xen/pvcalls: use GFP_ATOMIC under spin lock
authorWei Yongjun <weiyongjun1@huawei.com>
Thu, 28 Dec 2017 03:46:48 +0000 (03:46 +0000)
committerBoris Ostrovsky <boris.ostrovsky@oracle.com>
Tue, 2 Jan 2018 15:02:59 +0000 (10:02 -0500)
A spin lock is taken here so we should use GFP_ATOMIC.

Fixes: 9774c6cca266 ("xen/pvcalls: implement accept command")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
drivers/xen/pvcalls-front.c

index d1e1d8d2b9d545b00eb87085b6f941f04900cd74..4c789e61554b2d3a9d9c01c4384a5d77a2b978d4 100644 (file)
@@ -805,7 +805,7 @@ int pvcalls_front_accept(struct socket *sock, struct socket *newsock, int flags)
                pvcalls_exit();
                return ret;
        }
-       map2 = kzalloc(sizeof(*map2), GFP_KERNEL);
+       map2 = kzalloc(sizeof(*map2), GFP_ATOMIC);
        if (map2 == NULL) {
                clear_bit(PVCALLS_FLAG_ACCEPT_INFLIGHT,
                          (void *)&map->passive.flags);