receive_from_sock() is never called in atomic context.
The call chain ending up at receive_from_sock() is:
[1] receive_from_sock() <- process_recv_sockets()
process_recv_sockets() is only set as a parameter of INIT_WORK()
And receive_from_sock() also calls mutex_lock(), which indicates
this function is not called in atomic context.
Despite never getting called from atomic context,
receive_from_sock() calls alloc_page() with GFP_ATOMIC,
which waits busily for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL,
to avoid busy waiting and improve the possibility of sucessful allocation.
This is found by a static analysis tool named DCNS written by myself.
And I also manually check it.
Signed-off-by: Jia-Ju Bai <baijiaju1...@gmail.com>
---
fs/dlm/lowcomms.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 4813d0e..2d4e230 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -633,7 +633,7 @@ static int receive_from_sock(struct connection *con)
* This doesn't need to be atomic, but I think it should
* improve performance if it is.
*/
- con->rx_page = alloc_page(GFP_ATOMIC);
+ con->rx_page = alloc_page(GFP_KERNEL);
if (con->rx_page == NULL)
goto out_resched;
cbuf_init(&con->cb, PAGE_SIZE);
--
1.9.1
