are doing trylock on would be that > of a live dentry. The latter makes sure that anything assigned to > dentry->d_parent after we drop ->d_lock will not be freed until we drop > rcu_read_lock.
Although I don't know how to reproduce the problem for the latter case, I think the patch is good.
J. R. Okajima
| |