aboutsummaryrefslogtreecommitdiffstats
path: root/include/linux/dlm_plock.h
Commit message (Collapse)AuthorAge
* dlm: make plock operation killableDavid Teigland2011-05-23
| | | | | | | | | | | | | | | | | | | | | | | | Allow processes blocked on plock requests to be interrupted when they are killed. This leaves the problem of cleaning up the lock state in userspace. This has three parts: 1. Add a flag to unlock operations sent to userspace indicating the file is being closed. Userspace will then look for and clear any waiting plock operations that were abandoned by an interrupted process. 2. Queue an unlock-close operation (like in 1) to clean up userspace from an interrupted plock request. This is needed because the vfs will not send a cleanup-unlock if it sees no locks on the file, which it won't if the interrupted operation was the only one. 3. Do not use replies from userspace for unlock-close operations because they are unnecessary (they are just cleaning up for the process which did not make an unlock call). This also simplifies the new unlock-close generated from point 2. Signed-off-by: David Teigland <teigland@redhat.com>
* headers_check fix: linux/dlm_plock.hJaswinder Singh Rajput2009-01-30
| | | | | | | | fix the following 'make headers_check' warning: usr/include/linux/dlm_plock.h:25: found __[us]{8,16,32,64} type without #include <linux/types.h> Signed-off-by: Jaswinder Singh Rajput <jaswinderrajput@gmail.com>
* dlm: move plock code from gfs2David Teigland2008-04-21
Move the code that handles cluster posix locks from gfs2 into the dlm so that it can be used by both gfs2 and ocfs2. Signed-off-by: David Teigland <teigland@redhat.com>