mm,thp: introduce flush_pmd_tlb_range

ARCHes with special requirements for evicting THP backing TLB entries
can implement this.

Otherwise also, it can help optimize TLB flush in THP regime.
stock flush_tlb_range() typically has optimization to nuke the entire
TLB if flush span is greater than a certain threshhold, which will
likely be true for a single huge page. Thus a single thp flush will
invalidate the entrire TLB which is not desirable.

e.g. see arch/arc: flush_pmd_tlb_range
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Link: http://lkml.kernel.org/r/20151009100816.GC7873@nodeSigned-off-by: Vineet Gupta <vgupta@synopsys.com>

mm/huge_memory.c

@@ -1880,7 +1880,7 @@ static int __split_huge_page_map(struct page *page,
 		 * here). But it is generally safer to never allow
 		 * small and huge TLB entries for the same virtual
 		 * address to be loaded simultaneously. So instead of
-		 * doing "pmd_populate(); flush_tlb_range();" we first
+		 * doing "pmd_populate(); flush_pmd_tlb_range();" we first
 		 * mark the current pmd notpresent (atomically because
 		 * here the pmd_trans_huge and pmd_trans_splitting
 		 * must remain set at all times on the pmd until the

mm/pgtable-generic.c

@@ -84,6 +84,20 @@ pte_t ptep_clear_flush(struct vm_area_struct *vma, unsigned long address,
 
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 
+#ifndef __HAVE_ARCH_FLUSH_PMD_TLB_RANGE
+
+/*
+ * ARCHes with special requirements for evicting THP backing TLB entries can
+ * implement this. Otherwise also, it can help optimize normal TLB flush in
+ * THP regime. stock flush_tlb_range() typically has optimization to nuke the
+ * entire TLB TLB if flush span is greater than a threshhold, which will
+ * likely be true for a single huge page. Thus a single thp flush will
+ * invalidate the entire TLB which is not desitable.
+ * e.g. see arch/arc: flush_pmd_tlb_range
+ */
+#define flush_pmd_tlb_range(vma, addr, end)	flush_tlb_range(vma, addr, end)
+#endif
+
 #ifndef __HAVE_ARCH_PMDP_SET_ACCESS_FLAGS
 int pmdp_set_access_flags(struct vm_area_struct *vma,
 			  unsigned long address, pmd_t *pmdp,
@@ -93,7 +107,7 @@ int pmdp_set_access_flags(struct vm_area_struct *vma,
 	VM_BUG_ON(address & ~HPAGE_PMD_MASK);
 	if (changed) {
 		set_pmd_at(vma->vm_mm, address, pmdp, entry);
-		flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+		flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 	}
 	return changed;
 }
@@ -107,7 +121,7 @@ int pmdp_clear_flush_young(struct vm_area_struct *vma,
 	VM_BUG_ON(address & ~HPAGE_PMD_MASK);
 	young = pmdp_test_and_clear_young(vma, address, pmdp);
 	if (young)
-		flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+		flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 	return young;
 }
 #endif
@@ -120,7 +134,7 @@ pmd_t pmdp_huge_clear_flush(struct vm_area_struct *vma, unsigned long address,
 	VM_BUG_ON(address & ~HPAGE_PMD_MASK);
 	VM_BUG_ON(!pmd_trans_huge(*pmdp));
 	pmd = pmdp_huge_get_and_clear(vma->vm_mm, address, pmdp);
-	flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+	flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 	return pmd;
 }
 #endif
@@ -133,7 +147,7 @@ void pmdp_splitting_flush(struct vm_area_struct *vma, unsigned long address,
 	VM_BUG_ON(address & ~HPAGE_PMD_MASK);
 	set_pmd_at(vma->vm_mm, address, pmdp, pmd);
 	/* tlb flush only to serialize against gup-fast */
-	flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+	flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 }
 #endif
 
@@ -179,7 +193,7 @@ void pmdp_invalidate(struct vm_area_struct *vma, unsigned long address,
 {
 	pmd_t entry = *pmdp;
 	set_pmd_at(vma->vm_mm, address, pmdp, pmd_mknotpresent(entry));
-	flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+	flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 }
 #endif
 
@@ -196,7 +210,7 @@ pmd_t pmdp_collapse_flush(struct vm_area_struct *vma, unsigned long address,
 	VM_BUG_ON(address & ~HPAGE_PMD_MASK);
 	VM_BUG_ON(pmd_trans_huge(*pmdp));
 	pmd = pmdp_huge_get_and_clear(vma->vm_mm, address, pmdp);
-	flush_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
+	flush_pmd_tlb_range(vma, address, address + HPAGE_PMD_SIZE);
 	return pmd;
 }
 #endif