diff mbox

[RFC,1/9] mm/hugetlb: Make GENERAL_HUGETLB functions PGD implementation aware

Message ID 1457525450-4262-1-git-send-email-khandual@linux.vnet.ibm.com (mailing list archive)
State RFC
Headers show

Commit Message

Anshuman Khandual March 9, 2016, 12:10 p.m. UTC
Currently both the ARCH_WANT_GENERAL_HUGETLB functions 'huge_pte_alloc'
and 'huge_pte_offset' dont take into account huge page implementation
at the PGD level. With addition of PGD awareness into these functions,
more architectures like POWER which also implements huge pages at PGD
level (along with PMD level), can use ARCH_WANT_GENERAL_HUGETLB option.

Signed-off-by: Anshuman Khandual <khandual@linux.vnet.ibm.com>
---
 mm/hugetlb.c | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Anshuman Khandual March 11, 2016, 3:01 a.m. UTC | #1
On 03/09/2016 05:40 PM, Anshuman Khandual wrote:
> Currently both the ARCH_WANT_GENERAL_HUGETLB functions 'huge_pte_alloc'
> and 'huge_pte_offset' dont take into account huge page implementation
> at the PGD level. With addition of PGD awareness into these functions,
> more architectures like POWER which also implements huge pages at PGD
> level (along with PMD level), can use ARCH_WANT_GENERAL_HUGETLB option.

Hugh/Mel/Naoya/Andrew,

	Thoughts/inputs/suggestions ? Does this change looks okay ?
Andrew Morton March 14, 2016, 8:29 p.m. UTC | #2
On Fri, 11 Mar 2016 08:31:55 +0530 Anshuman Khandual <khandual@linux.vnet.ibm.com> wrote:

> On 03/09/2016 05:40 PM, Anshuman Khandual wrote:
> > Currently both the ARCH_WANT_GENERAL_HUGETLB functions 'huge_pte_alloc'
> > and 'huge_pte_offset' dont take into account huge page implementation
> > at the PGD level. With addition of PGD awareness into these functions,
> > more architectures like POWER which also implements huge pages at PGD
> > level (along with PMD level), can use ARCH_WANT_GENERAL_HUGETLB option.
> 
> Hugh/Mel/Naoya/Andrew,
> 
> 	Thoughts/inputs/suggestions ? Does this change looks okay ?

Patches 1, 2 and 3 look OK to me.  Please include them in the powerpc
merge when the patchset is considered ready.
diff mbox

Patch

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 01f2b48..a478b7b 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -4251,6 +4251,11 @@  pte_t *huge_pte_alloc(struct mm_struct *mm,
 	pte_t *pte = NULL;
 
 	pgd = pgd_offset(mm, addr);
+	if (sz == PGDIR_SIZE) {
+		pte = (pte_t *)pgd;
+		goto huge_pgd;
+	}
+
 	pud = pud_alloc(mm, pgd, addr);
 	if (pud) {
 		if (sz == PUD_SIZE) {
@@ -4263,6 +4268,8 @@  pte_t *huge_pte_alloc(struct mm_struct *mm,
 				pte = (pte_t *)pmd_alloc(mm, pud, addr);
 		}
 	}
+
+huge_pgd:
 	BUG_ON(pte && !pte_none(*pte) && !pte_huge(*pte));
 
 	return pte;
@@ -4276,6 +4283,8 @@  pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
 
 	pgd = pgd_offset(mm, addr);
 	if (pgd_present(*pgd)) {
+		if (pgd_huge(*pgd))
+			return (pte_t *)pgd;
 		pud = pud_offset(pgd, addr);
 		if (pud_present(*pud)) {
 			if (pud_huge(*pud))