x86-64: Fix page table accounting
authorJan Beulich <JBeulich@suse.com>
Thu, 4 Oct 2012 13:48:10 +0000 (14:48 +0100)
committerIngo Molnar <mingo@kernel.org>
Wed, 24 Oct 2012 08:50:25 +0000 (10:50 +0200)
Commit 20167d3421a089a1bf1bd680b150dc69c9506810 ("x86-64: Fix
accounting in kernel_physical_mapping_init()") went a little too
far by entirely removing the counting of pre-populated page
tables: this should be done at boot time (to cover the page
tables set up in early boot code), but shouldn't be done during
memory hot add.

Hence, re-add the removed increments of "pages", but make them
and the one in phys_pte_init() conditional upon !after_bootmem.

Reported-Acked-and-Tested-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Cc: <stable@kernel.org>
Link: http://lkml.kernel.org/r/506DAFBA020000780009FA8C@nat28.tlf.novell.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
arch/x86/mm/init_64.c

index 2b6b4a3c8beb8727d27b1cdba22411fe0d271efb..3baff255adac6472570298d7e9ee60df75e93dfc 100644 (file)
@@ -386,7 +386,8 @@ phys_pte_init(pte_t *pte_page, unsigned long addr, unsigned long end,
                 * these mappings are more intelligent.
                 */
                if (pte_val(*pte)) {
-                       pages++;
+                       if (!after_bootmem)
+                               pages++;
                        continue;
                }
 
@@ -451,6 +452,8 @@ phys_pmd_init(pmd_t *pmd_page, unsigned long address, unsigned long end,
                         * attributes.
                         */
                        if (page_size_mask & (1 << PG_LEVEL_2M)) {
+                               if (!after_bootmem)
+                                       pages++;
                                last_map_addr = next;
                                continue;
                        }
@@ -526,6 +529,8 @@ phys_pud_init(pud_t *pud_page, unsigned long addr, unsigned long end,
                         * attributes.
                         */
                        if (page_size_mask & (1 << PG_LEVEL_1G)) {
+                               if (!after_bootmem)
+                                       pages++;
                                last_map_addr = next;
                                continue;
                        }