summaryrefslogtreecommitdiff
path: root/fs/eventpoll.c
diff options
context:
space:
mode:
authorArjan van de Ven <arjan@infradead.org>2006-03-23 14:00:32 +0300
committerLinus Torvalds <torvalds@g5.osdl.org>2006-03-23 18:38:12 +0300
commit144efe3e3e5ad57af549bf800fa4560d7c74e9fe (patch)
tree332dd6364fe5122f489ab4352817f11737e7a258 /fs/eventpoll.c
parent70522e121a521aa09bd0f4e62e1aa68708b798e1 (diff)
downloadlinux-144efe3e3e5ad57af549bf800fa4560d7c74e9fe.tar.xz
[PATCH] sem2mutex: eventpoll
Semaphore to mutex conversion. The conversion was generated via scripts, and the result was validated automatically via a script as well. Signed-off-by: Arjan van de Ven <arjan@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'fs/eventpoll.c')
-rw-r--r--fs/eventpoll.c23
1 files changed, 12 insertions, 11 deletions
diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 4284cd31eba6..f5d69f46ba9b 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -34,6 +34,7 @@
#include <linux/eventpoll.h>
#include <linux/mount.h>
#include <linux/bitops.h>
+#include <linux/mutex.h>
#include <asm/uaccess.h>
#include <asm/system.h>
#include <asm/io.h>
@@ -46,7 +47,7 @@
* LOCKING:
* There are three level of locking required by epoll :
*
- * 1) epsem (semaphore)
+ * 1) epmutex (mutex)
* 2) ep->sem (rw_semaphore)
* 3) ep->lock (rw_lock)
*
@@ -67,9 +68,9 @@
* if a file has been pushed inside an epoll set and it is then
* close()d without a previous call toepoll_ctl(EPOLL_CTL_DEL).
* It is possible to drop the "ep->sem" and to use the global
- * semaphore "epsem" (together with "ep->lock") to have it working,
+ * semaphore "epmutex" (together with "ep->lock") to have it working,
* but having "ep->sem" will make the interface more scalable.
- * Events that require holding "epsem" are very rare, while for
+ * Events that require holding "epmutex" are very rare, while for
* normal operations the epoll private "ep->sem" will guarantee
* a greater scalability.
*/
@@ -274,7 +275,7 @@ static struct super_block *eventpollfs_get_sb(struct file_system_type *fs_type,
/*
* This semaphore is used to serialize ep_free() and eventpoll_release_file().
*/
-static struct semaphore epsem;
+static struct mutex epmutex;
/* Safe wake up implementation */
static struct poll_safewake psw;
@@ -477,10 +478,10 @@ void eventpoll_release_file(struct file *file)
* cleanup path, and this means that noone is using this file anymore.
* The only hit might come from ep_free() but by holding the semaphore
* will correctly serialize the operation. We do need to acquire
- * "ep->sem" after "epsem" because ep_remove() requires it when called
+ * "ep->sem" after "epmutex" because ep_remove() requires it when called
* from anywhere but ep_free().
*/
- down(&epsem);
+ mutex_lock(&epmutex);
while (!list_empty(lsthead)) {
epi = list_entry(lsthead->next, struct epitem, fllink);
@@ -492,7 +493,7 @@ void eventpoll_release_file(struct file *file)
up_write(&ep->sem);
}
- up(&epsem);
+ mutex_unlock(&epmutex);
}
@@ -819,9 +820,9 @@ static void ep_free(struct eventpoll *ep)
* We do not need to hold "ep->sem" here because the epoll file
* is on the way to be removed and no one has references to it
* anymore. The only hit might come from eventpoll_release_file() but
- * holding "epsem" is sufficent here.
+ * holding "epmutex" is sufficent here.
*/
- down(&epsem);
+ mutex_lock(&epmutex);
/*
* Walks through the whole tree by unregistering poll callbacks.
@@ -843,7 +844,7 @@ static void ep_free(struct eventpoll *ep)
ep_remove(ep, epi);
}
- up(&epsem);
+ mutex_unlock(&epmutex);
}
@@ -1615,7 +1616,7 @@ static int __init eventpoll_init(void)
{
int error;
- init_MUTEX(&epsem);
+ mutex_init(&epmutex);
/* Initialize the structure used to perform safe poll wait head wake ups */
ep_poll_safewake_init(&psw);