maps4: add /proc/kpagecount interface
authorMatt Mackall <mpm@selenic.com>
Tue, 5 Feb 2008 06:29:05 +0000 (22:29 -0800)
committerLinus Torvalds <torvalds@woody.linux-foundation.org>
Tue, 5 Feb 2008 17:44:17 +0000 (09:44 -0800)
This makes physical page map counts available to userspace. Together
with /proc/pid/pagemap and /proc/pid/clear_refs, this can be used to
monitor memory usage on a per-page basis.

[akpm@linux-foundation.org: remove unneeded access_ok()]
[bunk@stusta.de: make struct proc_kpagemap static]
Signed-off-by: Matt Mackall <mpm@selenic.com>
Cc: Jeremy Fitzhardinge <jeremy@goop.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/proc/proc_misc.c

index 3462bfde89f68224a4ec48629e26c81feee0e44b..19b69f931bef0b237acd1832303ed2a658e4364a 100644 (file)
@@ -46,6 +46,7 @@
 #include <linux/vmalloc.h>
 #include <linux/crash_dump.h>
 #include <linux/pid_namespace.h>
+#include <linux/bootmem.h>
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/io.h>
@@ -675,6 +676,54 @@ static const struct file_operations proc_sysrq_trigger_operations = {
 };
 #endif
 
+#define KPMSIZE sizeof(u64)
+#define KPMMASK (KPMSIZE - 1)
+/* /proc/kpagecount - an array exposing page counts
+ *
+ * Each entry is a u64 representing the corresponding
+ * physical page count.
+ */
+static ssize_t kpagecount_read(struct file *file, char __user *buf,
+                            size_t count, loff_t *ppos)
+{
+       u64 __user *out = (u64 __user *)buf;
+       struct page *ppage;
+       unsigned long src = *ppos;
+       unsigned long pfn;
+       ssize_t ret = 0;
+       u64 pcount;
+
+       pfn = src / KPMSIZE;
+       count = min_t(size_t, count, (max_pfn * KPMSIZE) - src);
+       if (src & KPMMASK || count & KPMMASK)
+               return -EIO;
+
+       while (count > 0) {
+               ppage = pfn_to_page(pfn++);
+               if (!ppage)
+                       pcount = 0;
+               else
+                       pcount = atomic_read(&ppage->_count);
+
+               if (put_user(pcount, out++)) {
+                       ret = -EFAULT;
+                       break;
+               }
+
+               count -= KPMSIZE;
+       }
+
+       *ppos += (char __user *)out - buf;
+       if (!ret)
+               ret = (char __user *)out - buf;
+       return ret;
+}
+
+static struct file_operations proc_kpagecount_operations = {
+       .llseek = mem_lseek,
+       .read = kpagecount_read,
+};
+
 struct proc_dir_entry *proc_root_kcore;
 
 void create_seq_entry(char *name, mode_t mode, const struct file_operations *f)
@@ -755,6 +804,7 @@ void __init proc_misc_init(void)
                                (size_t)high_memory - PAGE_OFFSET + PAGE_SIZE;
        }
 #endif
+       create_seq_entry("kpagecount", S_IRUSR, &proc_kpagecount_operations);
 #ifdef CONFIG_PROC_VMCORE
        proc_vmcore = create_proc_entry("vmcore", S_IRUSR, NULL);
        if (proc_vmcore)