[PATCH] add nowarn to a few pte chain allocators

From: Arjan van de Ven <arjanv@redhat.com>

Several of the pte_chain_alloc() allocators that use GFP_ATOMIC have a
fallback for failure that sleeps; they thus need to not warn on failure..
Seen during a big fork on a busy system.

mm/memory.c

@@ -222,7 +222,7 @@ int copy_page_range(struct mm_struct *dst, struct mm_struct *src,
 	if (is_vm_hugetlb_page(vma))
 		return copy_hugetlb_page_range(dst, src, vma);
 
-	pte_chain = pte_chain_alloc(GFP_ATOMIC);
+	pte_chain = pte_chain_alloc(GFP_ATOMIC | __GFP_NOWARN);
 	if (!pte_chain) {
 		spin_unlock(&dst->page_table_lock);
 		pte_chain = pte_chain_alloc(GFP_KERNEL);
@@ -335,7 +335,7 @@ skip_copy_pmd_range:	address = (address + PGDIR_SIZE) & PGDIR_MASK;
 							pte_chain);
 				if (pte_chain)
 					goto cont_copy_pte_range_noset;
-				pte_chain = pte_chain_alloc(GFP_ATOMIC);
+				pte_chain = pte_chain_alloc(GFP_ATOMIC | __GFP_NOWARN);
 				if (pte_chain)
 					goto cont_copy_pte_range_noset;
 
@@ -1336,7 +1336,7 @@ do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma,
 	struct pte_chain *pte_chain;
 	int ret;
 
-	pte_chain = pte_chain_alloc(GFP_ATOMIC);
+	pte_chain = pte_chain_alloc(GFP_ATOMIC | __GFP_NOWARN);
 	if (!pte_chain) {
 		pte_unmap(page_table);
 		spin_unlock(&mm->page_table_lock);