NFSv4: nfs4_proc_async_renew should use a GFP_NOFS allocation
Trond Myklebust [Wed, 24 Aug 2011 19:07:35 +0000 (15:07 -0400)]
We shouldn't allow the renew daemon to do direct reclaim on the NFS
partition.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

fs/nfs/nfs4proc.c

index 8c77039..776b41a 100644 (file)
@@ -3397,7 +3397,7 @@ int nfs4_proc_async_renew(struct nfs_client *clp, struct rpc_cred *cred)
 
        if (!atomic_inc_not_zero(&clp->cl_count))
                return -EIO;
-       data = kmalloc(sizeof(*data), GFP_KERNEL);
+       data = kmalloc(sizeof(*data), GFP_NOFS);
        if (data == NULL)
                return -ENOMEM;
        data->client = clp;