Skip to contentSkip to conversation

HT scheduler, sched-2.5.59-D7

PATCH Manfred Spraul 2 participants 2 messages Jan 20, 2003

lore
Share & tools
lore
#0 lore raw

fs/pipe.c line 123

Quoted 12 lines · Ingo
fs/pipe.c renamed from fs/pipe.c.orig +1 -1
@@ -117,7 +117,7 @@
 	up(PIPE_SEM(*inode));
 	/* Signal writers asynchronously that there is more room.  */
 	if (do_wakeup) {
-		wake_up_interruptible(PIPE_WAIT(*inode));
+		wake_up_interruptible_sync(PIPE_WAIT(*inode));
 		kill_fasync(PIPE_FASYNC_WRITERS(*inode), SIGIO, POLL_OUT);
 	}
 	if (ret > 0)

What's the purpose of this change?
I thought that the _sync functions should be called if it's guaranteed that schedule() will be called immediately, i.e. if the scheduler should not rebalance. You've added _sync() to the codepaths that lead to the end of the syscall.

--
Manfred

#1 lore raw
Quoted 5 lines · Manfred Spraul

Review

Quoted 3 lines
	if (do_wakeup) {
(quoted diff) +1 -1
-		wake_up_interruptible(PIPE_WAIT(*inode));
+		wake_up_interruptible_sync(PIPE_WAIT(*inode));

What's the purpose of this change?

It was a quick experiment i forgot to remove from my sources, will fix it in the next patch.

Ingo