Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1607450

[RFC PATCH 2/2] arm64: hugetlbpages: Correctly handle swap entries in huge_pte_offset()

From Punit Agrawal <punit.agrawal@arm.com>
Newsgroups linux.kernel
Subject [RFC PATCH 2/2] arm64: hugetlbpages: Correctly handle swap entries in huge_pte_offset()
Date 2017-03-23 14:00 +0100
Message-ID <toavo-2lS-19@gated-at.bofh.it> (permalink)
References <toavo-2lS-21@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


huge_pte_offset() does not correctly handle poisoned or migration page
table entries. Not knowing the size of the hugepage entry being
requested only compounded the problem.

The recently added hstate parameter can be used to determine the size of
hugepage being accessed. Use the size to find the correct page table
entry to return when coming across a swap page table entry.

Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Cc: David Woods <dwoods@mellanox.com>
---
 arch/arm64/mm/hugetlbpage.c | 31 ++++++++++++++++---------------
 1 file changed, 16 insertions(+), 15 deletions(-)

diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
index 75d8cc3e138b..db108fa6e197 100644
--- a/arch/arm64/mm/hugetlbpage.c
+++ b/arch/arm64/mm/hugetlbpage.c
@@ -191,38 +191,39 @@ pte_t *huge_pte_alloc(struct mm_struct *mm,
 
 pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr, struct hstate *h)
 {
+	unsigned long sz = huge_page_size(h);
 	pgd_t *pgd;
 	pud_t *pud;
-	pmd_t *pmd = NULL;
-	pte_t *pte = NULL;
+	pmd_t *pmd;
+	pte_t *pte;
 
 	pgd = pgd_offset(mm, addr);
 	pr_debug("%s: addr:0x%lx pgd:%p\n", __func__, addr, pgd);
 	if (!pgd_present(*pgd))
 		return NULL;
+
 	pud = pud_offset(pgd, addr);
-	if (!pud_present(*pud))
+	if (pud_none(*pud) && sz != PUD_SIZE)
 		return NULL;
-
-	if (pud_huge(*pud))
+	else if (!pud_table(*pud))
 		return (pte_t *)pud;
+
+	if (sz == CONT_PMD_SIZE)
+		addr &= CONT_PMD_MASK;
+
 	pmd = pmd_offset(pud, addr);
-	if (!pmd_present(*pmd))
+	if (pmd_none(*pmd) &&
+	    !(sz == PMD_SIZE || sz == CONT_PMD_SIZE))
 		return NULL;
-
-	if (pte_cont(pmd_pte(*pmd))) {
-		pmd = pmd_offset(
-			pud, (addr & CONT_PMD_MASK));
-		return (pte_t *)pmd;
-	}
-	if (pmd_huge(*pmd))
+	else if (!pmd_table(*pmd))
 		return (pte_t *)pmd;
-	pte = pte_offset_kernel(pmd, addr);
-	if (pte_present(*pte) && pte_cont(*pte)) {
+
+	if (sz == CONT_PTE_SIZE) {
 		pte = pte_offset_kernel(
 			pmd, (addr & CONT_PTE_MASK));
 		return pte;
 	}
+
 	return NULL;
 }
 
-- 
2.11.0

Back to linux.kernel | Previous | Next | Find similar | Unroll thread


Thread

[RFC PATCH 2/2] arm64: hugetlbpages: Correctly handle swap entries in huge_pte_offset() Punit Agrawal <punit.agrawal@arm.com> - 2017-03-23 14:00 +0100

csiph-web