diff mbox series

[RFC,v2,13/21] x86/pgtable: dont check PMD accessed bit

Message ID 20181226133351.828074959@intel.com (mailing list archive)
State New, archived
Headers show
Series PMEM NUMA node and hotness accounting/migration | expand

Commit Message

Fengguang Wu Dec. 26, 2018, 1:14 p.m. UTC 
From: Jingqi Liu <jingqi.liu@intel.com>

ept-idle will clear PMD accessed bit to speedup PTE scan -- if the bit
remains unset in the next scan, all the 512 PTEs can be skipped.

So don't complain on !_PAGE_ACCESSED in pmd_bad().

Note that clearing PMD accessed bit has its own cost, the optimization
may only be worthwhile for
- large idle area
- sparsely populated area

Signed-off-by: Jingqi Liu <jingqi.liu@intel.com>
Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
---
 arch/x86/include/asm/pgtable.h |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
diff mbox series

Patch

--- linux.orig/arch/x86/include/asm/pgtable.h	2018-12-23 19:50:50.917902600 +0800
+++ linux/arch/x86/include/asm/pgtable.h	2018-12-23 19:50:50.913902605 +0800
@@ -821,7 +821,8 @@  static inline pte_t *pte_offset_kernel(p
 
 static inline int pmd_bad(pmd_t pmd)
 {
-	return (pmd_flags(pmd) & ~_PAGE_USER) != _KERNPG_TABLE;
+	return (pmd_flags(pmd) & ~(_PAGE_USER | _PAGE_ACCESSED)) !=
+					(_KERNPG_TABLE & ~_PAGE_ACCESSED);
 }
 
 static inline unsigned long pages_to_mb(unsigned long npg)