From: Zach Brown <zach.brown@oracle.com>
Date: Mon, 14 Nov 2005 00:07:33 +0000 (-0800)
Subject: [PATCH] aio: remove kioctx from mm_struct
X-Git-Tag: v2.6.15-rc2~155
X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=20dcae32439384b6863c626bb3b2a09bed65b33e;p=linux-2.6

[PATCH] aio: remove kioctx from mm_struct

Sync iocbs have a life cycle that don't need a kioctx.  Their retrying, if
any, is done in the context of their owner who has allocated them on the
stack.

The sole user of a sync iocb's ctx reference was aio_complete() checking for
an elevated iocb ref count that could never happen.  No path which grabs an
iocb ref has access to sync iocbs.

If we were to implement sync iocb cancelation it would be done by the owner of
the iocb using its on-stack reference.

Removing this chunk from aio_complete allows us to remove the entire kioctx
instance from mm_struct, reducing its size by a third.  On a i386 testing box
the slab size went from 768 to 504 bytes and from 5 to 8 per page.

Signed-off-by: Zach Brown <zach.brown@oracle.com>
Acked-by: Benjamin LaHaise <bcrl@kvack.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
---

diff --git a/fs/aio.c b/fs/aio.c
index 20bb919eb1..e7cd40b626 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -937,28 +937,19 @@ int fastcall aio_complete(struct kiocb *iocb, long res, long res2)
 	unsigned long	tail;
 	int		ret;
 
-	/* Special case handling for sync iocbs: events go directly
-	 * into the iocb for fast handling.  Note that this will not 
-	 * work if we allow sync kiocbs to be cancelled. in which
-	 * case the usage count checks will have to move under ctx_lock
-	 * for all cases.
+	/*
+	 * Special case handling for sync iocbs:
+	 *  - events go directly into the iocb for fast handling
+	 *  - the sync task with the iocb in its stack holds the single iocb
+	 *    ref, no other paths have a way to get another ref
+	 *  - the sync task helpfully left a reference to itself in the iocb
 	 */
 	if (is_sync_kiocb(iocb)) {
-		int ret;
-
+		BUG_ON(iocb->ki_users != 1);
 		iocb->ki_user_data = res;
-		if (iocb->ki_users == 1) {
-			iocb->ki_users = 0;
-			ret = 1;
-		} else {
-			spin_lock_irq(&ctx->ctx_lock);
-			iocb->ki_users--;
-			ret = (0 == iocb->ki_users);
-			spin_unlock_irq(&ctx->ctx_lock);
-		}
-		/* sync iocbs put the task here for us */
+		iocb->ki_users = 0;
 		wake_up_process(iocb->ki_obj.tsk);
-		return ret;
+		return 1;
 	}
 
 	info = &ctx->ring_info;
diff --git a/include/linux/aio.h b/include/linux/aio.h
index 403d71dcb7..9e0ae87112 100644
--- a/include/linux/aio.h
+++ b/include/linux/aio.h
@@ -124,7 +124,7 @@ struct kiocb {
 		(x)->ki_users = 1;			\
 		(x)->ki_key = KIOCB_SYNC_KEY;		\
 		(x)->ki_filp = (filp);			\
-		(x)->ki_ctx = &tsk->active_mm->default_kioctx;	\
+		(x)->ki_ctx = NULL;			\
 		(x)->ki_cancel = NULL;			\
 		(x)->ki_dtor = NULL;			\
 		(x)->ki_obj.tsk = tsk;			\
diff --git a/include/linux/init_task.h b/include/linux/init_task.h
index 68ab5f2ab9..dcfd2ecccb 100644
--- a/include/linux/init_task.h
+++ b/include/linux/init_task.h
@@ -51,7 +51,6 @@
 	.page_table_lock =  SPIN_LOCK_UNLOCKED, 		\
 	.mmlist		= LIST_HEAD_INIT(name.mmlist),		\
 	.cpu_vm_mask	= CPU_MASK_ALL,				\
-	.default_kioctx = INIT_KIOCTX(name.default_kioctx, name),	\
 }
 
 #define INIT_SIGNALS(sig) {	\
diff --git a/include/linux/sched.h b/include/linux/sched.h
index 41df813957..2038bd27b0 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -357,7 +357,6 @@ struct mm_struct {
 	/* aio bits */
 	rwlock_t		ioctx_list_lock;
 	struct kioctx		*ioctx_list;
-	struct kioctx		default_kioctx;
 };
 
 struct sighand_struct {
diff --git a/kernel/fork.c b/kernel/fork.c
index 2c70c9cdf5..e0d0b77343 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -323,7 +323,6 @@ static struct mm_struct * mm_init(struct mm_struct * mm)
 	spin_lock_init(&mm->page_table_lock);
 	rwlock_init(&mm->ioctx_list_lock);
 	mm->ioctx_list = NULL;
-	mm->default_kioctx = (struct kioctx)INIT_KIOCTX(mm->default_kioctx, *mm);
 	mm->free_area_cache = TASK_UNMAPPED_BASE;
 	mm->cached_hole_size = ~0UL;