The backport of d35c99ff77ec ("netlink: do not enter direct reclaim from
netlink_dump()") to the 4.4 branch (first in 4.4.32) mistakenly removed
direct claim from the initial large allocation _and_ the fallback
allocation which means that allocations can spuriously fail.
Fix the issue by adding back the direct reclaim flag to the fallback
allocation.
Fixes: 6d123f1d396b ("netlink: do not enter direct reclaim from netlink_dump()")
Signed-off-by: Ross Lagerwall <ross.lagerwall@xxxxxxxxxx>
---
Note that this is only for the 4.4 branch as the regression is only in
this branch. Consequently, there is no corresponding upstream commit.
I'm resending this to the linux-stable list since I now understand the
netdev maintainer only handles backports for the last couple of versions
of Linux.
net/netlink/af_netlink.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/net/netlink/af_netlink.c b/net/netlink/af_netlink.c
index 8e33019..acfb16f 100644
--- a/net/netlink/af_netlink.c
+++ b/net/netlink/af_netlink.c
@@ -2107,7 +2107,7 @@ static int netlink_dump(struct sock *sk)
if (!skb) {
alloc_size = alloc_min_size;
skb = netlink_alloc_skb(sk, alloc_size, nlk->portid,
- (GFP_KERNEL & ~__GFP_DIRECT_RECLAIM));
+ GFP_KERNEL);
}
if (!skb)
goto errout_skb;
--
2.7.4