Yes, I think you are right, it might lead to a race condition because it uses flock() already. It is better to leave things as they are. Please skip this patch ! Regards, Guido > On the 20th of April 2017 at 17.56 Stephen Smalley <sds@xxxxxxxxxxxxx> wrote: > > > On Thu, 2017-04-20 at 17:45 +0200, Guido Trentalancia wrote: > > Hello Stephen. > > > > Usually, when a lock file is released, the corresponding file is > > removed from the filesystem for keeping it clean and tidy. > > > > I might be wrong... But why not ? > > > > If nothing is handling the semanage store, then there shouldn't be a > > reason for keeping it locked. The presence of a lock file, usually > > means that the lock is active. > > libsemanage doesn't use the lock files that way; it just uses them as > the object for flock() operations. So the presence of the lock file > means nothing. Removing it just means it will have to be re-created on > the next operation. Not fundamentally opposed, but someone would need > to validate that it doesn't cause any issues. It's been that way > forever. Maybe the original Tresys authors of this code have an > opinion on it.
