When an application is mixing sync and async IO it ends up crashing
later on in the sync path when calling lock_page_slow as the io_wait
queue has been set to NULL in a previous AIO request.

  Therefore after the retry method has been called the task io_wait
queue should be set to the default queue.

  This patch applies over Suparna's wait-bit patchset:

	- modify-wait-bit-action-args
	- lock_page_wait
	- init-wait-bit-key
	- tsk-default-io-wait
	- aio-wait-bit
	- aio-wait-page

and could be folded into aio-wait-bit.

Signed-off-by: Sébastien Dugué <sebastien.dugue@bull.net>


 aio.c |    2 +-
 1 files changed, 1 insertion(+), 1 deletion(-)

Index: linux-2.6.12/fs/aio.c
===================================================================
--- linux-2.6.12.orig/fs/aio.c	2005-06-30 17:48:47.000000000 +0200
+++ linux-2.6.12/fs/aio.c	2005-06-30 17:48:57.000000000 +0200
@@ -714,7 +714,7 @@
 	BUG_ON(!is_sync_wait(current->io_wait));
 	current->io_wait = &iocb->ki_wait.wait;
 	ret = retry(iocb);
-	current->io_wait = NULL;
+	current->io_wait = &current->__wait.wait;
 
 	if (-EIOCBRETRY != ret) {
  		if (-EIOCBQUEUED != ret) {