[PATCH] Fold numa_maps into mempolicies.c
Christoph Lameter [Sun, 8 Jan 2006 09:01:02 +0000 (01:01 -0800)]
First discussed at http://marc.theaimsgroup.com/?t=113149255100001&r=1&w=2

- Use the check_range() in mempolicy.c to gather statistics.

- Improve the numa_maps code in general and fix some comments.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

fs/proc/task_mmu.c
mm/mempolicy.c

index 50bd5a8..0eaad41 100644 (file)
@@ -390,129 +390,12 @@ struct seq_operations proc_pid_smaps_op = {
 };
 
 #ifdef CONFIG_NUMA
-
-struct numa_maps {
-       unsigned long pages;
-       unsigned long anon;
-       unsigned long mapped;
-       unsigned long mapcount_max;
-       unsigned long node[MAX_NUMNODES];
-};
-
-/*
- * Calculate numa node maps for a vma
- */
-static struct numa_maps *get_numa_maps(struct vm_area_struct *vma)
-{
-       int i;
-       struct page *page;
-       unsigned long vaddr;
-       struct numa_maps *md = kmalloc(sizeof(struct numa_maps), GFP_KERNEL);
-
-       if (!md)
-               return NULL;
-       md->pages = 0;
-       md->anon = 0;
-       md->mapped = 0;
-       md->mapcount_max = 0;
-       for_each_node(i)
-               md->node[i] =0;
-
-       for (vaddr = vma->vm_start; vaddr < vma->vm_end; vaddr += PAGE_SIZE) {
-               page = follow_page(vma, vaddr, 0);
-               if (page) {
-                       int count = page_mapcount(page);
-
-                       if (count)
-                               md->mapped++;
-                       if (count > md->mapcount_max)
-                               md->mapcount_max = count;
-                       md->pages++;
-                       if (PageAnon(page))
-                               md->anon++;
-                       md->node[page_to_nid(page)]++;
-               }
-               cond_resched();
-       }
-       return md;
-}
-
-static int show_numa_map(struct seq_file *m, void *v)
-{
-       struct task_struct *task = m->private;
-       struct vm_area_struct *vma = v;
-       struct mempolicy *pol;
-       struct numa_maps *md;
-       struct zone **z;
-       int n;
-       int first;
-
-       if (!vma->vm_mm)
-               return 0;
-
-       md = get_numa_maps(vma);
-       if (!md)
-               return 0;
-
-       seq_printf(m, "%08lx", vma->vm_start);
-       pol = get_vma_policy(task, vma, vma->vm_start);
-       /* Print policy */
-       switch (pol->policy) {
-       case MPOL_PREFERRED:
-               seq_printf(m, " prefer=%d", pol->v.preferred_node);
-               break;
-       case MPOL_BIND:
-               seq_printf(m, " bind={");
-               first = 1;
-               for (z = pol->v.zonelist->zones; *z; z++) {
-
-                       if (!first)
-                               seq_putc(m, ',');
-                       else
-                               first = 0;
-                       seq_printf(m, "%d/%s", (*z)->zone_pgdat->node_id,
-                                       (*z)->name);
-               }
-               seq_putc(m, '}');
-               break;
-       case MPOL_INTERLEAVE:
-               seq_printf(m, " interleave={");
-               first = 1;
-               for_each_node(n) {
-                       if (node_isset(n, pol->v.nodes)) {
-                               if (!first)
-                                       seq_putc(m,',');
-                               else
-                                       first = 0;
-                               seq_printf(m, "%d",n);
-                       }
-               }
-               seq_putc(m, '}');
-               break;
-       default:
-               seq_printf(m," default");
-               break;
-       }
-       seq_printf(m, " MaxRef=%lu Pages=%lu Mapped=%lu",
-                       md->mapcount_max, md->pages, md->mapped);
-       if (md->anon)
-               seq_printf(m," Anon=%lu",md->anon);
-
-       for_each_online_node(n) {
-               if (md->node[n])
-                       seq_printf(m, " N%d=%lu", n, md->node[n]);
-       }
-       seq_putc(m, '\n');
-       kfree(md);
-       if (m->count < m->size)  /* vma is copied successfully */
-               m->version = (vma != get_gate_vma(task)) ? vma->vm_start : 0;
-       return 0;
-}
+extern int show_numa_map(struct seq_file *m, void *v);
 
 struct seq_operations proc_pid_numa_maps_op = {
-       .start  = m_start,
-       .next   = m_next,
-       .stop   = m_stop,
-       .show   = show_numa_map
+        .start  = m_start,
+        .next   = m_next,
+        .stop   = m_stop,
+        .show   = show_numa_map
 };
 #endif
index 270e9a3..44b9d69 100644 (file)
@@ -84,6 +84,8 @@
 #include <linux/compat.h>
 #include <linux/mempolicy.h>
 #include <linux/swap.h>
+#include <linux/seq_file.h>
+#include <linux/proc_fs.h>
 
 #include <asm/tlbflush.h>
 #include <asm/uaccess.h>
@@ -91,6 +93,7 @@
 /* Internal flags */
 #define MPOL_MF_DISCONTIG_OK (MPOL_MF_INTERNAL << 0)   /* Skip checks for continuous vmas */
 #define MPOL_MF_INVERT (MPOL_MF_INTERNAL << 1)         /* Invert check for nodemask */
+#define MPOL_MF_STATS (MPOL_MF_INTERNAL << 2)          /* Gather statistics */
 
 static kmem_cache_t *policy_cache;
 static kmem_cache_t *sn_cache;
@@ -228,6 +231,8 @@ static void migrate_page_add(struct vm_area_struct *vma,
        }
 }
 
+static void gather_stats(struct page *, void *);
+
 /* Scan through pages checking if pages follow certain conditions. */
 static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
                unsigned long addr, unsigned long end,
@@ -252,7 +257,9 @@ static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
                if (node_isset(nid, *nodes) == !!(flags & MPOL_MF_INVERT))
                        continue;
 
-               if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
+               if (flags & MPOL_MF_STATS)
+                       gather_stats(page, private);
+               else if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
                        migrate_page_add(vma, page, private, flags);
                else
                        break;
@@ -1460,3 +1467,132 @@ void numa_policy_rebind(const nodemask_t *old, const nodemask_t *new)
 {
        rebind_policy(current->mempolicy, old, new);
 }
+
+/*
+ * Display pages allocated per node and memory policy via /proc.
+ */
+
+static const char *policy_types[] = { "default", "prefer", "bind",
+                                     "interleave" };
+
+/*
+ * Convert a mempolicy into a string.
+ * Returns the number of characters in buffer (if positive)
+ * or an error (negative)
+ */
+static inline int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol)
+{
+       char *p = buffer;
+       int l;
+       nodemask_t nodes;
+       int mode = pol ? pol->policy : MPOL_DEFAULT;
+
+       switch (mode) {
+       case MPOL_DEFAULT:
+               nodes_clear(nodes);
+               break;
+
+       case MPOL_PREFERRED:
+               nodes_clear(nodes);
+               node_set(pol->v.preferred_node, nodes);
+               break;
+
+       case MPOL_BIND:
+               get_zonemask(pol, &nodes);
+               break;
+
+       case MPOL_INTERLEAVE:
+               nodes = pol->v.nodes;
+               break;
+
+       default:
+               BUG();
+               return -EFAULT;
+       }
+
+       l = strlen(policy_types[mode]);
+       if (buffer + maxlen < p + l + 1)
+               return -ENOSPC;
+
+       strcpy(p, policy_types[mode]);
+       p += l;
+
+       if (!nodes_empty(nodes)) {
+               if (buffer + maxlen < p + 2)
+                       return -ENOSPC;
+               *p++ = '=';
+               p += nodelist_scnprintf(p, buffer + maxlen - p, nodes);
+       }
+       return p - buffer;
+}
+
+struct numa_maps {
+       unsigned long pages;
+       unsigned long anon;
+       unsigned long mapped;
+       unsigned long mapcount_max;
+       unsigned long node[MAX_NUMNODES];
+};
+
+static void gather_stats(struct page *page, void *private)
+{
+       struct numa_maps *md = private;
+       int count = page_mapcount(page);
+
+       if (count)
+               md->mapped++;
+
+       if (count > md->mapcount_max)
+               md->mapcount_max = count;
+
+       md->pages++;
+
+       if (PageAnon(page))
+               md->anon++;
+
+       md->node[page_to_nid(page)]++;
+       cond_resched();
+}
+
+int show_numa_map(struct seq_file *m, void *v)
+{
+       struct task_struct *task = m->private;
+       struct vm_area_struct *vma = v;
+       struct numa_maps *md;
+       int n;
+       char buffer[50];
+
+       if (!vma->vm_mm)
+               return 0;
+
+       md = kzalloc(sizeof(struct numa_maps), GFP_KERNEL);
+       if (!md)
+               return 0;
+
+       check_pgd_range(vma, vma->vm_start, vma->vm_end,
+                   &node_online_map, MPOL_MF_STATS, md);
+
+       if (md->pages) {
+               mpol_to_str(buffer, sizeof(buffer),
+                           get_vma_policy(task, vma, vma->vm_start));
+
+               seq_printf(m, "%08lx %s pages=%lu mapped=%lu maxref=%lu",
+                          vma->vm_start, buffer, md->pages,
+                          md->mapped, md->mapcount_max);
+
+               if (md->anon)
+                       seq_printf(m," anon=%lu",md->anon);
+
+               for_each_online_node(n)
+                       if (md->node[n])
+                               seq_printf(m, " N%d=%lu", n, md->node[n]);
+
+               seq_putc(m, '\n');
+       }
+       kfree(md);
+
+       if (m->count < m->size)
+               m->version = (vma != get_gate_vma(task)) ? vma->vm_start : 0;
+       return 0;
+}
+