提交 18ccee26 编写于 作者: D Davidlohr Bueso 提交者: Linus Torvalds

ipc: update locking scheme comments

The initial documentation was a bit incomplete, update accordingly.

[akpm@linux-foundation.org: make it more readable in 80 columns]
Signed-off-by: NDavidlohr Bueso <davidlohr@hp.com>
Acked-by: NManfred Spraul <manfred@colorfullife.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>
上级 9c567512
......@@ -17,12 +17,27 @@
* Pavel Emelianov <xemul@openvz.org>
*
* General sysv ipc locking scheme:
* when doing ipc id lookups, take the ids->rwsem
* rcu_read_lock()
* obtain the ipc object (kern_ipc_perm)
* perform security, capabilities, auditing and permission checks, etc.
* acquire the ipc lock (kern_ipc_perm.lock) throught ipc_lock_object()
* perform data updates (ie: SET, RMID, LOCK/UNLOCK commands)
* rcu_read_lock()
* obtain the ipc object (kern_ipc_perm) by looking up the id in an idr
* tree.
* - perform initial checks (capabilities, auditing and permission,
* etc).
* - perform read-only operations, such as STAT, INFO commands.
* acquire the ipc lock (kern_ipc_perm.lock) through
* ipc_lock_object()
* - perform data updates, such as SET, RMID commands and
* mechanism-specific operations (semop/semtimedop,
* msgsnd/msgrcv, shmat/shmdt).
* drop the ipc lock, through ipc_unlock_object().
* rcu_read_unlock()
*
* The ids->rwsem must be taken when:
* - creating, removing and iterating the existing entries in ipc
* identifier sets.
* - iterating through files under /proc/sysvipc/
*
* Note that sems have a special fast path that avoids kern_ipc_perm.lock -
* see sem_lock().
*/
#include <linux/mm.h>
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册