net/rose: Use GFP_ATOMIC

Author: Julia Lawall <julia@diku.dk>

The other calls to kmalloc in the same function use GFP_ATOMIC, and indeed
two locks are held within the body of the function.

The semantic match that finds this problem is as follows:
(http://coccinelle.lip6.fr/)

// 
@@ identifier f; @@

*f(...,GFP_ATOMIC,...)
... when != spin_unlock(...)
    when != read_unlock(...)
    when != write_unlock(...)
    when != read_unlock_irq(...)
    when != write_unlock_irq(...)
    when != read_unlock_irqrestore(...)
    when != write_unlock_irqrestore(...)
    when != spin_unlock_irq(...)
    when != spin_unlock_irqrestore(...)
*f(...,GFP_KERNEL,...)
// 

Signed-off-by: Julia Lawall 
Signed-off-by: David S. Miller 

---
 net/rose/rose_route.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
 
diff --git a/net/rose/rose_route.c b/net/rose/rose_route.c
index cbc244a1..b4fdaac 100644
--- a/net/rose/rose_route.c
+++ b/net/rose/rose_route.c
@@ -109,7 +109,9 @@ static int __must_check rose_add_node(struct rose_route_struct *rose_route,
 		init_timer(&rose_neigh->t0timer);
 
 		if (rose_route->ndigis != 0) {
-			if ((rose_neigh->digipeat = kmalloc(sizeof(ax25_digi), GFP_KERNEL)) == NULL) {
+			rose_neigh->digipeat =
+				kmalloc(sizeof(ax25_digi), GFP_ATOMIC);
+			if (rose_neigh->digipeat == NULL) {
 				kfree(rose_neigh);
 				res = -ENOMEM;
 				goto out;