X-Git-Url: http://pilppa.org/gitweb/gitweb.cgi?a=blobdiff_plain;f=mm%2Fmempolicy.c;h=3c3601121509d8a3c1d3ab2a8cf4a35953a9aa4e;hb=b7624ec1cfaa1218320faa00a061b9891ed28997;hp=c1592a94582f8c2950a6a4cf34711f02a35066c6;hpb=786d3693f46579c7cd982e65de9f43eba94e4a57;p=linux-2.6-omap-h63xx.git

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index c1592a94582..3c360112150 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -116,22 +116,51 @@ static void mpol_rebind_policy(struct mempolicy *pol,
 /* Do sanity checking on a policy */
 static int mpol_check_policy(int mode, nodemask_t *nodes)
 {
-	int empty = nodes_empty(*nodes);
+	int was_empty, is_empty;
+
+	if (!nodes)
+		return 0;
+
+	/*
+	 * "Contextualize" the in-coming nodemast for cpusets:
+	 * Remember whether in-coming nodemask was empty,  If not,
+	 * restrict the nodes to the allowed nodes in the cpuset.
+	 * This is guaranteed to be a subset of nodes with memory.
+	 */
+	cpuset_update_task_memory_state();
+	is_empty = was_empty = nodes_empty(*nodes);
+	if (!was_empty) {
+		nodes_and(*nodes, *nodes, cpuset_current_mems_allowed);
+		is_empty = nodes_empty(*nodes);	/* after "contextualization" */
+	}
 
 	switch (mode) {
 	case MPOL_DEFAULT:
-		if (!empty)
+		/*
+		 * require caller to specify an empty nodemask
+		 * before "contextualization"
+		 */
+		if (!was_empty)
 			return -EINVAL;
 		break;
 	case MPOL_BIND:
 	case MPOL_INTERLEAVE:
-		/* Preferred will only use the first bit, but allow
-		   more for now. */
-		if (empty)
+		/*
+		 * require at least 1 valid node after "contextualization"
+		 */
+		if (is_empty)
+			return -EINVAL;
+		break;
+	case MPOL_PREFERRED:
+		/*
+		 * Did caller specify invalid nodes?
+		 * Don't silently accept this as "local allocation".
+		 */
+		if (!was_empty && is_empty)
 			return -EINVAL;
 		break;
 	}
- 	return nodes_subset(*nodes, node_states[N_HIGH_MEMORY]) ? 0 : -EINVAL;
+	return 0;
 }
 
 /* Generate a custom zonelist for the BIND policy. */
@@ -188,8 +217,6 @@ static struct mempolicy *mpol_new(int mode, nodemask_t *nodes)
 	switch (mode) {
 	case MPOL_INTERLEAVE:
 		policy->v.nodes = *nodes;
-		nodes_and(policy->v.nodes, policy->v.nodes,
-					node_states[N_HIGH_MEMORY]);
 		if (nodes_weight(policy->v.nodes) == 0) {
 			kmem_cache_free(policy_cache, policy);
 			return ERR_PTR(-EINVAL);
@@ -421,18 +448,6 @@ static int mbind_range(struct vm_area_struct *vma, unsigned long start,
 	return err;
 }
 
-static int contextualize_policy(int mode, nodemask_t *nodes)
-{
-	if (!nodes)
-		return 0;
-
-	cpuset_update_task_memory_state();
-	if (!cpuset_nodes_subset_current_mems_allowed(*nodes))
-		return -EINVAL;
-	return mpol_check_policy(mode, nodes);
-}
-
-
 /*
  * Update task->flags PF_MEMPOLICY bit: set iff non-default
  * mempolicy.  Allows more rapid checking of this (combined perhaps
@@ -468,7 +483,7 @@ static long do_set_mempolicy(int mode, nodemask_t *nodes)
 {
 	struct mempolicy *new;
 
-	if (contextualize_policy(mode, nodes))
+	if (mpol_check_policy(mode, nodes))
 		return -EINVAL;
 	new = mpol_new(mode, nodes);
 	if (IS_ERR(new))
@@ -722,12 +737,29 @@ out:
 
 }
 
+/*
+ * Allocate a new page for page migration based on vma policy.
+ * Start assuming that page is mapped by vma pointed to by @private.
+ * Search forward from there, if not.  N.B., this assumes that the
+ * list of pages handed to migrate_pages()--which is how we get here--
+ * is in virtual address order.
+ */
 static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
 {
 	struct vm_area_struct *vma = (struct vm_area_struct *)private;
+	unsigned long uninitialized_var(address);
 
-	return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma,
-					page_address_in_vma(page, vma));
+	while (vma) {
+		address = page_address_in_vma(page, vma);
+		if (address != -EFAULT)
+			break;
+		vma = vma->vm_next;
+	}
+
+	/*
+	 * if !vma, alloc_page_vma() will use task or system default policy
+	 */
+	return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
 }
 #else
 
@@ -898,10 +930,6 @@ asmlinkage long sys_mbind(unsigned long start, unsigned long len,
 	err = get_nodes(&nodes, nmask, maxnode);
 	if (err)
 		return err;
-#ifdef CONFIG_CPUSETS
-	/* Restrict the nodes to the allowed nodes in the cpuset */
-	nodes_and(nodes, nodes, current->mems_allowed);
-#endif
 	return do_mbind(start, len, mode, &nodes, flags);
 }
 
@@ -1268,7 +1296,9 @@ struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr,
 		unsigned nid;
 
 		nid = interleave_nid(pol, vma, addr, HPAGE_SHIFT);
-		__mpol_free(pol);		/* finished with pol */
+		if (unlikely(pol != &default_policy &&
+				pol != current->mempolicy))
+			__mpol_free(pol);	/* finished with pol */
 		return NODE_DATA(nid)->node_zonelists + gfp_zone(gfp_flags);
 	}
 
@@ -1332,6 +1362,9 @@ alloc_page_vma(gfp_t gfp, struct vm_area_struct *vma, unsigned long addr)
 		unsigned nid;
 
 		nid = interleave_nid(pol, vma, addr, PAGE_SHIFT);
+		if (unlikely(pol != &default_policy &&
+				pol != current->mempolicy))
+			__mpol_free(pol);	/* finished with pol */
 		return alloc_page_interleave(gfp, 0, nid);
 	}
 	zl = zonelist_policy(gfp, pol);
@@ -1968,7 +2001,7 @@ int show_numa_map(struct seq_file *m, void *v)
 
 	if (file) {
 		seq_printf(m, " file=");
-		seq_path(m, file->f_path.mnt, file->f_path.dentry, "\n\t= ");
+		seq_path(m, &file->f_path, "\n\t= ");
 	} else if (vma->vm_start <= mm->brk && vma->vm_end >= mm->start_brk) {
 		seq_printf(m, " heap");
 	} else if (vma->vm_start <= mm->start_stack &&