Simplify fifo_open() locking logic
authorLinus Torvalds <torvalds@g5.osdl.org>
Tue, 7 Mar 2006 17:16:35 +0000 (09:16 -0800)
committerLinus Torvalds <torvalds@g5.osdl.org>
Tue, 7 Mar 2006 17:16:35 +0000 (09:16 -0800)
We don't do interruptible waits for the pipe mutex anywhere else any
more either, so don't do it in fifo_open() either.

Acked-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
fs/fifo.c

index 923371b753ab0c04d1d8c57c56072bf40042fa30..d13fcd3ec803bd52e335ef340d09c817421b05fb 100644 (file)
--- a/fs/fifo.c
+++ b/fs/fifo.c
@@ -34,10 +34,7 @@ static int fifo_open(struct inode *inode, struct file *filp)
 {
        int ret;
 
-       ret = -ERESTARTSYS;
-       if (mutex_lock_interruptible(PIPE_MUTEX(*inode)))
-               goto err_nolock_nocleanup;
-
+       mutex_lock(PIPE_MUTEX(*inode));
        if (!inode->i_pipe) {
                ret = -ENOMEM;
                if(!pipe_new(inode))
@@ -140,8 +137,6 @@ err:
 
 err_nocleanup:
        mutex_unlock(PIPE_MUTEX(*inode));
-
-err_nolock_nocleanup:
        return ret;
 }