[PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Kirill A. Shutemov]

Hi,

The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
to be usable for x86 and switches x86 over.

Please review and consider applying.

Kirill A. Shutemov (7):
  mm: Drop arch_pte_access_permitted() mmu hook
  mm/gup: Move permission checks into helpers
  mm/gup: Move page table entry dereference into helper
  mm/gup: Make pages referenced during generic get_user_pages_fast()
  mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()
  mm/gup: Provide hook to check if __GUP_fast() is allowed for the range
  x86/mm: Switch to generic get_user_page_fast() implementation

 arch/powerpc/include/asm/mmu_context.h   |   6 -
 arch/s390/include/asm/mmu_context.h      |   6 -
 arch/um/include/asm/mmu_context.h        |   6 -
 arch/unicore32/include/asm/mmu_context.h |   6 -
 arch/x86/Kconfig                         |   3 +
 arch/x86/include/asm/mmu_context.h       |  16 -
 arch/x86/include/asm/pgtable-3level.h    |  45 +++
 arch/x86/include/asm/pgtable.h           |  53 ++++
 arch/x86/include/asm/pgtable_64.h        |  16 +-
 arch/x86/mm/Makefile                     |   2 +-
 arch/x86/mm/gup.c                        | 496 -------------------------------
 include/asm-generic/mm_hooks.h           |   6 -
 include/asm-generic/pgtable.h            |  25 ++
 include/linux/mm.h                       |   4 +
 mm/gup.c                                 | 134 +++++++--
 15 files changed, 262 insertions(+), 562 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

-- 
2.11.0

[PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Kirill A. Shutemov]

Hi,

The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
to be usable for x86 and switches x86 over.

Please review and consider applying.

Kirill A. Shutemov (7):
  mm: Drop arch_pte_access_permitted() mmu hook
  mm/gup: Move permission checks into helpers
  mm/gup: Move page table entry dereference into helper
  mm/gup: Make pages referenced during generic get_user_pages_fast()
  mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()
  mm/gup: Provide hook to check if __GUP_fast() is allowed for the range
  x86/mm: Switch to generic get_user_page_fast() implementation

 arch/powerpc/include/asm/mmu_context.h   |   6 -
 arch/s390/include/asm/mmu_context.h      |   6 -
 arch/um/include/asm/mmu_context.h        |   6 -
 arch/unicore32/include/asm/mmu_context.h |   6 -
 arch/x86/Kconfig                         |   3 +
 arch/x86/include/asm/mmu_context.h       |  16 -
 arch/x86/include/asm/pgtable-3level.h    |  45 +++
 arch/x86/include/asm/pgtable.h           |  53 ++++
 arch/x86/include/asm/pgtable_64.h        |  16 +-
 arch/x86/mm/Makefile                     |   2 +-
 arch/x86/mm/gup.c                        | 496 -------------------------------
 include/asm-generic/mm_hooks.h           |   6 -
 include/asm-generic/pgtable.h            |  25 ++
 include/linux/mm.h                       |   4 +
 mm/gup.c                                 | 134 +++++++--
 15 files changed, 262 insertions(+), 562 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 1/7] mm: Drop arch_pte_access_permitted() mmu hook

[Kirill A. Shutemov]

The only arch that defines it to something meaningful is x86.
But x86 doesn't use generic GUP_fast() implementation -- the only place
where the hook is called.

Let's drop it.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 arch/powerpc/include/asm/mmu_context.h   | 6 ------
 arch/s390/include/asm/mmu_context.h      | 6 ------
 arch/um/include/asm/mmu_context.h        | 6 ------
 arch/unicore32/include/asm/mmu_context.h | 6 ------
 arch/x86/include/asm/mmu_context.h       | 4 ----
 include/asm-generic/mm_hooks.h           | 6 ------
 mm/gup.c                                 | 3 ---
 7 files changed, 37 deletions(-)

diff --git a/arch/powerpc/include/asm/mmu_context.h b/arch/powerpc/include/asm/mmu_context.h
index b9e3f0aca261..ecf9885ab660 100644
--- a/arch/powerpc/include/asm/mmu_context.h
+++ b/arch/powerpc/include/asm/mmu_context.h
@@ -163,11 +163,5 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __KERNEL__ */
 #endif /* __ASM_POWERPC_MMU_CONTEXT_H */
diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 6e31d87fb669..fa2bf69be182 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -156,10 +156,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __S390_MMU_CONTEXT_H */
diff --git a/arch/um/include/asm/mmu_context.h b/arch/um/include/asm/mmu_context.h
index 94ac2739918c..b668e351fd6c 100644
--- a/arch/um/include/asm/mmu_context.h
+++ b/arch/um/include/asm/mmu_context.h
@@ -37,12 +37,6 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return true;
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
-
 /*
  * end asm-generic/mm_hooks.h functions
  */
diff --git a/arch/unicore32/include/asm/mmu_context.h b/arch/unicore32/include/asm/mmu_context.h
index 62dfc644c908..59b06b48f27d 100644
--- a/arch/unicore32/include/asm/mmu_context.h
+++ b/arch/unicore32/include/asm/mmu_context.h
@@ -103,10 +103,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 306c7e12af55..68b329d77b3a 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -268,8 +268,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return __pkru_allows_pkey(vma_pkey(vma), write);
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	return __pkru_allows_pkey(pte_flags_pkey(pte_flags(pte)), write);
-}
 #endif /* _ASM_X86_MMU_CONTEXT_H */
diff --git a/include/asm-generic/mm_hooks.h b/include/asm-generic/mm_hooks.h
index cc5d9a1405df..41e5b6784b97 100644
--- a/include/asm-generic/mm_hooks.h
+++ b/include/asm-generic/mm_hooks.h
@@ -32,10 +32,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif	/* _ASM_GENERIC_MM_HOOKS_H */
diff --git a/mm/gup.c b/mm/gup.c
index 04aa405350dc..3f2338ba3402 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1216,9 +1216,6 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			pte_protnone(pte) || (write && !pte_write(pte)))
 			goto pte_unmap;
 
-		if (!arch_pte_access_permitted(pte, write))
-			goto pte_unmap;
-
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
 		page = pte_page(pte);
 		head = compound_head(page);
-- 
2.11.0

[PATCH 1/7] mm: Drop arch_pte_access_permitted() mmu hook

[Kirill A. Shutemov]

The only arch that defines it to something meaningful is x86.
But x86 doesn't use generic GUP_fast() implementation -- the only place
where the hook is called.

Let's drop it.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 arch/powerpc/include/asm/mmu_context.h   | 6 ------
 arch/s390/include/asm/mmu_context.h      | 6 ------
 arch/um/include/asm/mmu_context.h        | 6 ------
 arch/unicore32/include/asm/mmu_context.h | 6 ------
 arch/x86/include/asm/mmu_context.h       | 4 ----
 include/asm-generic/mm_hooks.h           | 6 ------
 mm/gup.c                                 | 3 ---
 7 files changed, 37 deletions(-)

diff --git a/arch/powerpc/include/asm/mmu_context.h b/arch/powerpc/include/asm/mmu_context.h
index b9e3f0aca261..ecf9885ab660 100644
--- a/arch/powerpc/include/asm/mmu_context.h
+++ b/arch/powerpc/include/asm/mmu_context.h
@@ -163,11 +163,5 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __KERNEL__ */
 #endif /* __ASM_POWERPC_MMU_CONTEXT_H */
diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 6e31d87fb669..fa2bf69be182 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -156,10 +156,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __S390_MMU_CONTEXT_H */
diff --git a/arch/um/include/asm/mmu_context.h b/arch/um/include/asm/mmu_context.h
index 94ac2739918c..b668e351fd6c 100644
--- a/arch/um/include/asm/mmu_context.h
+++ b/arch/um/include/asm/mmu_context.h
@@ -37,12 +37,6 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return true;
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
-
 /*
  * end asm-generic/mm_hooks.h functions
  */
diff --git a/arch/unicore32/include/asm/mmu_context.h b/arch/unicore32/include/asm/mmu_context.h
index 62dfc644c908..59b06b48f27d 100644
--- a/arch/unicore32/include/asm/mmu_context.h
+++ b/arch/unicore32/include/asm/mmu_context.h
@@ -103,10 +103,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 306c7e12af55..68b329d77b3a 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -268,8 +268,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return __pkru_allows_pkey(vma_pkey(vma), write);
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	return __pkru_allows_pkey(pte_flags_pkey(pte_flags(pte)), write);
-}
 #endif /* _ASM_X86_MMU_CONTEXT_H */
diff --git a/include/asm-generic/mm_hooks.h b/include/asm-generic/mm_hooks.h
index cc5d9a1405df..41e5b6784b97 100644
--- a/include/asm-generic/mm_hooks.h
+++ b/include/asm-generic/mm_hooks.h
@@ -32,10 +32,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif	/* _ASM_GENERIC_MM_HOOKS_H */
diff --git a/mm/gup.c b/mm/gup.c
index 04aa405350dc..3f2338ba3402 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1216,9 +1216,6 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			pte_protnone(pte) || (write && !pte_write(pte)))
 			goto pte_unmap;
 
-		if (!arch_pte_access_permitted(pte, write))
-			goto pte_unmap;
-
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
 		page = pte_page(pte);
 		head = compound_head(page);
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 2/7] mm/gup: Move permission checks into helpers

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86, we would need to do additional permission checks to determinate if
access is allowed.

Let's abstract it out into separate helpers.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 include/asm-generic/pgtable.h | 25 +++++++++++++++++++++++++
 mm/gup.c                      | 15 ++++++++++-----
 2 files changed, 35 insertions(+), 5 deletions(-)

diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
index 1fad160f35de..7dfa767dc680 100644
--- a/include/asm-generic/pgtable.h
+++ b/include/asm-generic/pgtable.h
@@ -341,6 +341,31 @@ static inline int pte_unused(pte_t pte)
 }
 #endif
 
+#ifndef pte_access_permitted
+#define pte_access_permitted(pte, write) \
+	(pte_present(pte) && (!(write) || pte_write(pte)))
+#endif
+
+#ifndef pmd_access_permitted
+#define pmd_access_permitted(pmd, write) \
+	(pmd_present(pmd) && (!(write) || pmd_write(pmd)))
+#endif
+
+#ifndef pud_access_permitted
+#define pud_access_permitted(pud, write) \
+	(pud_present(pud) && (!(write) || pud_write(pud)))
+#endif
+
+#ifndef p4d_access_permitted
+#define p4d_access_permitted(p4d, write) \
+	(p4d_present(p4d) && (!(write) || p4d_write(p4d)))
+#endif
+
+#ifndef pgd_access_permitted
+#define pgd_access_permitted(pgd, write) \
+	(pgd_present(pgd) && (!(write) || pgd_write(pgd)))
+#endif
+
 #ifndef __HAVE_ARCH_PMD_SAME
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
diff --git a/mm/gup.c b/mm/gup.c
index 3f2338ba3402..a62a778ce4ec 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1212,8 +1212,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		 * Similar to the PMD case below, NUMA hinting must take slow
 		 * path using the pte_protnone check.
 		 */
-		if (!pte_present(pte) || pte_special(pte) ||
-			pte_protnone(pte) || (write && !pte_write(pte)))
+		if (pte_protnone(pte))
+			goto pte_unmap;
+
+		if (!pte_access_permitted(pte, write))
+			goto pte_unmap;
+
+		if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1264,7 +1269,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pmd_write(orig))
+	if (!pmd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1299,7 +1304,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pud_write(orig))
+	if (!pud_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1335,7 +1340,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	int refs;
 	struct page *head, *page;
 
-	if (write && !pgd_write(orig))
+	if (!pgd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
-- 
2.11.0

[PATCH 2/7] mm/gup: Move permission checks into helpers

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86, we would need to do additional permission checks to determinate if
access is allowed.

Let's abstract it out into separate helpers.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 include/asm-generic/pgtable.h | 25 +++++++++++++++++++++++++
 mm/gup.c                      | 15 ++++++++++-----
 2 files changed, 35 insertions(+), 5 deletions(-)

diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
index 1fad160f35de..7dfa767dc680 100644
--- a/include/asm-generic/pgtable.h
+++ b/include/asm-generic/pgtable.h
@@ -341,6 +341,31 @@ static inline int pte_unused(pte_t pte)
 }
 #endif
 
+#ifndef pte_access_permitted
+#define pte_access_permitted(pte, write) \
+	(pte_present(pte) && (!(write) || pte_write(pte)))
+#endif
+
+#ifndef pmd_access_permitted
+#define pmd_access_permitted(pmd, write) \
+	(pmd_present(pmd) && (!(write) || pmd_write(pmd)))
+#endif
+
+#ifndef pud_access_permitted
+#define pud_access_permitted(pud, write) \
+	(pud_present(pud) && (!(write) || pud_write(pud)))
+#endif
+
+#ifndef p4d_access_permitted
+#define p4d_access_permitted(p4d, write) \
+	(p4d_present(p4d) && (!(write) || p4d_write(p4d)))
+#endif
+
+#ifndef pgd_access_permitted
+#define pgd_access_permitted(pgd, write) \
+	(pgd_present(pgd) && (!(write) || pgd_write(pgd)))
+#endif
+
 #ifndef __HAVE_ARCH_PMD_SAME
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
diff --git a/mm/gup.c b/mm/gup.c
index 3f2338ba3402..a62a778ce4ec 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1212,8 +1212,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		 * Similar to the PMD case below, NUMA hinting must take slow
 		 * path using the pte_protnone check.
 		 */
-		if (!pte_present(pte) || pte_special(pte) ||
-			pte_protnone(pte) || (write && !pte_write(pte)))
+		if (pte_protnone(pte))
+			goto pte_unmap;
+
+		if (!pte_access_permitted(pte, write))
+			goto pte_unmap;
+
+		if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1264,7 +1269,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pmd_write(orig))
+	if (!pmd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1299,7 +1304,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pud_write(orig))
+	if (!pud_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1335,7 +1340,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	int refs;
 	struct page *head, *page;
 
-	if (write && !pgd_write(orig))
+	if (!pgd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 3/7] mm/gup: Move page table entry dereference into helper

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86 PAE, page table entry is larger than sizeof(long) and we would
need to provide helper that can read the entry atomically.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/gup.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index a62a778ce4ec..ed2259dc4606 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1189,6 +1189,17 @@ struct page *get_dump_page(unsigned long addr)
  */
 #ifdef CONFIG_HAVE_GENERIC_RCU_GUP
 
+#ifndef gup_get_pte
+/*
+ * We assume that the pte can be read atomically. If this is not the case for
+ * your architecture, please provide the helper.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	return READ_ONCE(*ptep);
+}
+#endif
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
@@ -1198,14 +1209,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
-		/*
-		 * In the line below we are assuming that the pte can be read
-		 * atomically. If this is not the case for your architecture,
-		 * please wrap this in a helper function!
-		 *
-		 * for an example see gup_get_pte in arch/x86/mm/gup.c
-		 */
-		pte_t pte = READ_ONCE(*ptep);
+		pte_t pte = gup_get_pte(ptep);
 		struct page *head, *page;
 
 		/*
-- 
2.11.0

[PATCH 3/7] mm/gup: Move page table entry dereference into helper

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86 PAE, page table entry is larger than sizeof(long) and we would
need to provide helper that can read the entry atomically.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/gup.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index a62a778ce4ec..ed2259dc4606 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1189,6 +1189,17 @@ struct page *get_dump_page(unsigned long addr)
  */
 #ifdef CONFIG_HAVE_GENERIC_RCU_GUP
 
+#ifndef gup_get_pte
+/*
+ * We assume that the pte can be read atomically. If this is not the case for
+ * your architecture, please provide the helper.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	return READ_ONCE(*ptep);
+}
+#endif
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
@@ -1198,14 +1209,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
-		/*
-		 * In the line below we are assuming that the pte can be read
-		 * atomically. If this is not the case for your architecture,
-		 * please wrap this in a helper function!
-		 *
-		 * for an example see gup_get_pte in arch/x86/mm/gup.c
-		 */
-		pte_t pte = READ_ONCE(*ptep);
+		pte_t pte = gup_get_pte(ptep);
 		struct page *head, *page;
 
 		/*
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 4/7] mm/gup: Make pages referenced during generic get_user_pages_fast()

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

Unlike generic GUP_fast(), x86 version makes all pages it touches
referenced. It seems required for GRU and EPT.

See commit 8ee53820edfd ("thp: mmu_notifier_test_young").

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
---
 mm/gup.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/mm/gup.c b/mm/gup.c
index ed2259dc4606..f0cfe225c08f 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1238,6 +1238,8 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		}
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
+
+		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
 
@@ -1299,6 +1301,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1334,6 +1337,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1370,6 +1374,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
-- 
2.11.0

[PATCH 4/7] mm/gup: Make pages referenced during generic get_user_pages_fast()

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

Unlike generic GUP_fast(), x86 version makes all pages it touches
referenced. It seems required for GRU and EPT.

See commit 8ee53820edfd ("thp: mmu_notifier_test_young").

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
---
 mm/gup.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/mm/gup.c b/mm/gup.c
index ed2259dc4606..f0cfe225c08f 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1238,6 +1238,8 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		}
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
+
+		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
 
@@ -1299,6 +1301,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1334,6 +1337,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1370,6 +1374,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 5/7] mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

Prepare generic GUP_fast() to handle dev_mapmap. At the moment, it's
only implemented on x86. On non-x86, the new code will be compiled out.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
---
 include/linux/mm.h |  4 +++
 mm/gup.c           | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 75 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 5f01c88f0800..e197d3ca3e8a 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -430,6 +430,10 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 
 /*
diff --git a/mm/gup.c b/mm/gup.c
index f0cfe225c08f..e04f76594eb9 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1200,12 +1200,23 @@ static inline pte_t gup_get_pte(pte_t *ptep)
 }
 #endif
 
+static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
+{
+	while ((*nr) - nr_start) {
+		struct page *page = pages[--(*nr)];
+
+		ClearPageReferenced(page);
+		put_page(page);
+	}
+}
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
 {
+	struct dev_pagemap *pgmap = NULL;
+	int nr_start = *nr, ret = 0;
 	pte_t *ptep, *ptem;
-	int ret = 0;
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
@@ -1222,7 +1233,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		if (!pte_access_permitted(pte, write))
 			goto pte_unmap;
 
-		if (pte_special(pte))
+		if (pte_devmap(pte)) {
+			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
+			if (unlikely(!pgmap)) {
+				undo_dev_pagemap(nr, nr_start, pages);
+				goto pte_unmap;
+			}
+		} else if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1239,6 +1256,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
 
+		put_dev_pagemap(pgmap);
 		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
@@ -1269,6 +1287,48 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 }
 #endif /* __HAVE_ARCH_PTE_SPECIAL */
 
+static int __gup_device_huge(unsigned long pfn, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	int nr_start = *nr;
+	struct dev_pagemap *pgmap = NULL;
+
+	do {
+		struct page *page = pfn_to_page(pfn);
+
+		pgmap = get_dev_pagemap(pfn, pgmap);
+		if (unlikely(!pgmap)) {
+			undo_dev_pagemap(nr, nr_start, pages);
+			return 0;
+		}
+		SetPageReferenced(page);
+		pages[*nr] = page;
+		get_page(page);
+		put_dev_pagemap(pgmap);
+		(*nr)++;
+		pfn++;
+	} while (addr += PAGE_SIZE, addr != end);
+	return 1;
+}
+
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+
 static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, int write, struct page **pages, int *nr)
 {
@@ -1278,6 +1338,10 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	if (!pmd_access_permitted(orig, write))
 		return 0;
 
+	VM_BUG_ON(!pfn_valid(pmd_pfn(orig)));
+	if (pmd_devmap(orig))
+		return __gup_device_huge_pmd(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pmd_page(orig);
 	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
@@ -1314,6 +1378,10 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	if (!pud_access_permitted(orig, write))
 		return 0;
 
+	VM_BUG_ON(!pfn_valid(pud_pfn(orig)));
+	if (pud_devmap(orig))
+		return __gup_device_huge_pud(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pud_page(orig);
 	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
@@ -1351,6 +1419,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	if (!pgd_access_permitted(orig, write))
 		return 0;
 
+	BUILD_BUG_ON(pgd_devmap(orig));
 	refs = 0;
 	head = pgd_page(orig);
 	page = head + ((addr & ~PGDIR_MASK) >> PAGE_SHIFT);
-- 
2.11.0

[PATCH 5/7] mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

Prepare generic GUP_fast() to handle dev_mapmap. At the moment, it's
only implemented on x86. On non-x86, the new code will be compiled out.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
---
 include/linux/mm.h |  4 +++
 mm/gup.c           | 73 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 75 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 5f01c88f0800..e197d3ca3e8a 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -430,6 +430,10 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 
 /*
diff --git a/mm/gup.c b/mm/gup.c
index f0cfe225c08f..e04f76594eb9 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1200,12 +1200,23 @@ static inline pte_t gup_get_pte(pte_t *ptep)
 }
 #endif
 
+static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
+{
+	while ((*nr) - nr_start) {
+		struct page *page = pages[--(*nr)];
+
+		ClearPageReferenced(page);
+		put_page(page);
+	}
+}
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
 {
+	struct dev_pagemap *pgmap = NULL;
+	int nr_start = *nr, ret = 0;
 	pte_t *ptep, *ptem;
-	int ret = 0;
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
@@ -1222,7 +1233,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		if (!pte_access_permitted(pte, write))
 			goto pte_unmap;
 
-		if (pte_special(pte))
+		if (pte_devmap(pte)) {
+			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
+			if (unlikely(!pgmap)) {
+				undo_dev_pagemap(nr, nr_start, pages);
+				goto pte_unmap;
+			}
+		} else if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1239,6 +1256,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
 
+		put_dev_pagemap(pgmap);
 		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
@@ -1269,6 +1287,48 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 }
 #endif /* __HAVE_ARCH_PTE_SPECIAL */
 
+static int __gup_device_huge(unsigned long pfn, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	int nr_start = *nr;
+	struct dev_pagemap *pgmap = NULL;
+
+	do {
+		struct page *page = pfn_to_page(pfn);
+
+		pgmap = get_dev_pagemap(pfn, pgmap);
+		if (unlikely(!pgmap)) {
+			undo_dev_pagemap(nr, nr_start, pages);
+			return 0;
+		}
+		SetPageReferenced(page);
+		pages[*nr] = page;
+		get_page(page);
+		put_dev_pagemap(pgmap);
+		(*nr)++;
+		pfn++;
+	} while (addr += PAGE_SIZE, addr != end);
+	return 1;
+}
+
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+
 static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, int write, struct page **pages, int *nr)
 {
@@ -1278,6 +1338,10 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	if (!pmd_access_permitted(orig, write))
 		return 0;
 
+	VM_BUG_ON(!pfn_valid(pmd_pfn(orig)));
+	if (pmd_devmap(orig))
+		return __gup_device_huge_pmd(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pmd_page(orig);
 	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
@@ -1314,6 +1378,10 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	if (!pud_access_permitted(orig, write))
 		return 0;
 
+	VM_BUG_ON(!pfn_valid(pud_pfn(orig)));
+	if (pud_devmap(orig))
+		return __gup_device_huge_pud(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pud_page(orig);
 	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
@@ -1351,6 +1419,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	if (!pgd_access_permitted(orig, write))
 		return 0;
 
+	BUILD_BUG_ON(pgd_devmap(orig));
 	refs = 0;
 	head = pgd_page(orig);
 	page = head + ((addr & ~PGDIR_MASK) >> PAGE_SHIFT);
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 6/7] mm/gup: Provide hook to check if __GUP_fast() is allowed for the range

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86, get_user_pages_fast() does few sanity checks to see if we can
call __get_user_pages_fast() for the range.

Wrapping protection should be useful for generic code too.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/gup.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index e04f76594eb9..a6e21380c8b0 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1600,6 +1600,21 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return nr;
 }
 
+#ifndef gup_fast_permitted
+/*
+ * Check if it's allowed to use __get_user_pages_fast() for the range or
+ * need to fallback to non-fast version.
+ */
+bool gup_fast_permitted(unsigned long start, int nr_pages, int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long) nr_pages << PAGE_SHIFT;
+	end = start + len;
+	return end >= start;
+}
+#endif
+
 /**
  * get_user_pages_fast() - pin user pages in memory
  * @start:	starting user address
@@ -1619,11 +1634,14 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 			struct page **pages)
 {
-	int nr, ret;
+	int nr = 0, ret = 0;
 
 	start &= PAGE_MASK;
-	nr = __get_user_pages_fast(start, nr_pages, write, pages);
-	ret = nr;
+
+	if (gup_fast_permitted(start, nr_pages, write)) {
+		nr = __get_user_pages_fast(start, nr_pages, write, pages);
+		ret = nr;
+	}
 
 	if (nr < nr_pages) {
 		/* Try to get the remaining pages with get_user_pages */
-- 
2.11.0

[PATCH 6/7] mm/gup: Provide hook to check if __GUP_fast() is allowed for the range

[Kirill A. Shutemov]

This is preparation patch for transition of x86 to generic GUP_fast()
implementation.

On x86, get_user_pages_fast() does few sanity checks to see if we can
call __get_user_pages_fast() for the range.

Wrapping protection should be useful for generic code too.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/gup.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index e04f76594eb9..a6e21380c8b0 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1600,6 +1600,21 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return nr;
 }
 
+#ifndef gup_fast_permitted
+/*
+ * Check if it's allowed to use __get_user_pages_fast() for the range or
+ * need to fallback to non-fast version.
+ */
+bool gup_fast_permitted(unsigned long start, int nr_pages, int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long) nr_pages << PAGE_SHIFT;
+	end = start + len;
+	return end >= start;
+}
+#endif
+
 /**
  * get_user_pages_fast() - pin user pages in memory
  * @start:	starting user address
@@ -1619,11 +1634,14 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 			struct page **pages)
 {
-	int nr, ret;
+	int nr = 0, ret = 0;
 
 	start &= PAGE_MASK;
-	nr = __get_user_pages_fast(start, nr_pages, write, pages);
-	ret = nr;
+
+	if (gup_fast_permitted(start, nr_pages, write)) {
+		nr = __get_user_pages_fast(start, nr_pages, write, pages);
+		ret = nr;
+	}
 
 	if (nr < nr_pages) {
 		/* Try to get the remaining pages with get_user_pages */
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

The patch provides all required hooks to match generic
get_user_pages_fast() behaviour to x86 and switch x86 over.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 arch/x86/Kconfig                      |   3 +
 arch/x86/include/asm/mmu_context.h    |  12 -
 arch/x86/include/asm/pgtable-3level.h |  45 +++
 arch/x86/include/asm/pgtable.h        |  53 ++++
 arch/x86/include/asm/pgtable_64.h     |  16 +-
 arch/x86/mm/Makefile                  |   2 +-
 arch/x86/mm/gup.c                     | 496 ----------------------------------
 7 files changed, 117 insertions(+), 510 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2bab9d093b51..f9d859180803 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2788,6 +2788,9 @@ config X86_DMA_REMAP
 	bool
 	depends on STA2X11
 
+config HAVE_GENERIC_RCU_GUP
+	def_bool y
+
 source "net/Kconfig"
 
 source "drivers/Kconfig"
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 68b329d77b3a..6e933d2d88d9 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -220,18 +220,6 @@ static inline int vma_pkey(struct vm_area_struct *vma)
 }
 #endif
 
-static inline bool __pkru_allows_pkey(u16 pkey, bool write)
-{
-	u32 pkru = read_pkru();
-
-	if (!__pkru_allows_read(pkru, pkey))
-		return false;
-	if (write && !__pkru_allows_write(pkru, pkey))
-		return false;
-
-	return true;
-}
-
 /*
  * We only want to enforce protection keys on the current process
  * because we effectively have no access to PKRU for other
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index 72277b1028a5..698060a0dfb6 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -215,4 +215,49 @@ static inline pud_t native_pudp_get_and_clear(pud_t *pudp)
 #define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t){ { .pte_high = (x).val } })
 
+#define gup_get_pte gup_get_pte
+/*
+ * WARNING: only to be used in get_user_pages_fast() implementation.
+ *
+ * With get_user_pages_fast, we walk down the pagetables without taking
+ * any locks.  For this we would like to load the pointers atomically,
+ * but that is not possible (without expensive cmpxchg8b) on PAE.  What
+ * we do have is the guarantee that a pte will only either go from not
+ * present to present, or present to not present or both -- it will not
+ * switch to a completely different present page without a TLB flush in
+ * between; something that we are blocking by holding interrupts off.
+ *
+ * Setting ptes from not present to present goes:
+ * ptep->pte_high = h;
+ * smp_wmb();
+ * ptep->pte_low = l;
+ *
+ * And present to not present goes:
+ * ptep->pte_low = 0;
+ * smp_wmb();
+ * ptep->pte_high = 0;
+ *
+ * We must ensure here that the load of pte_low sees l iff pte_high
+ * sees h. We load pte_high *after* loading pte_low, which ensures we
+ * don't see an older value of pte_high.  *Then* we recheck pte_low,
+ * which ensures that we haven't picked up a changed pte high. We might
+ * have got rubbish values from pte_low and pte_high, but we are
+ * guaranteed that pte_low will not have the present bit set *unless*
+ * it is 'l'. And get_user_pages_fast only operates on present ptes, so
+ * we're safe.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	pte_t pte;
+
+	do {
+		pte.pte_low = ptep->pte_low;
+		smp_rmb();
+		pte.pte_high = ptep->pte_high;
+		smp_rmb();
+	} while (unlikely(pte.pte_low != ptep->pte_low));
+
+	return pte;
+}
+
 #endif /* _ASM_X86_PGTABLE_3LEVEL_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 83f17218e66a..96545878f180 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -232,6 +232,11 @@ static inline int pud_devmap(pud_t pud)
 }
 #endif
 #endif
+
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 static inline pte_t pte_set_flags(pte_t pte, pteval_t set)
@@ -1135,6 +1140,54 @@ static inline u16 pte_flags_pkey(unsigned long pte_flags)
 #endif
 }
 
+static inline bool __pkru_allows_pkey(u16 pkey, bool write)
+{
+	u32 pkru = read_pkru();
+
+	if (!__pkru_allows_read(pkru, pkey))
+		return false;
+	if (write && !__pkru_allows_write(pkru, pkey))
+		return false;
+
+	return true;
+}
+
+/*
+ * 'pteval' can come from a pte, pmd or pud.  We only check
+ * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
+ * same value on all 3 types.
+ */
+static inline bool __pte_access_permitted(unsigned long pteval, bool write)
+{
+	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
+
+	if (write)
+		need_pte_bits |= _PAGE_RW;
+
+	if ((pteval & need_pte_bits) != need_pte_bits)
+		return 0;
+
+	return __pkru_allows_pkey(pte_flags_pkey(pteval), write);
+}
+
+#define pte_access_permitted pte_access_permitted
+static inline bool pte_access_permitted(pte_t pte, bool write)
+{
+	return __pte_access_permitted(pte_val(pte), write);
+}
+
+#define pmd_access_permitted pmd_access_permitted
+static inline bool pmd_access_permitted(pmd_t pmd, bool write)
+{
+	return __pte_access_permitted(pmd_val(pmd), write);
+}
+
+#define pud_access_permitted pud_access_permitted
+static inline bool pud_access_permitted(pud_t pud, bool write)
+{
+	return __pte_access_permitted(pud_val(pud), write);
+}
+
 #include <asm-generic/pgtable.h>
 #endif	/* __ASSEMBLY__ */
 
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 73c7ccc38912..1a4bc71534d4 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -206,6 +206,20 @@ extern void cleanup_highmap(void);
 extern void init_extra_mapping_uc(unsigned long phys, unsigned long size);
 extern void init_extra_mapping_wb(unsigned long phys, unsigned long size);
 
-#endif /* !__ASSEMBLY__ */
+#define gup_fast_permitted gup_fast_permitted
+static inline bool gup_fast_permitted(unsigned long start, int nr_pages,
+		int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long)nr_pages << PAGE_SHIFT;
+	end = start + len;
+	if (end < start)
+		return false;
+	if (end >> __VIRTUAL_MASK_SHIFT)
+		return false;
+	return true;
+}
 
+#endif /* !__ASSEMBLY__ */
 #endif /* _ASM_X86_PGTABLE_64_H */
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 96d2b847e09e..0fbdcb64f9f8 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -2,7 +2,7 @@
 KCOV_INSTRUMENT_tlb.o	:= n
 
 obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o gup.o setup_nx.o tlb.o
+	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
deleted file mode 100644
index 456dfdfd2249..000000000000
--- a/arch/x86/mm/gup.c
+++ /dev/null
@@ -1,496 +0,0 @@
-/*
- * Lockless get_user_pages_fast for x86
- *
- * Copyright (C) 2008 Nick Piggin
- * Copyright (C) 2008 Novell Inc.
- */
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmstat.h>
-#include <linux/highmem.h>
-#include <linux/swap.h>
-#include <linux/memremap.h>
-
-#include <asm/mmu_context.h>
-#include <asm/pgtable.h>
-
-static inline pte_t gup_get_pte(pte_t *ptep)
-{
-#ifndef CONFIG_X86_PAE
-	return READ_ONCE(*ptep);
-#else
-	/*
-	 * With get_user_pages_fast, we walk down the pagetables without taking
-	 * any locks.  For this we would like to load the pointers atomically,
-	 * but that is not possible (without expensive cmpxchg8b) on PAE.  What
-	 * we do have is the guarantee that a pte will only either go from not
-	 * present to present, or present to not present or both -- it will not
-	 * switch to a completely different present page without a TLB flush in
-	 * between; something that we are blocking by holding interrupts off.
-	 *
-	 * Setting ptes from not present to present goes:
-	 * ptep->pte_high = h;
-	 * smp_wmb();
-	 * ptep->pte_low = l;
-	 *
-	 * And present to not present goes:
-	 * ptep->pte_low = 0;
-	 * smp_wmb();
-	 * ptep->pte_high = 0;
-	 *
-	 * We must ensure here that the load of pte_low sees l iff pte_high
-	 * sees h. We load pte_high *after* loading pte_low, which ensures we
-	 * don't see an older value of pte_high.  *Then* we recheck pte_low,
-	 * which ensures that we haven't picked up a changed pte high. We might
-	 * have got rubbish values from pte_low and pte_high, but we are
-	 * guaranteed that pte_low will not have the present bit set *unless*
-	 * it is 'l'. And get_user_pages_fast only operates on present ptes, so
-	 * we're safe.
-	 *
-	 * gup_get_pte should not be used or copied outside gup.c without being
-	 * very careful -- it does not atomically load the pte or anything that
-	 * is likely to be useful for you.
-	 */
-	pte_t pte;
-
-retry:
-	pte.pte_low = ptep->pte_low;
-	smp_rmb();
-	pte.pte_high = ptep->pte_high;
-	smp_rmb();
-	if (unlikely(pte.pte_low != ptep->pte_low))
-		goto retry;
-
-	return pte;
-#endif
-}
-
-static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
-{
-	while ((*nr) - nr_start) {
-		struct page *page = pages[--(*nr)];
-
-		ClearPageReferenced(page);
-		put_page(page);
-	}
-}
-
-/*
- * 'pteval' can come from a pte, pmd, pud or p4d.  We only check
- * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
- * same value on all 4 types.
- */
-static inline int pte_allows_gup(unsigned long pteval, int write)
-{
-	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
-
-	if (write)
-		need_pte_bits |= _PAGE_RW;
-
-	if ((pteval & need_pte_bits) != need_pte_bits)
-		return 0;
-
-	/* Check memory protection keys permissions. */
-	if (!__pkru_allows_pkey(pte_flags_pkey(pteval), write))
-		return 0;
-
-	return 1;
-}
-
-/*
- * The performance critical leaf functions are made noinline otherwise gcc
- * inlines everything into a single function which results in too much
- * register pressure.
- */
-static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct dev_pagemap *pgmap = NULL;
-	int nr_start = *nr, ret = 0;
-	pte_t *ptep, *ptem;
-
-	/*
-	 * Keep the original mapped PTE value (ptem) around since we
-	 * might increment ptep off the end of the page when finishing
-	 * our loop iteration.
-	 */
-	ptem = ptep = pte_offset_map(&pmd, addr);
-	do {
-		pte_t pte = gup_get_pte(ptep);
-		struct page *page;
-
-		/* Similar to the PMD case, NUMA hinting must take slow path */
-		if (pte_protnone(pte))
-			break;
-
-		if (!pte_allows_gup(pte_val(pte), write))
-			break;
-
-		if (pte_devmap(pte)) {
-			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
-			if (unlikely(!pgmap)) {
-				undo_dev_pagemap(nr, nr_start, pages);
-				break;
-			}
-		} else if (pte_special(pte))
-			break;
-
-		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
-		page = pte_page(pte);
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		(*nr)++;
-
-	} while (ptep++, addr += PAGE_SIZE, addr != end);
-	if (addr == end)
-		ret = 1;
-	pte_unmap(ptem);
-
-	return ret;
-}
-
-static inline void get_head_page_multiple(struct page *page, int nr)
-{
-	VM_BUG_ON_PAGE(page != compound_head(page), page);
-	VM_BUG_ON_PAGE(page_count(page) == 0, page);
-	page_ref_add(page, nr);
-	SetPageReferenced(page);
-}
-
-static int __gup_device_huge(unsigned long pfn, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	int nr_start = *nr;
-	struct dev_pagemap *pgmap = NULL;
-
-	do {
-		struct page *page = pfn_to_page(pfn);
-
-		pgmap = get_dev_pagemap(pfn, pgmap);
-		if (unlikely(!pgmap)) {
-			undo_dev_pagemap(nr, nr_start, pages);
-			return 0;
-		}
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		(*nr)++;
-		pfn++;
-	} while (addr += PAGE_SIZE, addr != end);
-	return 1;
-}
-
-static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static noinline int gup_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pmd_val(pmd), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pmd_pfn(pmd)));
-	if (pmd_devmap(pmd))
-		return __gup_device_huge_pmd(pmd, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pmd_flags(pmd) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pmd_page(pmd);
-	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end,
-		int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pmd_t *pmdp;
-
-	pmdp = pmd_offset(&pud, addr);
-	do {
-		pmd_t pmd = *pmdp;
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			return 0;
-		if (unlikely(pmd_large(pmd) || !pmd_present(pmd))) {
-			/*
-			 * NUMA hinting faults need to be handled in the GUP
-			 * slowpath for accounting purposes and so that they
-			 * can be serialised against THP migration.
-			 */
-			if (pmd_protnone(pmd))
-				return 0;
-			if (!gup_huge_pmd(pmd, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pte_range(pmd, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pmdp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static noinline int gup_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pud_val(pud), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pud_pfn(pud)));
-	if (pud_devmap(pud))
-		return __gup_device_huge_pud(pud, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pud_flags(pud) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pud_page(pud);
-	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pud_range(p4d_t p4d, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pud_t *pudp;
-
-	pudp = pud_offset(&p4d, addr);
-	do {
-		pud_t pud = *pudp;
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			return 0;
-		if (unlikely(pud_large(pud))) {
-			if (!gup_huge_pud(pud, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pmd_range(pud, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pudp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static int gup_p4d_range(pgd_t pgd, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	p4d_t *p4dp;
-
-	p4dp = p4d_offset(&pgd, addr);
-	do {
-		p4d_t p4d = *p4dp;
-
-		next = p4d_addr_end(addr, end);
-		if (p4d_none(p4d))
-			return 0;
-		BUILD_BUG_ON(p4d_large(p4d));
-		if (!gup_pud_range(p4d, addr, next, write, pages, nr))
-			return 0;
-	} while (p4dp++, addr = next, addr != end);
-
-	return 1;
-}
-
-/*
- * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
- * back to the regular GUP.
- */
-int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			  struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	unsigned long flags;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
-					(void __user *)start, len)))
-		return 0;
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_save(flags);
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			break;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			break;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_restore(flags);
-
-	return nr;
-}
-
-/**
- * get_user_pages_fast() - pin user pages in memory
- * @start:	starting user address
- * @nr_pages:	number of pages from start to pin
- * @write:	whether pages will be written to
- * @pages:	array that receives pointers to the pages pinned.
- * 		Should be at least nr_pages long.
- *
- * Attempt to pin user pages in memory without taking mm->mmap_sem.
- * If not successful, it will fall back to taking the lock and
- * calling get_user_pages().
- *
- * Returns number of pages pinned. This may be fewer than the number
- * requested. If nr_pages is 0 or negative, returns 0. If no pages
- * were pinned, returns -errno.
- */
-int get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-
-	end = start + len;
-	if (end < start)
-		goto slow_irqon;
-
-#ifdef CONFIG_X86_64
-	if (end >> __VIRTUAL_MASK_SHIFT)
-		goto slow_irqon;
-#endif
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		ret = get_user_pages_unlocked(start,
-					      (end - start) >> PAGE_SHIFT,
-					      pages, write ? FOLL_WRITE : 0);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
-}
-- 
2.11.0

[PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

The patch provides all required hooks to match generic
get_user_pages_fast() behaviour to x86 and switch x86 over.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 arch/x86/Kconfig                      |   3 +
 arch/x86/include/asm/mmu_context.h    |  12 -
 arch/x86/include/asm/pgtable-3level.h |  45 +++
 arch/x86/include/asm/pgtable.h        |  53 ++++
 arch/x86/include/asm/pgtable_64.h     |  16 +-
 arch/x86/mm/Makefile                  |   2 +-
 arch/x86/mm/gup.c                     | 496 ----------------------------------
 7 files changed, 117 insertions(+), 510 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2bab9d093b51..f9d859180803 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2788,6 +2788,9 @@ config X86_DMA_REMAP
 	bool
 	depends on STA2X11
 
+config HAVE_GENERIC_RCU_GUP
+	def_bool y
+
 source "net/Kconfig"
 
 source "drivers/Kconfig"
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 68b329d77b3a..6e933d2d88d9 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -220,18 +220,6 @@ static inline int vma_pkey(struct vm_area_struct *vma)
 }
 #endif
 
-static inline bool __pkru_allows_pkey(u16 pkey, bool write)
-{
-	u32 pkru = read_pkru();
-
-	if (!__pkru_allows_read(pkru, pkey))
-		return false;
-	if (write && !__pkru_allows_write(pkru, pkey))
-		return false;
-
-	return true;
-}
-
 /*
  * We only want to enforce protection keys on the current process
  * because we effectively have no access to PKRU for other
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index 72277b1028a5..698060a0dfb6 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -215,4 +215,49 @@ static inline pud_t native_pudp_get_and_clear(pud_t *pudp)
 #define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t){ { .pte_high = (x).val } })
 
+#define gup_get_pte gup_get_pte
+/*
+ * WARNING: only to be used in get_user_pages_fast() implementation.
+ *
+ * With get_user_pages_fast, we walk down the pagetables without taking
+ * any locks.  For this we would like to load the pointers atomically,
+ * but that is not possible (without expensive cmpxchg8b) on PAE.  What
+ * we do have is the guarantee that a pte will only either go from not
+ * present to present, or present to not present or both -- it will not
+ * switch to a completely different present page without a TLB flush in
+ * between; something that we are blocking by holding interrupts off.
+ *
+ * Setting ptes from not present to present goes:
+ * ptep->pte_high = h;
+ * smp_wmb();
+ * ptep->pte_low = l;
+ *
+ * And present to not present goes:
+ * ptep->pte_low = 0;
+ * smp_wmb();
+ * ptep->pte_high = 0;
+ *
+ * We must ensure here that the load of pte_low sees l iff pte_high
+ * sees h. We load pte_high *after* loading pte_low, which ensures we
+ * don't see an older value of pte_high.  *Then* we recheck pte_low,
+ * which ensures that we haven't picked up a changed pte high. We might
+ * have got rubbish values from pte_low and pte_high, but we are
+ * guaranteed that pte_low will not have the present bit set *unless*
+ * it is 'l'. And get_user_pages_fast only operates on present ptes, so
+ * we're safe.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	pte_t pte;
+
+	do {
+		pte.pte_low = ptep->pte_low;
+		smp_rmb();
+		pte.pte_high = ptep->pte_high;
+		smp_rmb();
+	} while (unlikely(pte.pte_low != ptep->pte_low));
+
+	return pte;
+}
+
 #endif /* _ASM_X86_PGTABLE_3LEVEL_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 83f17218e66a..96545878f180 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -232,6 +232,11 @@ static inline int pud_devmap(pud_t pud)
 }
 #endif
 #endif
+
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 static inline pte_t pte_set_flags(pte_t pte, pteval_t set)
@@ -1135,6 +1140,54 @@ static inline u16 pte_flags_pkey(unsigned long pte_flags)
 #endif
 }
 
+static inline bool __pkru_allows_pkey(u16 pkey, bool write)
+{
+	u32 pkru = read_pkru();
+
+	if (!__pkru_allows_read(pkru, pkey))
+		return false;
+	if (write && !__pkru_allows_write(pkru, pkey))
+		return false;
+
+	return true;
+}
+
+/*
+ * 'pteval' can come from a pte, pmd or pud.  We only check
+ * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
+ * same value on all 3 types.
+ */
+static inline bool __pte_access_permitted(unsigned long pteval, bool write)
+{
+	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
+
+	if (write)
+		need_pte_bits |= _PAGE_RW;
+
+	if ((pteval & need_pte_bits) != need_pte_bits)
+		return 0;
+
+	return __pkru_allows_pkey(pte_flags_pkey(pteval), write);
+}
+
+#define pte_access_permitted pte_access_permitted
+static inline bool pte_access_permitted(pte_t pte, bool write)
+{
+	return __pte_access_permitted(pte_val(pte), write);
+}
+
+#define pmd_access_permitted pmd_access_permitted
+static inline bool pmd_access_permitted(pmd_t pmd, bool write)
+{
+	return __pte_access_permitted(pmd_val(pmd), write);
+}
+
+#define pud_access_permitted pud_access_permitted
+static inline bool pud_access_permitted(pud_t pud, bool write)
+{
+	return __pte_access_permitted(pud_val(pud), write);
+}
+
 #include <asm-generic/pgtable.h>
 #endif	/* __ASSEMBLY__ */
 
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 73c7ccc38912..1a4bc71534d4 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -206,6 +206,20 @@ extern void cleanup_highmap(void);
 extern void init_extra_mapping_uc(unsigned long phys, unsigned long size);
 extern void init_extra_mapping_wb(unsigned long phys, unsigned long size);
 
-#endif /* !__ASSEMBLY__ */
+#define gup_fast_permitted gup_fast_permitted
+static inline bool gup_fast_permitted(unsigned long start, int nr_pages,
+		int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long)nr_pages << PAGE_SHIFT;
+	end = start + len;
+	if (end < start)
+		return false;
+	if (end >> __VIRTUAL_MASK_SHIFT)
+		return false;
+	return true;
+}
 
+#endif /* !__ASSEMBLY__ */
 #endif /* _ASM_X86_PGTABLE_64_H */
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 96d2b847e09e..0fbdcb64f9f8 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -2,7 +2,7 @@
 KCOV_INSTRUMENT_tlb.o	:= n
 
 obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o gup.o setup_nx.o tlb.o
+	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
deleted file mode 100644
index 456dfdfd2249..000000000000
--- a/arch/x86/mm/gup.c
+++ /dev/null
@@ -1,496 +0,0 @@
-/*
- * Lockless get_user_pages_fast for x86
- *
- * Copyright (C) 2008 Nick Piggin
- * Copyright (C) 2008 Novell Inc.
- */
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmstat.h>
-#include <linux/highmem.h>
-#include <linux/swap.h>
-#include <linux/memremap.h>
-
-#include <asm/mmu_context.h>
-#include <asm/pgtable.h>
-
-static inline pte_t gup_get_pte(pte_t *ptep)
-{
-#ifndef CONFIG_X86_PAE
-	return READ_ONCE(*ptep);
-#else
-	/*
-	 * With get_user_pages_fast, we walk down the pagetables without taking
-	 * any locks.  For this we would like to load the pointers atomically,
-	 * but that is not possible (without expensive cmpxchg8b) on PAE.  What
-	 * we do have is the guarantee that a pte will only either go from not
-	 * present to present, or present to not present or both -- it will not
-	 * switch to a completely different present page without a TLB flush in
-	 * between; something that we are blocking by holding interrupts off.
-	 *
-	 * Setting ptes from not present to present goes:
-	 * ptep->pte_high = h;
-	 * smp_wmb();
-	 * ptep->pte_low = l;
-	 *
-	 * And present to not present goes:
-	 * ptep->pte_low = 0;
-	 * smp_wmb();
-	 * ptep->pte_high = 0;
-	 *
-	 * We must ensure here that the load of pte_low sees l iff pte_high
-	 * sees h. We load pte_high *after* loading pte_low, which ensures we
-	 * don't see an older value of pte_high.  *Then* we recheck pte_low,
-	 * which ensures that we haven't picked up a changed pte high. We might
-	 * have got rubbish values from pte_low and pte_high, but we are
-	 * guaranteed that pte_low will not have the present bit set *unless*
-	 * it is 'l'. And get_user_pages_fast only operates on present ptes, so
-	 * we're safe.
-	 *
-	 * gup_get_pte should not be used or copied outside gup.c without being
-	 * very careful -- it does not atomically load the pte or anything that
-	 * is likely to be useful for you.
-	 */
-	pte_t pte;
-
-retry:
-	pte.pte_low = ptep->pte_low;
-	smp_rmb();
-	pte.pte_high = ptep->pte_high;
-	smp_rmb();
-	if (unlikely(pte.pte_low != ptep->pte_low))
-		goto retry;
-
-	return pte;
-#endif
-}
-
-static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
-{
-	while ((*nr) - nr_start) {
-		struct page *page = pages[--(*nr)];
-
-		ClearPageReferenced(page);
-		put_page(page);
-	}
-}
-
-/*
- * 'pteval' can come from a pte, pmd, pud or p4d.  We only check
- * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
- * same value on all 4 types.
- */
-static inline int pte_allows_gup(unsigned long pteval, int write)
-{
-	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
-
-	if (write)
-		need_pte_bits |= _PAGE_RW;
-
-	if ((pteval & need_pte_bits) != need_pte_bits)
-		return 0;
-
-	/* Check memory protection keys permissions. */
-	if (!__pkru_allows_pkey(pte_flags_pkey(pteval), write))
-		return 0;
-
-	return 1;
-}
-
-/*
- * The performance critical leaf functions are made noinline otherwise gcc
- * inlines everything into a single function which results in too much
- * register pressure.
- */
-static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct dev_pagemap *pgmap = NULL;
-	int nr_start = *nr, ret = 0;
-	pte_t *ptep, *ptem;
-
-	/*
-	 * Keep the original mapped PTE value (ptem) around since we
-	 * might increment ptep off the end of the page when finishing
-	 * our loop iteration.
-	 */
-	ptem = ptep = pte_offset_map(&pmd, addr);
-	do {
-		pte_t pte = gup_get_pte(ptep);
-		struct page *page;
-
-		/* Similar to the PMD case, NUMA hinting must take slow path */
-		if (pte_protnone(pte))
-			break;
-
-		if (!pte_allows_gup(pte_val(pte), write))
-			break;
-
-		if (pte_devmap(pte)) {
-			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
-			if (unlikely(!pgmap)) {
-				undo_dev_pagemap(nr, nr_start, pages);
-				break;
-			}
-		} else if (pte_special(pte))
-			break;
-
-		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
-		page = pte_page(pte);
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		(*nr)++;
-
-	} while (ptep++, addr += PAGE_SIZE, addr != end);
-	if (addr == end)
-		ret = 1;
-	pte_unmap(ptem);
-
-	return ret;
-}
-
-static inline void get_head_page_multiple(struct page *page, int nr)
-{
-	VM_BUG_ON_PAGE(page != compound_head(page), page);
-	VM_BUG_ON_PAGE(page_count(page) == 0, page);
-	page_ref_add(page, nr);
-	SetPageReferenced(page);
-}
-
-static int __gup_device_huge(unsigned long pfn, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	int nr_start = *nr;
-	struct dev_pagemap *pgmap = NULL;
-
-	do {
-		struct page *page = pfn_to_page(pfn);
-
-		pgmap = get_dev_pagemap(pfn, pgmap);
-		if (unlikely(!pgmap)) {
-			undo_dev_pagemap(nr, nr_start, pages);
-			return 0;
-		}
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		(*nr)++;
-		pfn++;
-	} while (addr += PAGE_SIZE, addr != end);
-	return 1;
-}
-
-static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static noinline int gup_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pmd_val(pmd), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pmd_pfn(pmd)));
-	if (pmd_devmap(pmd))
-		return __gup_device_huge_pmd(pmd, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pmd_flags(pmd) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pmd_page(pmd);
-	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end,
-		int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pmd_t *pmdp;
-
-	pmdp = pmd_offset(&pud, addr);
-	do {
-		pmd_t pmd = *pmdp;
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			return 0;
-		if (unlikely(pmd_large(pmd) || !pmd_present(pmd))) {
-			/*
-			 * NUMA hinting faults need to be handled in the GUP
-			 * slowpath for accounting purposes and so that they
-			 * can be serialised against THP migration.
-			 */
-			if (pmd_protnone(pmd))
-				return 0;
-			if (!gup_huge_pmd(pmd, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pte_range(pmd, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pmdp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static noinline int gup_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pud_val(pud), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pud_pfn(pud)));
-	if (pud_devmap(pud))
-		return __gup_device_huge_pud(pud, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pud_flags(pud) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pud_page(pud);
-	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pud_range(p4d_t p4d, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pud_t *pudp;
-
-	pudp = pud_offset(&p4d, addr);
-	do {
-		pud_t pud = *pudp;
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			return 0;
-		if (unlikely(pud_large(pud))) {
-			if (!gup_huge_pud(pud, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pmd_range(pud, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pudp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static int gup_p4d_range(pgd_t pgd, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	p4d_t *p4dp;
-
-	p4dp = p4d_offset(&pgd, addr);
-	do {
-		p4d_t p4d = *p4dp;
-
-		next = p4d_addr_end(addr, end);
-		if (p4d_none(p4d))
-			return 0;
-		BUILD_BUG_ON(p4d_large(p4d));
-		if (!gup_pud_range(p4d, addr, next, write, pages, nr))
-			return 0;
-	} while (p4dp++, addr = next, addr != end);
-
-	return 1;
-}
-
-/*
- * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
- * back to the regular GUP.
- */
-int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			  struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	unsigned long flags;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
-					(void __user *)start, len)))
-		return 0;
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_save(flags);
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			break;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			break;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_restore(flags);
-
-	return nr;
-}
-
-/**
- * get_user_pages_fast() - pin user pages in memory
- * @start:	starting user address
- * @nr_pages:	number of pages from start to pin
- * @write:	whether pages will be written to
- * @pages:	array that receives pointers to the pages pinned.
- * 		Should be at least nr_pages long.
- *
- * Attempt to pin user pages in memory without taking mm->mmap_sem.
- * If not successful, it will fall back to taking the lock and
- * calling get_user_pages().
- *
- * Returns number of pages pinned. This may be fewer than the number
- * requested. If nr_pages is 0 or negative, returns 0. If no pages
- * were pinned, returns -errno.
- */
-int get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-
-	end = start + len;
-	if (end < start)
-		goto slow_irqon;
-
-#ifdef CONFIG_X86_64
-	if (end >> __VIRTUAL_MASK_SHIFT)
-		goto slow_irqon;
-#endif
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		ret = get_user_pages_unlocked(start,
-					      (end - start) >> PAGE_SHIFT,
-					      pages, write ? FOLL_WRITE : 0);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
-}
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Peter Zijlstra]

On Thu, Mar 16, 2017 at 06:26:55PM +0300, Kirill A. Shutemov wrote:
> +config HAVE_GENERIC_RCU_GUP
> +	def_bool y
> +

Nothing immediately jumped out to me; except that this option might be
misnamed.

AFAICT that code does not in fact rely on HAVE_RCU_TABLE_FREE; it will
happily work with the (x86) broadcast IPI invalidate model, as you show
here.

Architectures that do not do that obviously need HAVE_RCU_TABLE_FREE,
but that is not the point I feel.

Also, this code hard relies on IRQ-disable delaying grace periods, which
is mostly true I think, but has always been something Paul didn't really
want to commit too firmly to.

Re: [PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Peter Zijlstra]

On Thu, Mar 16, 2017 at 06:26:55PM +0300, Kirill A. Shutemov wrote:
> +config HAVE_GENERIC_RCU_GUP
> +	def_bool y
> +

Nothing immediately jumped out to me; except that this option might be
misnamed.

AFAICT that code does not in fact rely on HAVE_RCU_TABLE_FREE; it will
happily work with the (x86) broadcast IPI invalidate model, as you show
here.

Architectures that do not do that obviously need HAVE_RCU_TABLE_FREE,
but that is not the point I feel.

Also, this code hard relies on IRQ-disable delaying grace periods, which
is mostly true I think, but has always been something Paul didn't really
want to commit too firmly to.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Linus Torvalds]

On Thu, Mar 16, 2017 at 8:26 AM, Kirill A. Shutemov
<kirill.shutemov@linux.intel.com> wrote:
>
> The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
> to be usable for x86 and switches x86 over.

Thanks for doing this, it looks good and removes more lines than it adds.

And despite removing lines, it should make it easier for other
architectures to support devmap if they ever want to. I hadn't even
noticed that difference in the GUP implementations.

             Linus

Re: [PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Linus Torvalds]

On Thu, Mar 16, 2017 at 8:26 AM, Kirill A. Shutemov
<kirill.shutemov@linux.intel.com> wrote:
>
> The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
> to be usable for x86 and switches x86 over.

Thanks for doing this, it looks good and removes more lines than it adds.

And despite removing lines, it should make it easier for other
architectures to support devmap if they ever want to. I hadn't even
noticed that difference in the GUP implementations.

             Linus

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Paul E. McKenney]

On Thu, Mar 16, 2017 at 06:20:46PM +0100, Peter Zijlstra wrote:
> On Thu, Mar 16, 2017 at 06:26:55PM +0300, Kirill A. Shutemov wrote:
> > +config HAVE_GENERIC_RCU_GUP
> > +	def_bool y
> > +
> 
> Nothing immediately jumped out to me; except that this option might be
> misnamed.
> 
> AFAICT that code does not in fact rely on HAVE_RCU_TABLE_FREE; it will
> happily work with the (x86) broadcast IPI invalidate model, as you show
> here.
> 
> Architectures that do not do that obviously need HAVE_RCU_TABLE_FREE,
> but that is not the point I feel.
> 
> Also, this code hard relies on IRQ-disable delaying grace periods, which
> is mostly true I think, but has always been something Paul didn't really
> want to commit too firmly to.

That is quite true!

The only case where IRQ-disable is guaranteed to delay grace periods is
when you are using RCU-sched, in other words synchronize_sched() and
call_rcu_sched().  And even then, the CPU cannot be in the idle loop,
cannot be offline, and cannot be a nohz_full CPU on its way to/from
userspace execution.

							Thanx, Paul

Re: [PATCH 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Paul E. McKenney]

On Thu, Mar 16, 2017 at 06:20:46PM +0100, Peter Zijlstra wrote:
> On Thu, Mar 16, 2017 at 06:26:55PM +0300, Kirill A. Shutemov wrote:
> > +config HAVE_GENERIC_RCU_GUP
> > +	def_bool y
> > +
> 
> Nothing immediately jumped out to me; except that this option might be
> misnamed.
> 
> AFAICT that code does not in fact rely on HAVE_RCU_TABLE_FREE; it will
> happily work with the (x86) broadcast IPI invalidate model, as you show
> here.
> 
> Architectures that do not do that obviously need HAVE_RCU_TABLE_FREE,
> but that is not the point I feel.
> 
> Also, this code hard relies on IRQ-disable delaying grace periods, which
> is mostly true I think, but has always been something Paul didn't really
> want to commit too firmly to.

That is quite true!

The only case where IRQ-disable is guaranteed to delay grace periods is
when you are using RCU-sched, in other words synchronize_sched() and
call_rcu_sched().  And even then, the CPU cannot be in the idle loop,
cannot be offline, and cannot be a nohz_full CPU on its way to/from
userspace execution.

							Thanx, Paul

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCHv2 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

The patch provides all required hooks to match generic
get_user_pages_fast() behaviour to x86 and switch x86 over.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 v2:
    - Fix build on non-PAE 32-bit x86;
    - Rename HAVE_GENERIC_RCU_GUP to HAVE_GENERIC_GUP and update comment;
---
 arch/arm/Kconfig                      |   2 +-
 arch/arm64/Kconfig                    |   2 +-
 arch/powerpc/Kconfig                  |   2 +-
 arch/x86/Kconfig                      |   3 +
 arch/x86/include/asm/mmu_context.h    |  12 -
 arch/x86/include/asm/pgtable-3level.h |  45 +++
 arch/x86/include/asm/pgtable.h        |  53 ++++
 arch/x86/include/asm/pgtable_64.h     |  16 +-
 arch/x86/mm/Makefile                  |   2 +-
 arch/x86/mm/gup.c                     | 496 ----------------------------------
 mm/Kconfig                            |   2 +-
 mm/gup.c                              |  10 +-
 12 files changed, 126 insertions(+), 519 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 0d4e71b42c77..454fadd077ad 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -1666,7 +1666,7 @@ config ARCH_SELECT_MEMORY_MODEL
 config HAVE_ARCH_PFN_VALID
 	def_bool ARCH_HAS_HOLES_MEMORYMODEL || !SPARSEMEM
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 	depends on ARM_LPAE
 
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 8c7c244247b6..e0306e4b6cc8 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -205,7 +205,7 @@ config GENERIC_CALIBRATE_DELAY
 config ZONE_DMA
 	def_bool y
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 
 config ARCH_DMA_ADDR_T_64BIT
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 97a8bc8a095c..3a716b2dcde9 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -135,7 +135,7 @@ config PPC
 	select HAVE_FUNCTION_GRAPH_TRACER
 	select HAVE_FUNCTION_TRACER
 	select HAVE_GCC_PLUGINS
-	select HAVE_GENERIC_RCU_GUP
+	select HAVE_GENERIC_GUP
 	select HAVE_HW_BREAKPOINT		if PERF_EVENTS && (PPC_BOOK3S || PPC_8xx)
 	select HAVE_IDE
 	select HAVE_IOREMAP_PROT
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2bab9d093b51..8977d9c77373 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2788,6 +2788,9 @@ config X86_DMA_REMAP
 	bool
 	depends on STA2X11
 
+config HAVE_GENERIC_GUP
+	def_bool y
+
 source "net/Kconfig"
 
 source "drivers/Kconfig"
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 68b329d77b3a..6e933d2d88d9 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -220,18 +220,6 @@ static inline int vma_pkey(struct vm_area_struct *vma)
 }
 #endif
 
-static inline bool __pkru_allows_pkey(u16 pkey, bool write)
-{
-	u32 pkru = read_pkru();
-
-	if (!__pkru_allows_read(pkru, pkey))
-		return false;
-	if (write && !__pkru_allows_write(pkru, pkey))
-		return false;
-
-	return true;
-}
-
 /*
  * We only want to enforce protection keys on the current process
  * because we effectively have no access to PKRU for other
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index 72277b1028a5..698060a0dfb6 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -215,4 +215,49 @@ static inline pud_t native_pudp_get_and_clear(pud_t *pudp)
 #define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t){ { .pte_high = (x).val } })
 
+#define gup_get_pte gup_get_pte
+/*
+ * WARNING: only to be used in get_user_pages_fast() implementation.
+ *
+ * With get_user_pages_fast, we walk down the pagetables without taking
+ * any locks.  For this we would like to load the pointers atomically,
+ * but that is not possible (without expensive cmpxchg8b) on PAE.  What
+ * we do have is the guarantee that a pte will only either go from not
+ * present to present, or present to not present or both -- it will not
+ * switch to a completely different present page without a TLB flush in
+ * between; something that we are blocking by holding interrupts off.
+ *
+ * Setting ptes from not present to present goes:
+ * ptep->pte_high = h;
+ * smp_wmb();
+ * ptep->pte_low = l;
+ *
+ * And present to not present goes:
+ * ptep->pte_low = 0;
+ * smp_wmb();
+ * ptep->pte_high = 0;
+ *
+ * We must ensure here that the load of pte_low sees l iff pte_high
+ * sees h. We load pte_high *after* loading pte_low, which ensures we
+ * don't see an older value of pte_high.  *Then* we recheck pte_low,
+ * which ensures that we haven't picked up a changed pte high. We might
+ * have got rubbish values from pte_low and pte_high, but we are
+ * guaranteed that pte_low will not have the present bit set *unless*
+ * it is 'l'. And get_user_pages_fast only operates on present ptes, so
+ * we're safe.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	pte_t pte;
+
+	do {
+		pte.pte_low = ptep->pte_low;
+		smp_rmb();
+		pte.pte_high = ptep->pte_high;
+		smp_rmb();
+	} while (unlikely(pte.pte_low != ptep->pte_low));
+
+	return pte;
+}
+
 #endif /* _ASM_X86_PGTABLE_3LEVEL_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 83f17218e66a..e55fe9475979 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -230,6 +230,11 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 #endif
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
@@ -1135,6 +1140,54 @@ static inline u16 pte_flags_pkey(unsigned long pte_flags)
 #endif
 }
 
+static inline bool __pkru_allows_pkey(u16 pkey, bool write)
+{
+	u32 pkru = read_pkru();
+
+	if (!__pkru_allows_read(pkru, pkey))
+		return false;
+	if (write && !__pkru_allows_write(pkru, pkey))
+		return false;
+
+	return true;
+}
+
+/*
+ * 'pteval' can come from a pte, pmd or pud.  We only check
+ * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
+ * same value on all 3 types.
+ */
+static inline bool __pte_access_permitted(unsigned long pteval, bool write)
+{
+	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
+
+	if (write)
+		need_pte_bits |= _PAGE_RW;
+
+	if ((pteval & need_pte_bits) != need_pte_bits)
+		return 0;
+
+	return __pkru_allows_pkey(pte_flags_pkey(pteval), write);
+}
+
+#define pte_access_permitted pte_access_permitted
+static inline bool pte_access_permitted(pte_t pte, bool write)
+{
+	return __pte_access_permitted(pte_val(pte), write);
+}
+
+#define pmd_access_permitted pmd_access_permitted
+static inline bool pmd_access_permitted(pmd_t pmd, bool write)
+{
+	return __pte_access_permitted(pmd_val(pmd), write);
+}
+
+#define pud_access_permitted pud_access_permitted
+static inline bool pud_access_permitted(pud_t pud, bool write)
+{
+	return __pte_access_permitted(pud_val(pud), write);
+}
+
 #include <asm-generic/pgtable.h>
 #endif	/* __ASSEMBLY__ */
 
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 73c7ccc38912..1a4bc71534d4 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -206,6 +206,20 @@ extern void cleanup_highmap(void);
 extern void init_extra_mapping_uc(unsigned long phys, unsigned long size);
 extern void init_extra_mapping_wb(unsigned long phys, unsigned long size);
 
-#endif /* !__ASSEMBLY__ */
+#define gup_fast_permitted gup_fast_permitted
+static inline bool gup_fast_permitted(unsigned long start, int nr_pages,
+		int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long)nr_pages << PAGE_SHIFT;
+	end = start + len;
+	if (end < start)
+		return false;
+	if (end >> __VIRTUAL_MASK_SHIFT)
+		return false;
+	return true;
+}
 
+#endif /* !__ASSEMBLY__ */
 #endif /* _ASM_X86_PGTABLE_64_H */
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 96d2b847e09e..0fbdcb64f9f8 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -2,7 +2,7 @@
 KCOV_INSTRUMENT_tlb.o	:= n
 
 obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o gup.o setup_nx.o tlb.o
+	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
deleted file mode 100644
index 456dfdfd2249..000000000000
--- a/arch/x86/mm/gup.c
+++ /dev/null
@@ -1,496 +0,0 @@
-/*
- * Lockless get_user_pages_fast for x86
- *
- * Copyright (C) 2008 Nick Piggin
- * Copyright (C) 2008 Novell Inc.
- */
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmstat.h>
-#include <linux/highmem.h>
-#include <linux/swap.h>
-#include <linux/memremap.h>
-
-#include <asm/mmu_context.h>
-#include <asm/pgtable.h>
-
-static inline pte_t gup_get_pte(pte_t *ptep)
-{
-#ifndef CONFIG_X86_PAE
-	return READ_ONCE(*ptep);
-#else
-	/*
-	 * With get_user_pages_fast, we walk down the pagetables without taking
-	 * any locks.  For this we would like to load the pointers atomically,
-	 * but that is not possible (without expensive cmpxchg8b) on PAE.  What
-	 * we do have is the guarantee that a pte will only either go from not
-	 * present to present, or present to not present or both -- it will not
-	 * switch to a completely different present page without a TLB flush in
-	 * between; something that we are blocking by holding interrupts off.
-	 *
-	 * Setting ptes from not present to present goes:
-	 * ptep->pte_high = h;
-	 * smp_wmb();
-	 * ptep->pte_low = l;
-	 *
-	 * And present to not present goes:
-	 * ptep->pte_low = 0;
-	 * smp_wmb();
-	 * ptep->pte_high = 0;
-	 *
-	 * We must ensure here that the load of pte_low sees l iff pte_high
-	 * sees h. We load pte_high *after* loading pte_low, which ensures we
-	 * don't see an older value of pte_high.  *Then* we recheck pte_low,
-	 * which ensures that we haven't picked up a changed pte high. We might
-	 * have got rubbish values from pte_low and pte_high, but we are
-	 * guaranteed that pte_low will not have the present bit set *unless*
-	 * it is 'l'. And get_user_pages_fast only operates on present ptes, so
-	 * we're safe.
-	 *
-	 * gup_get_pte should not be used or copied outside gup.c without being
-	 * very careful -- it does not atomically load the pte or anything that
-	 * is likely to be useful for you.
-	 */
-	pte_t pte;
-
-retry:
-	pte.pte_low = ptep->pte_low;
-	smp_rmb();
-	pte.pte_high = ptep->pte_high;
-	smp_rmb();
-	if (unlikely(pte.pte_low != ptep->pte_low))
-		goto retry;
-
-	return pte;
-#endif
-}
-
-static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
-{
-	while ((*nr) - nr_start) {
-		struct page *page = pages[--(*nr)];
-
-		ClearPageReferenced(page);
-		put_page(page);
-	}
-}
-
-/*
- * 'pteval' can come from a pte, pmd, pud or p4d.  We only check
- * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
- * same value on all 4 types.
- */
-static inline int pte_allows_gup(unsigned long pteval, int write)
-{
-	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
-
-	if (write)
-		need_pte_bits |= _PAGE_RW;
-
-	if ((pteval & need_pte_bits) != need_pte_bits)
-		return 0;
-
-	/* Check memory protection keys permissions. */
-	if (!__pkru_allows_pkey(pte_flags_pkey(pteval), write))
-		return 0;
-
-	return 1;
-}
-
-/*
- * The performance critical leaf functions are made noinline otherwise gcc
- * inlines everything into a single function which results in too much
- * register pressure.
- */
-static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct dev_pagemap *pgmap = NULL;
-	int nr_start = *nr, ret = 0;
-	pte_t *ptep, *ptem;
-
-	/*
-	 * Keep the original mapped PTE value (ptem) around since we
-	 * might increment ptep off the end of the page when finishing
-	 * our loop iteration.
-	 */
-	ptem = ptep = pte_offset_map(&pmd, addr);
-	do {
-		pte_t pte = gup_get_pte(ptep);
-		struct page *page;
-
-		/* Similar to the PMD case, NUMA hinting must take slow path */
-		if (pte_protnone(pte))
-			break;
-
-		if (!pte_allows_gup(pte_val(pte), write))
-			break;
-
-		if (pte_devmap(pte)) {
-			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
-			if (unlikely(!pgmap)) {
-				undo_dev_pagemap(nr, nr_start, pages);
-				break;
-			}
-		} else if (pte_special(pte))
-			break;
-
-		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
-		page = pte_page(pte);
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		(*nr)++;
-
-	} while (ptep++, addr += PAGE_SIZE, addr != end);
-	if (addr == end)
-		ret = 1;
-	pte_unmap(ptem);
-
-	return ret;
-}
-
-static inline void get_head_page_multiple(struct page *page, int nr)
-{
-	VM_BUG_ON_PAGE(page != compound_head(page), page);
-	VM_BUG_ON_PAGE(page_count(page) == 0, page);
-	page_ref_add(page, nr);
-	SetPageReferenced(page);
-}
-
-static int __gup_device_huge(unsigned long pfn, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	int nr_start = *nr;
-	struct dev_pagemap *pgmap = NULL;
-
-	do {
-		struct page *page = pfn_to_page(pfn);
-
-		pgmap = get_dev_pagemap(pfn, pgmap);
-		if (unlikely(!pgmap)) {
-			undo_dev_pagemap(nr, nr_start, pages);
-			return 0;
-		}
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		(*nr)++;
-		pfn++;
-	} while (addr += PAGE_SIZE, addr != end);
-	return 1;
-}
-
-static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static noinline int gup_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pmd_val(pmd), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pmd_pfn(pmd)));
-	if (pmd_devmap(pmd))
-		return __gup_device_huge_pmd(pmd, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pmd_flags(pmd) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pmd_page(pmd);
-	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end,
-		int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pmd_t *pmdp;
-
-	pmdp = pmd_offset(&pud, addr);
-	do {
-		pmd_t pmd = *pmdp;
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			return 0;
-		if (unlikely(pmd_large(pmd) || !pmd_present(pmd))) {
-			/*
-			 * NUMA hinting faults need to be handled in the GUP
-			 * slowpath for accounting purposes and so that they
-			 * can be serialised against THP migration.
-			 */
-			if (pmd_protnone(pmd))
-				return 0;
-			if (!gup_huge_pmd(pmd, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pte_range(pmd, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pmdp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static noinline int gup_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pud_val(pud), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pud_pfn(pud)));
-	if (pud_devmap(pud))
-		return __gup_device_huge_pud(pud, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pud_flags(pud) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pud_page(pud);
-	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pud_range(p4d_t p4d, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pud_t *pudp;
-
-	pudp = pud_offset(&p4d, addr);
-	do {
-		pud_t pud = *pudp;
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			return 0;
-		if (unlikely(pud_large(pud))) {
-			if (!gup_huge_pud(pud, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pmd_range(pud, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pudp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static int gup_p4d_range(pgd_t pgd, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	p4d_t *p4dp;
-
-	p4dp = p4d_offset(&pgd, addr);
-	do {
-		p4d_t p4d = *p4dp;
-
-		next = p4d_addr_end(addr, end);
-		if (p4d_none(p4d))
-			return 0;
-		BUILD_BUG_ON(p4d_large(p4d));
-		if (!gup_pud_range(p4d, addr, next, write, pages, nr))
-			return 0;
-	} while (p4dp++, addr = next, addr != end);
-
-	return 1;
-}
-
-/*
- * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
- * back to the regular GUP.
- */
-int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			  struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	unsigned long flags;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
-					(void __user *)start, len)))
-		return 0;
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_save(flags);
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			break;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			break;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_restore(flags);
-
-	return nr;
-}
-
-/**
- * get_user_pages_fast() - pin user pages in memory
- * @start:	starting user address
- * @nr_pages:	number of pages from start to pin
- * @write:	whether pages will be written to
- * @pages:	array that receives pointers to the pages pinned.
- * 		Should be at least nr_pages long.
- *
- * Attempt to pin user pages in memory without taking mm->mmap_sem.
- * If not successful, it will fall back to taking the lock and
- * calling get_user_pages().
- *
- * Returns number of pages pinned. This may be fewer than the number
- * requested. If nr_pages is 0 or negative, returns 0. If no pages
- * were pinned, returns -errno.
- */
-int get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-
-	end = start + len;
-	if (end < start)
-		goto slow_irqon;
-
-#ifdef CONFIG_X86_64
-	if (end >> __VIRTUAL_MASK_SHIFT)
-		goto slow_irqon;
-#endif
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		ret = get_user_pages_unlocked(start,
-					      (end - start) >> PAGE_SHIFT,
-					      pages, write ? FOLL_WRITE : 0);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
-}
diff --git a/mm/Kconfig b/mm/Kconfig
index 9b8fccb969dc..c89f472b658c 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -137,7 +137,7 @@ config HAVE_MEMBLOCK_NODE_MAP
 config HAVE_MEMBLOCK_PHYS_MAP
 	bool
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	bool
 
 config ARCH_DISCARD_MEMBLOCK
diff --git a/mm/gup.c b/mm/gup.c
index a6e21380c8b0..5cc489d98562 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1155,7 +1155,7 @@ struct page *get_dump_page(unsigned long addr)
 #endif /* CONFIG_ELF_CORE */
 
 /*
- * Generic RCU Fast GUP
+ * Generic Fast GUP
  *
  * get_user_pages_fast attempts to pin user pages by walking the page
  * tables directly and avoids taking locks. Thus the walker needs to be
@@ -1176,8 +1176,8 @@ struct page *get_dump_page(unsigned long addr)
  * Before activating this code, please be aware that the following assumptions
  * are currently made:
  *
- *  *) HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table is used to free
- *      pages containing page tables.
+ *  *) Either HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table is used to
+ *  free pages containing page tables or TLB flushing requires IPI broadcast.
  *
  *  *) ptes can be read atomically by the architecture.
  *
@@ -1187,7 +1187,7 @@ struct page *get_dump_page(unsigned long addr)
  *
  * This code is based heavily on the PowerPC implementation by Nick Piggin.
  */
-#ifdef CONFIG_HAVE_GENERIC_RCU_GUP
+#ifdef CONFIG_HAVE_GENERIC_GUP
 
 #ifndef gup_get_pte
 /*
@@ -1663,4 +1663,4 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return ret;
 }
 
-#endif /* CONFIG_HAVE_GENERIC_RCU_GUP */
+#endif /* CONFIG_HAVE_GENERIC_GUP */
-- 
2.11.0

[PATCHv2 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

The patch provides all required hooks to match generic
get_user_pages_fast() behaviour to x86 and switch x86 over.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 v2:
    - Fix build on non-PAE 32-bit x86;
    - Rename HAVE_GENERIC_RCU_GUP to HAVE_GENERIC_GUP and update comment;
---
 arch/arm/Kconfig                      |   2 +-
 arch/arm64/Kconfig                    |   2 +-
 arch/powerpc/Kconfig                  |   2 +-
 arch/x86/Kconfig                      |   3 +
 arch/x86/include/asm/mmu_context.h    |  12 -
 arch/x86/include/asm/pgtable-3level.h |  45 +++
 arch/x86/include/asm/pgtable.h        |  53 ++++
 arch/x86/include/asm/pgtable_64.h     |  16 +-
 arch/x86/mm/Makefile                  |   2 +-
 arch/x86/mm/gup.c                     | 496 ----------------------------------
 mm/Kconfig                            |   2 +-
 mm/gup.c                              |  10 +-
 12 files changed, 126 insertions(+), 519 deletions(-)
 delete mode 100644 arch/x86/mm/gup.c

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 0d4e71b42c77..454fadd077ad 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -1666,7 +1666,7 @@ config ARCH_SELECT_MEMORY_MODEL
 config HAVE_ARCH_PFN_VALID
 	def_bool ARCH_HAS_HOLES_MEMORYMODEL || !SPARSEMEM
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 	depends on ARM_LPAE
 
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 8c7c244247b6..e0306e4b6cc8 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -205,7 +205,7 @@ config GENERIC_CALIBRATE_DELAY
 config ZONE_DMA
 	def_bool y
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 
 config ARCH_DMA_ADDR_T_64BIT
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 97a8bc8a095c..3a716b2dcde9 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -135,7 +135,7 @@ config PPC
 	select HAVE_FUNCTION_GRAPH_TRACER
 	select HAVE_FUNCTION_TRACER
 	select HAVE_GCC_PLUGINS
-	select HAVE_GENERIC_RCU_GUP
+	select HAVE_GENERIC_GUP
 	select HAVE_HW_BREAKPOINT		if PERF_EVENTS && (PPC_BOOK3S || PPC_8xx)
 	select HAVE_IDE
 	select HAVE_IOREMAP_PROT
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2bab9d093b51..8977d9c77373 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2788,6 +2788,9 @@ config X86_DMA_REMAP
 	bool
 	depends on STA2X11
 
+config HAVE_GENERIC_GUP
+	def_bool y
+
 source "net/Kconfig"
 
 source "drivers/Kconfig"
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 68b329d77b3a..6e933d2d88d9 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -220,18 +220,6 @@ static inline int vma_pkey(struct vm_area_struct *vma)
 }
 #endif
 
-static inline bool __pkru_allows_pkey(u16 pkey, bool write)
-{
-	u32 pkru = read_pkru();
-
-	if (!__pkru_allows_read(pkru, pkey))
-		return false;
-	if (write && !__pkru_allows_write(pkru, pkey))
-		return false;
-
-	return true;
-}
-
 /*
  * We only want to enforce protection keys on the current process
  * because we effectively have no access to PKRU for other
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index 72277b1028a5..698060a0dfb6 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -215,4 +215,49 @@ static inline pud_t native_pudp_get_and_clear(pud_t *pudp)
 #define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t){ { .pte_high = (x).val } })
 
+#define gup_get_pte gup_get_pte
+/*
+ * WARNING: only to be used in get_user_pages_fast() implementation.
+ *
+ * With get_user_pages_fast, we walk down the pagetables without taking
+ * any locks.  For this we would like to load the pointers atomically,
+ * but that is not possible (without expensive cmpxchg8b) on PAE.  What
+ * we do have is the guarantee that a pte will only either go from not
+ * present to present, or present to not present or both -- it will not
+ * switch to a completely different present page without a TLB flush in
+ * between; something that we are blocking by holding interrupts off.
+ *
+ * Setting ptes from not present to present goes:
+ * ptep->pte_high = h;
+ * smp_wmb();
+ * ptep->pte_low = l;
+ *
+ * And present to not present goes:
+ * ptep->pte_low = 0;
+ * smp_wmb();
+ * ptep->pte_high = 0;
+ *
+ * We must ensure here that the load of pte_low sees l iff pte_high
+ * sees h. We load pte_high *after* loading pte_low, which ensures we
+ * don't see an older value of pte_high.  *Then* we recheck pte_low,
+ * which ensures that we haven't picked up a changed pte high. We might
+ * have got rubbish values from pte_low and pte_high, but we are
+ * guaranteed that pte_low will not have the present bit set *unless*
+ * it is 'l'. And get_user_pages_fast only operates on present ptes, so
+ * we're safe.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	pte_t pte;
+
+	do {
+		pte.pte_low = ptep->pte_low;
+		smp_rmb();
+		pte.pte_high = ptep->pte_high;
+		smp_rmb();
+	} while (unlikely(pte.pte_low != ptep->pte_low));
+
+	return pte;
+}
+
 #endif /* _ASM_X86_PGTABLE_3LEVEL_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 83f17218e66a..e55fe9475979 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -230,6 +230,11 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 #endif
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
@@ -1135,6 +1140,54 @@ static inline u16 pte_flags_pkey(unsigned long pte_flags)
 #endif
 }
 
+static inline bool __pkru_allows_pkey(u16 pkey, bool write)
+{
+	u32 pkru = read_pkru();
+
+	if (!__pkru_allows_read(pkru, pkey))
+		return false;
+	if (write && !__pkru_allows_write(pkru, pkey))
+		return false;
+
+	return true;
+}
+
+/*
+ * 'pteval' can come from a pte, pmd or pud.  We only check
+ * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
+ * same value on all 3 types.
+ */
+static inline bool __pte_access_permitted(unsigned long pteval, bool write)
+{
+	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
+
+	if (write)
+		need_pte_bits |= _PAGE_RW;
+
+	if ((pteval & need_pte_bits) != need_pte_bits)
+		return 0;
+
+	return __pkru_allows_pkey(pte_flags_pkey(pteval), write);
+}
+
+#define pte_access_permitted pte_access_permitted
+static inline bool pte_access_permitted(pte_t pte, bool write)
+{
+	return __pte_access_permitted(pte_val(pte), write);
+}
+
+#define pmd_access_permitted pmd_access_permitted
+static inline bool pmd_access_permitted(pmd_t pmd, bool write)
+{
+	return __pte_access_permitted(pmd_val(pmd), write);
+}
+
+#define pud_access_permitted pud_access_permitted
+static inline bool pud_access_permitted(pud_t pud, bool write)
+{
+	return __pte_access_permitted(pud_val(pud), write);
+}
+
 #include <asm-generic/pgtable.h>
 #endif	/* __ASSEMBLY__ */
 
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 73c7ccc38912..1a4bc71534d4 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -206,6 +206,20 @@ extern void cleanup_highmap(void);
 extern void init_extra_mapping_uc(unsigned long phys, unsigned long size);
 extern void init_extra_mapping_wb(unsigned long phys, unsigned long size);
 
-#endif /* !__ASSEMBLY__ */
+#define gup_fast_permitted gup_fast_permitted
+static inline bool gup_fast_permitted(unsigned long start, int nr_pages,
+		int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long)nr_pages << PAGE_SHIFT;
+	end = start + len;
+	if (end < start)
+		return false;
+	if (end >> __VIRTUAL_MASK_SHIFT)
+		return false;
+	return true;
+}
 
+#endif /* !__ASSEMBLY__ */
 #endif /* _ASM_X86_PGTABLE_64_H */
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 96d2b847e09e..0fbdcb64f9f8 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -2,7 +2,7 @@
 KCOV_INSTRUMENT_tlb.o	:= n
 
 obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o gup.o setup_nx.o tlb.o
+	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
deleted file mode 100644
index 456dfdfd2249..000000000000
--- a/arch/x86/mm/gup.c
+++ /dev/null
@@ -1,496 +0,0 @@
-/*
- * Lockless get_user_pages_fast for x86
- *
- * Copyright (C) 2008 Nick Piggin
- * Copyright (C) 2008 Novell Inc.
- */
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmstat.h>
-#include <linux/highmem.h>
-#include <linux/swap.h>
-#include <linux/memremap.h>
-
-#include <asm/mmu_context.h>
-#include <asm/pgtable.h>
-
-static inline pte_t gup_get_pte(pte_t *ptep)
-{
-#ifndef CONFIG_X86_PAE
-	return READ_ONCE(*ptep);
-#else
-	/*
-	 * With get_user_pages_fast, we walk down the pagetables without taking
-	 * any locks.  For this we would like to load the pointers atomically,
-	 * but that is not possible (without expensive cmpxchg8b) on PAE.  What
-	 * we do have is the guarantee that a pte will only either go from not
-	 * present to present, or present to not present or both -- it will not
-	 * switch to a completely different present page without a TLB flush in
-	 * between; something that we are blocking by holding interrupts off.
-	 *
-	 * Setting ptes from not present to present goes:
-	 * ptep->pte_high = h;
-	 * smp_wmb();
-	 * ptep->pte_low = l;
-	 *
-	 * And present to not present goes:
-	 * ptep->pte_low = 0;
-	 * smp_wmb();
-	 * ptep->pte_high = 0;
-	 *
-	 * We must ensure here that the load of pte_low sees l iff pte_high
-	 * sees h. We load pte_high *after* loading pte_low, which ensures we
-	 * don't see an older value of pte_high.  *Then* we recheck pte_low,
-	 * which ensures that we haven't picked up a changed pte high. We might
-	 * have got rubbish values from pte_low and pte_high, but we are
-	 * guaranteed that pte_low will not have the present bit set *unless*
-	 * it is 'l'. And get_user_pages_fast only operates on present ptes, so
-	 * we're safe.
-	 *
-	 * gup_get_pte should not be used or copied outside gup.c without being
-	 * very careful -- it does not atomically load the pte or anything that
-	 * is likely to be useful for you.
-	 */
-	pte_t pte;
-
-retry:
-	pte.pte_low = ptep->pte_low;
-	smp_rmb();
-	pte.pte_high = ptep->pte_high;
-	smp_rmb();
-	if (unlikely(pte.pte_low != ptep->pte_low))
-		goto retry;
-
-	return pte;
-#endif
-}
-
-static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
-{
-	while ((*nr) - nr_start) {
-		struct page *page = pages[--(*nr)];
-
-		ClearPageReferenced(page);
-		put_page(page);
-	}
-}
-
-/*
- * 'pteval' can come from a pte, pmd, pud or p4d.  We only check
- * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
- * same value on all 4 types.
- */
-static inline int pte_allows_gup(unsigned long pteval, int write)
-{
-	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
-
-	if (write)
-		need_pte_bits |= _PAGE_RW;
-
-	if ((pteval & need_pte_bits) != need_pte_bits)
-		return 0;
-
-	/* Check memory protection keys permissions. */
-	if (!__pkru_allows_pkey(pte_flags_pkey(pteval), write))
-		return 0;
-
-	return 1;
-}
-
-/*
- * The performance critical leaf functions are made noinline otherwise gcc
- * inlines everything into a single function which results in too much
- * register pressure.
- */
-static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct dev_pagemap *pgmap = NULL;
-	int nr_start = *nr, ret = 0;
-	pte_t *ptep, *ptem;
-
-	/*
-	 * Keep the original mapped PTE value (ptem) around since we
-	 * might increment ptep off the end of the page when finishing
-	 * our loop iteration.
-	 */
-	ptem = ptep = pte_offset_map(&pmd, addr);
-	do {
-		pte_t pte = gup_get_pte(ptep);
-		struct page *page;
-
-		/* Similar to the PMD case, NUMA hinting must take slow path */
-		if (pte_protnone(pte))
-			break;
-
-		if (!pte_allows_gup(pte_val(pte), write))
-			break;
-
-		if (pte_devmap(pte)) {
-			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
-			if (unlikely(!pgmap)) {
-				undo_dev_pagemap(nr, nr_start, pages);
-				break;
-			}
-		} else if (pte_special(pte))
-			break;
-
-		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
-		page = pte_page(pte);
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		(*nr)++;
-
-	} while (ptep++, addr += PAGE_SIZE, addr != end);
-	if (addr == end)
-		ret = 1;
-	pte_unmap(ptem);
-
-	return ret;
-}
-
-static inline void get_head_page_multiple(struct page *page, int nr)
-{
-	VM_BUG_ON_PAGE(page != compound_head(page), page);
-	VM_BUG_ON_PAGE(page_count(page) == 0, page);
-	page_ref_add(page, nr);
-	SetPageReferenced(page);
-}
-
-static int __gup_device_huge(unsigned long pfn, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	int nr_start = *nr;
-	struct dev_pagemap *pgmap = NULL;
-
-	do {
-		struct page *page = pfn_to_page(pfn);
-
-		pgmap = get_dev_pagemap(pfn, pgmap);
-		if (unlikely(!pgmap)) {
-			undo_dev_pagemap(nr, nr_start, pages);
-			return 0;
-		}
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		(*nr)++;
-		pfn++;
-	} while (addr += PAGE_SIZE, addr != end);
-	return 1;
-}
-
-static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static noinline int gup_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pmd_val(pmd), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pmd_pfn(pmd)));
-	if (pmd_devmap(pmd))
-		return __gup_device_huge_pmd(pmd, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pmd_flags(pmd) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pmd_page(pmd);
-	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end,
-		int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pmd_t *pmdp;
-
-	pmdp = pmd_offset(&pud, addr);
-	do {
-		pmd_t pmd = *pmdp;
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			return 0;
-		if (unlikely(pmd_large(pmd) || !pmd_present(pmd))) {
-			/*
-			 * NUMA hinting faults need to be handled in the GUP
-			 * slowpath for accounting purposes and so that they
-			 * can be serialised against THP migration.
-			 */
-			if (pmd_protnone(pmd))
-				return 0;
-			if (!gup_huge_pmd(pmd, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pte_range(pmd, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pmdp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static noinline int gup_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pud_val(pud), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pud_pfn(pud)));
-	if (pud_devmap(pud))
-		return __gup_device_huge_pud(pud, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pud_flags(pud) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pud_page(pud);
-	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pud_range(p4d_t p4d, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pud_t *pudp;
-
-	pudp = pud_offset(&p4d, addr);
-	do {
-		pud_t pud = *pudp;
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			return 0;
-		if (unlikely(pud_large(pud))) {
-			if (!gup_huge_pud(pud, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pmd_range(pud, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pudp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static int gup_p4d_range(pgd_t pgd, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	p4d_t *p4dp;
-
-	p4dp = p4d_offset(&pgd, addr);
-	do {
-		p4d_t p4d = *p4dp;
-
-		next = p4d_addr_end(addr, end);
-		if (p4d_none(p4d))
-			return 0;
-		BUILD_BUG_ON(p4d_large(p4d));
-		if (!gup_pud_range(p4d, addr, next, write, pages, nr))
-			return 0;
-	} while (p4dp++, addr = next, addr != end);
-
-	return 1;
-}
-
-/*
- * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
- * back to the regular GUP.
- */
-int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			  struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	unsigned long flags;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
-					(void __user *)start, len)))
-		return 0;
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_save(flags);
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			break;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			break;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_restore(flags);
-
-	return nr;
-}
-
-/**
- * get_user_pages_fast() - pin user pages in memory
- * @start:	starting user address
- * @nr_pages:	number of pages from start to pin
- * @write:	whether pages will be written to
- * @pages:	array that receives pointers to the pages pinned.
- * 		Should be at least nr_pages long.
- *
- * Attempt to pin user pages in memory without taking mm->mmap_sem.
- * If not successful, it will fall back to taking the lock and
- * calling get_user_pages().
- *
- * Returns number of pages pinned. This may be fewer than the number
- * requested. If nr_pages is 0 or negative, returns 0. If no pages
- * were pinned, returns -errno.
- */
-int get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-
-	end = start + len;
-	if (end < start)
-		goto slow_irqon;
-
-#ifdef CONFIG_X86_64
-	if (end >> __VIRTUAL_MASK_SHIFT)
-		goto slow_irqon;
-#endif
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		ret = get_user_pages_unlocked(start,
-					      (end - start) >> PAGE_SHIFT,
-					      pages, write ? FOLL_WRITE : 0);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
-}
diff --git a/mm/Kconfig b/mm/Kconfig
index 9b8fccb969dc..c89f472b658c 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -137,7 +137,7 @@ config HAVE_MEMBLOCK_NODE_MAP
 config HAVE_MEMBLOCK_PHYS_MAP
 	bool
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	bool
 
 config ARCH_DISCARD_MEMBLOCK
diff --git a/mm/gup.c b/mm/gup.c
index a6e21380c8b0..5cc489d98562 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1155,7 +1155,7 @@ struct page *get_dump_page(unsigned long addr)
 #endif /* CONFIG_ELF_CORE */
 
 /*
- * Generic RCU Fast GUP
+ * Generic Fast GUP
  *
  * get_user_pages_fast attempts to pin user pages by walking the page
  * tables directly and avoids taking locks. Thus the walker needs to be
@@ -1176,8 +1176,8 @@ struct page *get_dump_page(unsigned long addr)
  * Before activating this code, please be aware that the following assumptions
  * are currently made:
  *
- *  *) HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table is used to free
- *      pages containing page tables.
+ *  *) Either HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table is used to
+ *  free pages containing page tables or TLB flushing requires IPI broadcast.
  *
  *  *) ptes can be read atomically by the architecture.
  *
@@ -1187,7 +1187,7 @@ struct page *get_dump_page(unsigned long addr)
  *
  * This code is based heavily on the PowerPC implementation by Nick Piggin.
  */
-#ifdef CONFIG_HAVE_GENERIC_RCU_GUP
+#ifdef CONFIG_HAVE_GENERIC_GUP
 
 #ifndef gup_get_pte
 /*
@@ -1663,4 +1663,4 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return ret;
 }
 
-#endif /* CONFIG_HAVE_GENERIC_RCU_GUP */
+#endif /* CONFIG_HAVE_GENERIC_GUP */
-- 
2.11.0

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Ingo Molnar]

* Kirill A. Shutemov <kirill.shutemov@linux.intel.com> wrote:

> Hi,
> 
> The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
> to be usable for x86 and switches x86 over.
> 
> Please review and consider applying.
> 
> Kirill A. Shutemov (7):
>   mm: Drop arch_pte_access_permitted() mmu hook
>   mm/gup: Move permission checks into helpers
>   mm/gup: Move page table entry dereference into helper
>   mm/gup: Make pages referenced during generic get_user_pages_fast()
>   mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()
>   mm/gup: Provide hook to check if __GUP_fast() is allowed for the range
>   x86/mm: Switch to generic get_user_page_fast() implementation
> 
>  arch/powerpc/include/asm/mmu_context.h   |   6 -
>  arch/s390/include/asm/mmu_context.h      |   6 -
>  arch/um/include/asm/mmu_context.h        |   6 -
>  arch/unicore32/include/asm/mmu_context.h |   6 -
>  arch/x86/Kconfig                         |   3 +
>  arch/x86/include/asm/mmu_context.h       |  16 -
>  arch/x86/include/asm/pgtable-3level.h    |  45 +++
>  arch/x86/include/asm/pgtable.h           |  53 ++++
>  arch/x86/include/asm/pgtable_64.h        |  16 +-
>  arch/x86/mm/Makefile                     |   2 +-
>  arch/x86/mm/gup.c                        | 496 -------------------------------
>  include/asm-generic/mm_hooks.h           |   6 -
>  include/asm-generic/pgtable.h            |  25 ++
>  include/linux/mm.h                       |   4 +
>  mm/gup.c                                 | 134 +++++++--
>  15 files changed, 262 insertions(+), 562 deletions(-)
>  delete mode 100644 arch/x86/mm/gup.c

It fails to build on x86-64 defconfig/allyesconfig/allmodconfig:

  mm/gup.c:1422:15: error: implicit declaration of function ‘pgd_devmap’ [-Werror=implicit-function-declaration]

The PowerPC allnoconfig build broke as well:

/home/mingo/tip/mm/gup.c: In function '__gup_device_huge_pmd':
/home/mingo/tip/mm/gup.c:1319:2: error: implicit declaration of function 'pmd_pfn' [-Werror=implicit-function-declaration]

Please send delta fixes, because I've already done many small readability edits to 
the patches.

Thanks,

	Ingo

Re: [PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Ingo Molnar]

* Kirill A. Shutemov <kirill.shutemov@linux.intel.com> wrote:

> Hi,
> 
> The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
> to be usable for x86 and switches x86 over.
> 
> Please review and consider applying.
> 
> Kirill A. Shutemov (7):
>   mm: Drop arch_pte_access_permitted() mmu hook
>   mm/gup: Move permission checks into helpers
>   mm/gup: Move page table entry dereference into helper
>   mm/gup: Make pages referenced during generic get_user_pages_fast()
>   mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()
>   mm/gup: Provide hook to check if __GUP_fast() is allowed for the range
>   x86/mm: Switch to generic get_user_page_fast() implementation
> 
>  arch/powerpc/include/asm/mmu_context.h   |   6 -
>  arch/s390/include/asm/mmu_context.h      |   6 -
>  arch/um/include/asm/mmu_context.h        |   6 -
>  arch/unicore32/include/asm/mmu_context.h |   6 -
>  arch/x86/Kconfig                         |   3 +
>  arch/x86/include/asm/mmu_context.h       |  16 -
>  arch/x86/include/asm/pgtable-3level.h    |  45 +++
>  arch/x86/include/asm/pgtable.h           |  53 ++++
>  arch/x86/include/asm/pgtable_64.h        |  16 +-
>  arch/x86/mm/Makefile                     |   2 +-
>  arch/x86/mm/gup.c                        | 496 -------------------------------
>  include/asm-generic/mm_hooks.h           |   6 -
>  include/asm-generic/pgtable.h            |  25 ++
>  include/linux/mm.h                       |   4 +
>  mm/gup.c                                 | 134 +++++++--
>  15 files changed, 262 insertions(+), 562 deletions(-)
>  delete mode 100644 arch/x86/mm/gup.c

It fails to build on x86-64 defconfig/allyesconfig/allmodconfig:

  mm/gup.c:1422:15: error: implicit declaration of function ‘pgd_devmap’ [-Werror=implicit-function-declaration]

The PowerPC allnoconfig build broke as well:

/home/mingo/tip/mm/gup.c: In function '__gup_device_huge_pmd':
/home/mingo/tip/mm/gup.c:1319:2: error: implicit declaration of function 'pmd_pfn' [-Werror=implicit-function-declaration]

Please send delta fixes, because I've already done many small readability edits to 
the patches.

Thanks,

	Ingo

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 0/7] Switch x86 to generic get_user_pages_fast() implementation

[Ingo Molnar]

* Kirill A. Shutemov <kirill.shutemov@linux.intel.com> wrote:

> Hi,
> 
> The patcheset generalize mm/gup.c implementation of get_user_pages_fast()
> to be usable for x86 and switches x86 over.
> 
> Please review and consider applying.
> 
> Kirill A. Shutemov (7):
>   mm: Drop arch_pte_access_permitted() mmu hook
>   mm/gup: Move permission checks into helpers
>   mm/gup: Move page table entry dereference into helper
>   mm/gup: Make pages referenced during generic get_user_pages_fast()
>   mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()
>   mm/gup: Provide hook to check if __GUP_fast() is allowed for the range
>   x86/mm: Switch to generic get_user_page_fast() implementation
> 
>  arch/powerpc/include/asm/mmu_context.h   |   6 -
>  arch/s390/include/asm/mmu_context.h      |   6 -
>  arch/um/include/asm/mmu_context.h        |   6 -
>  arch/unicore32/include/asm/mmu_context.h |   6 -
>  arch/x86/Kconfig                         |   3 +
>  arch/x86/include/asm/mmu_context.h       |  16 -
>  arch/x86/include/asm/pgtable-3level.h    |  45 +++
>  arch/x86/include/asm/pgtable.h           |  53 ++++
>  arch/x86/include/asm/pgtable_64.h        |  16 +-
>  arch/x86/mm/Makefile                     |   2 +-
>  arch/x86/mm/gup.c                        | 496 -------------------------------
>  include/asm-generic/mm_hooks.h           |   6 -
>  include/asm-generic/pgtable.h            |  25 ++
>  include/linux/mm.h                       |   4 +
>  mm/gup.c                                 | 134 +++++++--
>  15 files changed, 262 insertions(+), 562 deletions(-)
>  delete mode 100644 arch/x86/mm/gup.c

It fails to build on x86-64 defconfig/allyesconfig/allmodconfig:

  mm/gup.c:1422:15: error: implicit declaration of function a??pgd_devmapa?? [-Werror=implicit-function-declaration]

The PowerPC allnoconfig build broke as well:

/home/mingo/tip/mm/gup.c: In function '__gup_device_huge_pmd':
/home/mingo/tip/mm/gup.c:1319:2: error: implicit declaration of function 'pmd_pfn' [-Werror=implicit-function-declaration]

Please send delta fixes, because I've already done many small readability edits to 
the patches.

Thanks,

	Ingo

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH 5/7] mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()

[Kirill A. Shutemov]

On Thu, Mar 16, 2017 at 06:26:53PM +0300, Kirill A. Shutemov wrote:
> +static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
> +		unsigned long end, struct page **pages, int *nr)
> +{
> +	unsigned long fault_pfn;
> +
> +	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
> +	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
> +}
> +
> +static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
> +		unsigned long end, struct page **pages, int *nr)
> +{
> +	unsigned long fault_pfn;
> +
> +	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
> +	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
> +}
> +

PowerPC doesn't [always] provide pmd_pfn() and pud_pfn().

Fixup:

diff --git a/mm/gup.c b/mm/gup.c
index 5cc489d98562..6f36cbc294cf 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1287,6 +1287,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 }
 #endif /* __HAVE_ARCH_PTE_SPECIAL */
 
+#ifdef __HAVE_ARCH_PTE_DEVMAP
 static int __gup_device_huge(unsigned long pfn, unsigned long addr,
 		unsigned long end, struct page **pages, int *nr)
 {
@@ -1328,6 +1329,21 @@ static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
 	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
 	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
 }
+#else
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+#endif
 
 static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, int write, struct page **pages, int *nr)
@@ -1338,7 +1354,6 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	if (!pmd_access_permitted(orig, write))
 		return 0;
 
-	VM_BUG_ON(!pfn_valid(pmd_pfn(orig)));
 	if (pmd_devmap(orig))
 		return __gup_device_huge_pmd(orig, addr, end, pages, nr);
 
@@ -1378,7 +1393,6 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	if (!pud_access_permitted(orig, write))
 		return 0;
 
-	VM_BUG_ON(!pfn_valid(pud_pfn(orig)));
 	if (pud_devmap(orig))
 		return __gup_device_huge_pud(orig, addr, end, pages, nr);
 
-- 
 Kirill A. Shutemov

Re: [PATCH 5/7] mm/gup: Implement dev_pagemap logic in generic get_user_pages_fast()

[Kirill A. Shutemov]

On Thu, Mar 16, 2017 at 06:26:53PM +0300, Kirill A. Shutemov wrote:
> +static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
> +		unsigned long end, struct page **pages, int *nr)
> +{
> +	unsigned long fault_pfn;
> +
> +	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
> +	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
> +}
> +
> +static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
> +		unsigned long end, struct page **pages, int *nr)
> +{
> +	unsigned long fault_pfn;
> +
> +	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
> +	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
> +}
> +

PowerPC doesn't [always] provide pmd_pfn() and pud_pfn().

Fixup:

diff --git a/mm/gup.c b/mm/gup.c
index 5cc489d98562..6f36cbc294cf 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1287,6 +1287,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 }
 #endif /* __HAVE_ARCH_PTE_SPECIAL */
 
+#ifdef __HAVE_ARCH_PTE_DEVMAP
 static int __gup_device_huge(unsigned long pfn, unsigned long addr,
 		unsigned long end, struct page **pages, int *nr)
 {
@@ -1328,6 +1329,21 @@ static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
 	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
 	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
 }
+#else
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+#endif
 
 static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, int write, struct page **pages, int *nr)
@@ -1338,7 +1354,6 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	if (!pmd_access_permitted(orig, write))
 		return 0;
 
-	VM_BUG_ON(!pfn_valid(pmd_pfn(orig)));
 	if (pmd_devmap(orig))
 		return __gup_device_huge_pmd(orig, addr, end, pages, nr);
 
@@ -1378,7 +1393,6 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	if (!pud_access_permitted(orig, write))
 		return 0;
 
-	VM_BUG_ON(!pfn_valid(pud_pfn(orig)));
 	if (pud_devmap(orig))
 		return __gup_device_huge_pud(orig, addr, end, pages, nr);
 
-- 
 Kirill A. Shutemov

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCHv2 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

On Fri, Mar 17, 2017 at 12:39:06AM +0300, Kirill A. Shutemov wrote:
> The patch provides all required hooks to match generic
> get_user_pages_fast() behaviour to x86 and switch x86 over.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> ---
>  v2:
>     - Fix build on non-PAE 32-bit x86;

Fixup for allmodconfig on x86-64:

diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index e55fe9475979..e1a0a1fb5971 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -230,13 +230,13 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+#endif
 
 static inline int pgd_devmap(pgd_t pgd)
 {
 	return 0;
 }
 #endif
-#endif
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 static inline pte_t pte_set_flags(pte_t pte, pteval_t set)
-- 
 Kirill A. Shutemov

Re: [PATCHv2 7/7] x86/mm: Switch to generic get_user_page_fast() implementation

[Kirill A. Shutemov]

On Fri, Mar 17, 2017 at 12:39:06AM +0300, Kirill A. Shutemov wrote:
> The patch provides all required hooks to match generic
> get_user_pages_fast() behaviour to x86 and switch x86 over.
> 
> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
> ---
>  v2:
>     - Fix build on non-PAE 32-bit x86;

Fixup for allmodconfig on x86-64:

diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index e55fe9475979..e1a0a1fb5971 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -230,13 +230,13 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+#endif
 
 static inline int pgd_devmap(pgd_t pgd)
 {
 	return 0;
 }
 #endif
-#endif
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
 static inline pte_t pte_set_flags(pte_t pte, pteval_t set)
-- 
 Kirill A. Shutemov

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[tip:x86/mm] mm/gup: Drop the arch_pte_access_permitted() MMU callback

[tip-bot for Kirill A. Shutemov]

Commit-ID:  9a804fecee232e71b47ac37d62fd3d5d66b08b91
Gitweb:     http://git.kernel.org/tip/9a804fecee232e71b47ac37d62fd3d5d66b08b91
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:49 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:01 +0100

mm/gup: Drop the arch_pte_access_permitted() MMU callback

The only arch that defines it to something meaningful is x86.
But x86 doesn't use the generic GUP_fast() implementation -- the
only place where the callback is called.

Let's drop it.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-2-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/powerpc/include/asm/mmu_context.h   | 6 ------
 arch/s390/include/asm/mmu_context.h      | 6 ------
 arch/um/include/asm/mmu_context.h        | 6 ------
 arch/unicore32/include/asm/mmu_context.h | 6 ------
 arch/x86/include/asm/mmu_context.h       | 4 ----
 include/asm-generic/mm_hooks.h           | 6 ------
 mm/gup.c                                 | 3 ---
 7 files changed, 37 deletions(-)

diff --git a/arch/powerpc/include/asm/mmu_context.h b/arch/powerpc/include/asm/mmu_context.h
index b9e3f0a..ecf9885 100644
--- a/arch/powerpc/include/asm/mmu_context.h
+++ b/arch/powerpc/include/asm/mmu_context.h
@@ -163,11 +163,5 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __KERNEL__ */
 #endif /* __ASM_POWERPC_MMU_CONTEXT_H */
diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 6e31d87..fa2bf69 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -156,10 +156,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif /* __S390_MMU_CONTEXT_H */
diff --git a/arch/um/include/asm/mmu_context.h b/arch/um/include/asm/mmu_context.h
index 94ac273..b668e35 100644
--- a/arch/um/include/asm/mmu_context.h
+++ b/arch/um/include/asm/mmu_context.h
@@ -37,12 +37,6 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return true;
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
-
 /*
  * end asm-generic/mm_hooks.h functions
  */
diff --git a/arch/unicore32/include/asm/mmu_context.h b/arch/unicore32/include/asm/mmu_context.h
index 62dfc64..59b06b4 100644
--- a/arch/unicore32/include/asm/mmu_context.h
+++ b/arch/unicore32/include/asm/mmu_context.h
@@ -103,10 +103,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 306c7e1..68b329d 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -268,8 +268,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	return __pkru_allows_pkey(vma_pkey(vma), write);
 }
 
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	return __pkru_allows_pkey(pte_flags_pkey(pte_flags(pte)), write);
-}
 #endif /* _ASM_X86_MMU_CONTEXT_H */
diff --git a/include/asm-generic/mm_hooks.h b/include/asm-generic/mm_hooks.h
index cc5d9a1..41e5b67 100644
--- a/include/asm-generic/mm_hooks.h
+++ b/include/asm-generic/mm_hooks.h
@@ -32,10 +32,4 @@ static inline bool arch_vma_access_permitted(struct vm_area_struct *vma,
 	/* by default, allow everything */
 	return true;
 }
-
-static inline bool arch_pte_access_permitted(pte_t pte, bool write)
-{
-	/* by default, allow everything */
-	return true;
-}
 #endif	/* _ASM_GENERIC_MM_HOOKS_H */
diff --git a/mm/gup.c b/mm/gup.c
index 04aa405..3f2338b 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1216,9 +1216,6 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			pte_protnone(pte) || (write && !pte_write(pte)))
 			goto pte_unmap;
 
-		if (!arch_pte_access_permitted(pte, write))
-			goto pte_unmap;
-
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
 		page = pte_page(pte);
 		head = compound_head(page);

[tip:x86/mm] mm/gup: Move permission checks into helpers

[tip-bot for Kirill A. Shutemov]

Commit-ID:  e7884f8ead4a301b04687a3238527b06feef8ea0
Gitweb:     http://git.kernel.org/tip/e7884f8ead4a301b04687a3238527b06feef8ea0
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:50 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:01 +0100

mm/gup: Move permission checks into helpers

This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

On x86, we would need to do additional permission checks to determine if
access is allowed.

Let's abstract it out into separate helpers.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-3-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/asm-generic/pgtable.h | 25 +++++++++++++++++++++++++
 mm/gup.c                      | 15 ++++++++++-----
 2 files changed, 35 insertions(+), 5 deletions(-)

diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
index 1fad160..7dfa767 100644
--- a/include/asm-generic/pgtable.h
+++ b/include/asm-generic/pgtable.h
@@ -341,6 +341,31 @@ static inline int pte_unused(pte_t pte)
 }
 #endif
 
+#ifndef pte_access_permitted
+#define pte_access_permitted(pte, write) \
+	(pte_present(pte) && (!(write) || pte_write(pte)))
+#endif
+
+#ifndef pmd_access_permitted
+#define pmd_access_permitted(pmd, write) \
+	(pmd_present(pmd) && (!(write) || pmd_write(pmd)))
+#endif
+
+#ifndef pud_access_permitted
+#define pud_access_permitted(pud, write) \
+	(pud_present(pud) && (!(write) || pud_write(pud)))
+#endif
+
+#ifndef p4d_access_permitted
+#define p4d_access_permitted(p4d, write) \
+	(p4d_present(p4d) && (!(write) || p4d_write(p4d)))
+#endif
+
+#ifndef pgd_access_permitted
+#define pgd_access_permitted(pgd, write) \
+	(pgd_present(pgd) && (!(write) || pgd_write(pgd)))
+#endif
+
 #ifndef __HAVE_ARCH_PMD_SAME
 #ifdef CONFIG_TRANSPARENT_HUGEPAGE
 static inline int pmd_same(pmd_t pmd_a, pmd_t pmd_b)
diff --git a/mm/gup.c b/mm/gup.c
index 3f2338b..a62a778c 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1212,8 +1212,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		 * Similar to the PMD case below, NUMA hinting must take slow
 		 * path using the pte_protnone check.
 		 */
-		if (!pte_present(pte) || pte_special(pte) ||
-			pte_protnone(pte) || (write && !pte_write(pte)))
+		if (pte_protnone(pte))
+			goto pte_unmap;
+
+		if (!pte_access_permitted(pte, write))
+			goto pte_unmap;
+
+		if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1264,7 +1269,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pmd_write(orig))
+	if (!pmd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1299,7 +1304,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	struct page *head, *page;
 	int refs;
 
-	if (write && !pud_write(orig))
+	if (!pud_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;
@@ -1335,7 +1340,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	int refs;
 	struct page *head, *page;
 
-	if (write && !pgd_write(orig))
+	if (!pgd_access_permitted(orig, write))
 		return 0;
 
 	refs = 0;

[tip:x86/mm] mm/gup: Move page table entry dereference into helper function

[tip-bot for Kirill A. Shutemov]

Commit-ID:  0005d20b2ff1e501e186b3b3bc587085ac305fdc
Gitweb:     http://git.kernel.org/tip/0005d20b2ff1e501e186b3b3bc587085ac305fdc
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:51 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:02 +0100

mm/gup: Move page table entry dereference into helper function

This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

On x86 PAE, page table entry is larger than sizeof(long) and we would
need to provide a helper that can read the entry atomically.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-4-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 mm/gup.c | 20 ++++++++++++--------
 1 file changed, 12 insertions(+), 8 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index a62a778c..e83db38 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1189,6 +1189,17 @@ struct page *get_dump_page(unsigned long addr)
  */
 #ifdef CONFIG_HAVE_GENERIC_RCU_GUP
 
+#ifndef gup_get_pte
+/*
+ * We assume that the PTE can be read atomically. If this is not the case for
+ * your architecture, please provide the helper.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	return READ_ONCE(*ptep);
+}
+#endif
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
@@ -1198,14 +1209,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
-		/*
-		 * In the line below we are assuming that the pte can be read
-		 * atomically. If this is not the case for your architecture,
-		 * please wrap this in a helper function!
-		 *
-		 * for an example see gup_get_pte in arch/x86/mm/gup.c
-		 */
-		pte_t pte = READ_ONCE(*ptep);
+		pte_t pte = gup_get_pte(ptep);
 		struct page *head, *page;
 
 		/*

[tip:x86/mm] mm/gup: Mark all pages PageReferenced in generic get_user_pages_fast()

[tip-bot for Kirill A. Shutemov]

Commit-ID:  e93480537fd7ecaf5ed1a662a979376f6fee50e3
Gitweb:     http://git.kernel.org/tip/e93480537fd7ecaf5ed1a662a979376f6fee50e3
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:52 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:02 +0100

mm/gup: Mark all pages PageReferenced in generic get_user_pages_fast()

This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

Unlike generic GUP_fast(), the x86 version makes all pages it touches
referenced. It seems required for GRU and EPT.

See the following commit:

  8ee53820edfd ("thp: mmu_notifier_test_young")

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-5-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 mm/gup.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/mm/gup.c b/mm/gup.c
index e83db38..2b6cd357 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1238,6 +1238,8 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		}
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
+
+		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
 
@@ -1299,6 +1301,7 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1334,6 +1337,7 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 
@@ -1370,6 +1374,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 		return 0;
 	}
 
+	SetPageReferenced(head);
 	return 1;
 }
 

[tip:x86/mm] mm/gup: Implement the dev_pagemap() logic in the generic get_user_pages_fast() function

[tip-bot for Kirill A. Shutemov]

Commit-ID:  b59f65fa076a8eac2ff3a8ab7f8e1705b9fa86cb
Gitweb:     http://git.kernel.org/tip/b59f65fa076a8eac2ff3a8ab7f8e1705b9fa86cb
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:53 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:02 +0100

mm/gup: Implement the dev_pagemap() logic in the generic get_user_pages_fast() function

This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

Prepare generic GUP_fast() to handle dev_pagemap(). At the moment, it's
only implemented on x86. On non-x86, the new code will be compiled out.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-6-kirill.shutemov@linux.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 include/linux/mm.h |  4 +++
 mm/gup.c           | 87 ++++++++++++++++++++++++++++++++++++++++++++++++++++--
 2 files changed, 89 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 5f01c88..e197d3c 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -430,6 +430,10 @@ static inline int pud_devmap(pud_t pud)
 {
 	return 0;
 }
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 
 /*
diff --git a/mm/gup.c b/mm/gup.c
index 2b6cd357..e3d1e80 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1200,12 +1200,23 @@ static inline pte_t gup_get_pte(pte_t *ptep)
 }
 #endif
 
+static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
+{
+	while ((*nr) - nr_start) {
+		struct page *page = pages[--(*nr)];
+
+		ClearPageReferenced(page);
+		put_page(page);
+	}
+}
+
 #ifdef __HAVE_ARCH_PTE_SPECIAL
 static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 			 int write, struct page **pages, int *nr)
 {
+	struct dev_pagemap *pgmap = NULL;
+	int nr_start = *nr, ret = 0;
 	pte_t *ptep, *ptem;
-	int ret = 0;
 
 	ptem = ptep = pte_offset_map(&pmd, addr);
 	do {
@@ -1222,7 +1233,13 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 		if (!pte_access_permitted(pte, write))
 			goto pte_unmap;
 
-		if (pte_special(pte))
+		if (pte_devmap(pte)) {
+			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
+			if (unlikely(!pgmap)) {
+				undo_dev_pagemap(nr, nr_start, pages);
+				goto pte_unmap;
+			}
+		} else if (pte_special(pte))
 			goto pte_unmap;
 
 		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
@@ -1239,6 +1256,7 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 
 		VM_BUG_ON_PAGE(compound_head(page) != head, page);
 
+		put_dev_pagemap(pgmap);
 		SetPageReferenced(page);
 		pages[*nr] = page;
 		(*nr)++;
@@ -1269,6 +1287,64 @@ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end,
 }
 #endif /* __HAVE_ARCH_PTE_SPECIAL */
 
+#ifdef __HAVE_ARCH_PTE_DEVMAP
+static int __gup_device_huge(unsigned long pfn, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	int nr_start = *nr;
+	struct dev_pagemap *pgmap = NULL;
+
+	do {
+		struct page *page = pfn_to_page(pfn);
+
+		pgmap = get_dev_pagemap(pfn, pgmap);
+		if (unlikely(!pgmap)) {
+			undo_dev_pagemap(nr, nr_start, pages);
+			return 0;
+		}
+		SetPageReferenced(page);
+		pages[*nr] = page;
+		get_page(page);
+		put_dev_pagemap(pgmap);
+		(*nr)++;
+		pfn++;
+	} while (addr += PAGE_SIZE, addr != end);
+	return 1;
+}
+
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	unsigned long fault_pfn;
+
+	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
+	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
+}
+#else
+static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+
+static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
+		unsigned long end, struct page **pages, int *nr)
+{
+	BUILD_BUG();
+	return 0;
+}
+#endif
+
 static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 		unsigned long end, int write, struct page **pages, int *nr)
 {
@@ -1278,6 +1354,9 @@ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr,
 	if (!pmd_access_permitted(orig, write))
 		return 0;
 
+	if (pmd_devmap(orig))
+		return __gup_device_huge_pmd(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pmd_page(orig);
 	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
@@ -1314,6 +1393,9 @@ static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr,
 	if (!pud_access_permitted(orig, write))
 		return 0;
 
+	if (pud_devmap(orig))
+		return __gup_device_huge_pud(orig, addr, end, pages, nr);
+
 	refs = 0;
 	head = pud_page(orig);
 	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
@@ -1351,6 +1433,7 @@ static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr,
 	if (!pgd_access_permitted(orig, write))
 		return 0;
 
+	BUILD_BUG_ON(pgd_devmap(orig));
 	refs = 0;
 	head = pgd_page(orig);
 	page = head + ((addr & ~PGDIR_MASK) >> PAGE_SHIFT);

[tip:x86/mm] mm/gup: Provide callback to check if __GUP_fast() is allowed for the range

[tip-bot for Kirill A. Shutemov]

Commit-ID:  73e10a61817dfc97fe7418bfad1f608e562d7348
Gitweb:     http://git.kernel.org/tip/73e10a61817dfc97fe7418bfad1f608e562d7348
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Thu, 16 Mar 2017 18:26:54 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:03 +0100

mm/gup: Provide callback to check if __GUP_fast() is allowed for the range

This is a preparation patch for the transition of x86 to the generic GUP_fast()
implementation.

On x86, get_user_pages_fast() does a couple of sanity checks to see if we can
call __get_user_pages_fast() for the range.

This kind of wrapping protection should be useful for the generic code too.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316152655.37789-7-kirill.shutemov@linux.intel.com
[ Small readability edits. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 mm/gup.c | 24 +++++++++++++++++++++---
 1 file changed, 21 insertions(+), 3 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index e3d1e80..527ec2c 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1614,6 +1614,21 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return nr;
 }
 
+#ifndef gup_fast_permitted
+/*
+ * Check if it's allowed to use __get_user_pages_fast() for the range, or
+ * we need to fall back to the slow version:
+ */
+bool gup_fast_permitted(unsigned long start, int nr_pages, int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long) nr_pages << PAGE_SHIFT;
+	end = start + len;
+	return end >= start;
+}
+#endif
+
 /**
  * get_user_pages_fast() - pin user pages in memory
  * @start:	starting user address
@@ -1633,11 +1648,14 @@ int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
 int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 			struct page **pages)
 {
-	int nr, ret;
+	int nr = 0, ret = 0;
 
 	start &= PAGE_MASK;
-	nr = __get_user_pages_fast(start, nr_pages, write, pages);
-	ret = nr;
+
+	if (gup_fast_permitted(start, nr_pages, write)) {
+		nr = __get_user_pages_fast(start, nr_pages, write, pages);
+		ret = nr;
+	}
 
 	if (nr < nr_pages) {
 		/* Try to get the remaining pages with get_user_pages */

[tip:x86/mm] x86/mm/gup: Switch GUP to the generic get_user_page_fast() implementation

[tip-bot for Kirill A. Shutemov]

Commit-ID:  2947ba054a4dabbd82848728d765346886050029
Gitweb:     http://git.kernel.org/tip/2947ba054a4dabbd82848728d765346886050029
Author:     Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
AuthorDate: Fri, 17 Mar 2017 00:39:06 +0300
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 18 Mar 2017 09:48:03 +0100

x86/mm/gup: Switch GUP to the generic get_user_page_fast() implementation

This patch provides all required callbacks required by the generic
get_user_pages_fast() code and switches x86 over - and removes
the platform specific implementation.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K . V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Dann Frazier <dann.frazier@canonical.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Steve Capper <steve.capper@linaro.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-arch@vger.kernel.org
Cc: linux-mm@kvack.org
Link: http://lkml.kernel.org/r/20170316213906.89528-1-kirill.shutemov@linux.intel.com
[ Minor readability edits. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/arm/Kconfig                      |   2 +-
 arch/arm64/Kconfig                    |   2 +-
 arch/powerpc/Kconfig                  |   2 +-
 arch/x86/Kconfig                      |   3 +
 arch/x86/include/asm/mmu_context.h    |  12 -
 arch/x86/include/asm/pgtable-3level.h |  47 ++++
 arch/x86/include/asm/pgtable.h        |  53 ++++
 arch/x86/include/asm/pgtable_64.h     |  16 +-
 arch/x86/mm/Makefile                  |   2 +-
 arch/x86/mm/gup.c                     | 496 ----------------------------------
 mm/Kconfig                            |   2 +-
 mm/gup.c                              |  10 +-
 12 files changed, 128 insertions(+), 519 deletions(-)

diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 0d4e71b..454fadd 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -1666,7 +1666,7 @@ config ARCH_SELECT_MEMORY_MODEL
 config HAVE_ARCH_PFN_VALID
 	def_bool ARCH_HAS_HOLES_MEMORYMODEL || !SPARSEMEM
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 	depends on ARM_LPAE
 
diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
index 3741859..af62bf7 100644
--- a/arch/arm64/Kconfig
+++ b/arch/arm64/Kconfig
@@ -205,7 +205,7 @@ config GENERIC_CALIBRATE_DELAY
 config ZONE_DMA
 	def_bool y
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	def_bool y
 
 config ARCH_DMA_ADDR_T_64BIT
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index 97a8bc8..3a716b2 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -135,7 +135,7 @@ config PPC
 	select HAVE_FUNCTION_GRAPH_TRACER
 	select HAVE_FUNCTION_TRACER
 	select HAVE_GCC_PLUGINS
-	select HAVE_GENERIC_RCU_GUP
+	select HAVE_GENERIC_GUP
 	select HAVE_HW_BREAKPOINT		if PERF_EVENTS && (PPC_BOOK3S || PPC_8xx)
 	select HAVE_IDE
 	select HAVE_IOREMAP_PROT
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 2bab9d0..8977d9c 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2788,6 +2788,9 @@ config X86_DMA_REMAP
 	bool
 	depends on STA2X11
 
+config HAVE_GENERIC_GUP
+	def_bool y
+
 source "net/Kconfig"
 
 source "drivers/Kconfig"
diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h
index 68b329d..6e933d2 100644
--- a/arch/x86/include/asm/mmu_context.h
+++ b/arch/x86/include/asm/mmu_context.h
@@ -220,18 +220,6 @@ static inline int vma_pkey(struct vm_area_struct *vma)
 }
 #endif
 
-static inline bool __pkru_allows_pkey(u16 pkey, bool write)
-{
-	u32 pkru = read_pkru();
-
-	if (!__pkru_allows_read(pkru, pkey))
-		return false;
-	if (write && !__pkru_allows_write(pkru, pkey))
-		return false;
-
-	return true;
-}
-
 /*
  * We only want to enforce protection keys on the current process
  * because we effectively have no access to PKRU for other
diff --git a/arch/x86/include/asm/pgtable-3level.h b/arch/x86/include/asm/pgtable-3level.h
index 72277b1..29eb577 100644
--- a/arch/x86/include/asm/pgtable-3level.h
+++ b/arch/x86/include/asm/pgtable-3level.h
@@ -215,4 +215,51 @@ static inline pud_t native_pudp_get_and_clear(pud_t *pudp)
 #define __pte_to_swp_entry(pte)		((swp_entry_t){ (pte).pte_high })
 #define __swp_entry_to_pte(x)		((pte_t){ { .pte_high = (x).val } })
 
+#define gup_get_pte gup_get_pte
+/*
+ * WARNING: only to be used in the get_user_pages_fast() implementation.
+ *
+ * With get_user_pages_fast(), we walk down the pagetables without taking
+ * any locks.  For this we would like to load the pointers atomically,
+ * but that is not possible (without expensive cmpxchg8b) on PAE.  What
+ * we do have is the guarantee that a PTE will only either go from not
+ * present to present, or present to not present or both -- it will not
+ * switch to a completely different present page without a TLB flush in
+ * between; something that we are blocking by holding interrupts off.
+ *
+ * Setting ptes from not present to present goes:
+ *
+ *   ptep->pte_high = h;
+ *   smp_wmb();
+ *   ptep->pte_low = l;
+ *
+ * And present to not present goes:
+ *
+ *   ptep->pte_low = 0;
+ *   smp_wmb();
+ *   ptep->pte_high = 0;
+ *
+ * We must ensure here that the load of pte_low sees 'l' iff pte_high
+ * sees 'h'. We load pte_high *after* loading pte_low, which ensures we
+ * don't see an older value of pte_high.  *Then* we recheck pte_low,
+ * which ensures that we haven't picked up a changed pte high. We might
+ * have gotten rubbish values from pte_low and pte_high, but we are
+ * guaranteed that pte_low will not have the present bit set *unless*
+ * it is 'l'. Because get_user_pages_fast() only operates on present ptes
+ * we're safe.
+ */
+static inline pte_t gup_get_pte(pte_t *ptep)
+{
+	pte_t pte;
+
+	do {
+		pte.pte_low = ptep->pte_low;
+		smp_rmb();
+		pte.pte_high = ptep->pte_high;
+		smp_rmb();
+	} while (unlikely(pte.pte_low != ptep->pte_low));
+
+	return pte;
+}
+
 #endif /* _ASM_X86_PGTABLE_3LEVEL_H */
diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h
index 6f6f351..160256b 100644
--- a/arch/x86/include/asm/pgtable.h
+++ b/arch/x86/include/asm/pgtable.h
@@ -233,6 +233,11 @@ static inline int pud_devmap(pud_t pud)
 	return 0;
 }
 #endif
+
+static inline int pgd_devmap(pgd_t pgd)
+{
+	return 0;
+}
 #endif
 #endif /* CONFIG_TRANSPARENT_HUGEPAGE */
 
@@ -1136,6 +1141,54 @@ static inline u16 pte_flags_pkey(unsigned long pte_flags)
 #endif
 }
 
+static inline bool __pkru_allows_pkey(u16 pkey, bool write)
+{
+	u32 pkru = read_pkru();
+
+	if (!__pkru_allows_read(pkru, pkey))
+		return false;
+	if (write && !__pkru_allows_write(pkru, pkey))
+		return false;
+
+	return true;
+}
+
+/*
+ * 'pteval' can come from a PTE, PMD or PUD.  We only check
+ * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
+ * same value on all 3 types.
+ */
+static inline bool __pte_access_permitted(unsigned long pteval, bool write)
+{
+	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
+
+	if (write)
+		need_pte_bits |= _PAGE_RW;
+
+	if ((pteval & need_pte_bits) != need_pte_bits)
+		return 0;
+
+	return __pkru_allows_pkey(pte_flags_pkey(pteval), write);
+}
+
+#define pte_access_permitted pte_access_permitted
+static inline bool pte_access_permitted(pte_t pte, bool write)
+{
+	return __pte_access_permitted(pte_val(pte), write);
+}
+
+#define pmd_access_permitted pmd_access_permitted
+static inline bool pmd_access_permitted(pmd_t pmd, bool write)
+{
+	return __pte_access_permitted(pmd_val(pmd), write);
+}
+
+#define pud_access_permitted pud_access_permitted
+static inline bool pud_access_permitted(pud_t pud, bool write)
+{
+	return __pte_access_permitted(pud_val(pud), write);
+}
+
 #include <asm-generic/pgtable.h>
 #endif	/* __ASSEMBLY__ */
 
diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h
index 67608d4..13709cf 100644
--- a/arch/x86/include/asm/pgtable_64.h
+++ b/arch/x86/include/asm/pgtable_64.h
@@ -207,6 +207,20 @@ extern void cleanup_highmap(void);
 extern void init_extra_mapping_uc(unsigned long phys, unsigned long size);
 extern void init_extra_mapping_wb(unsigned long phys, unsigned long size);
 
-#endif /* !__ASSEMBLY__ */
+#define gup_fast_permitted gup_fast_permitted
+static inline bool gup_fast_permitted(unsigned long start, int nr_pages,
+		int write)
+{
+	unsigned long len, end;
+
+	len = (unsigned long)nr_pages << PAGE_SHIFT;
+	end = start + len;
+	if (end < start)
+		return false;
+	if (end >> __VIRTUAL_MASK_SHIFT)
+		return false;
+	return true;
+}
 
+#endif /* !__ASSEMBLY__ */
 #endif /* _ASM_X86_PGTABLE_64_H */
diff --git a/arch/x86/mm/Makefile b/arch/x86/mm/Makefile
index 96d2b84..0fbdcb6 100644
--- a/arch/x86/mm/Makefile
+++ b/arch/x86/mm/Makefile
@@ -2,7 +2,7 @@
 KCOV_INSTRUMENT_tlb.o	:= n
 
 obj-y	:=  init.o init_$(BITS).o fault.o ioremap.o extable.o pageattr.o mmap.o \
-	    pat.o pgtable.o physaddr.o gup.o setup_nx.o tlb.o
+	    pat.o pgtable.o physaddr.o setup_nx.o tlb.o
 
 # Make sure __phys_addr has no stackprotector
 nostackp := $(call cc-option, -fno-stack-protector)
diff --git a/arch/x86/mm/gup.c b/arch/x86/mm/gup.c
deleted file mode 100644
index 456dfdf..0000000
--- a/arch/x86/mm/gup.c
+++ /dev/null
@@ -1,496 +0,0 @@
-/*
- * Lockless get_user_pages_fast for x86
- *
- * Copyright (C) 2008 Nick Piggin
- * Copyright (C) 2008 Novell Inc.
- */
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/vmstat.h>
-#include <linux/highmem.h>
-#include <linux/swap.h>
-#include <linux/memremap.h>
-
-#include <asm/mmu_context.h>
-#include <asm/pgtable.h>
-
-static inline pte_t gup_get_pte(pte_t *ptep)
-{
-#ifndef CONFIG_X86_PAE
-	return READ_ONCE(*ptep);
-#else
-	/*
-	 * With get_user_pages_fast, we walk down the pagetables without taking
-	 * any locks.  For this we would like to load the pointers atomically,
-	 * but that is not possible (without expensive cmpxchg8b) on PAE.  What
-	 * we do have is the guarantee that a pte will only either go from not
-	 * present to present, or present to not present or both -- it will not
-	 * switch to a completely different present page without a TLB flush in
-	 * between; something that we are blocking by holding interrupts off.
-	 *
-	 * Setting ptes from not present to present goes:
-	 * ptep->pte_high = h;
-	 * smp_wmb();
-	 * ptep->pte_low = l;
-	 *
-	 * And present to not present goes:
-	 * ptep->pte_low = 0;
-	 * smp_wmb();
-	 * ptep->pte_high = 0;
-	 *
-	 * We must ensure here that the load of pte_low sees l iff pte_high
-	 * sees h. We load pte_high *after* loading pte_low, which ensures we
-	 * don't see an older value of pte_high.  *Then* we recheck pte_low,
-	 * which ensures that we haven't picked up a changed pte high. We might
-	 * have got rubbish values from pte_low and pte_high, but we are
-	 * guaranteed that pte_low will not have the present bit set *unless*
-	 * it is 'l'. And get_user_pages_fast only operates on present ptes, so
-	 * we're safe.
-	 *
-	 * gup_get_pte should not be used or copied outside gup.c without being
-	 * very careful -- it does not atomically load the pte or anything that
-	 * is likely to be useful for you.
-	 */
-	pte_t pte;
-
-retry:
-	pte.pte_low = ptep->pte_low;
-	smp_rmb();
-	pte.pte_high = ptep->pte_high;
-	smp_rmb();
-	if (unlikely(pte.pte_low != ptep->pte_low))
-		goto retry;
-
-	return pte;
-#endif
-}
-
-static void undo_dev_pagemap(int *nr, int nr_start, struct page **pages)
-{
-	while ((*nr) - nr_start) {
-		struct page *page = pages[--(*nr)];
-
-		ClearPageReferenced(page);
-		put_page(page);
-	}
-}
-
-/*
- * 'pteval' can come from a pte, pmd, pud or p4d.  We only check
- * _PAGE_PRESENT, _PAGE_USER, and _PAGE_RW in here which are the
- * same value on all 4 types.
- */
-static inline int pte_allows_gup(unsigned long pteval, int write)
-{
-	unsigned long need_pte_bits = _PAGE_PRESENT|_PAGE_USER;
-
-	if (write)
-		need_pte_bits |= _PAGE_RW;
-
-	if ((pteval & need_pte_bits) != need_pte_bits)
-		return 0;
-
-	/* Check memory protection keys permissions. */
-	if (!__pkru_allows_pkey(pte_flags_pkey(pteval), write))
-		return 0;
-
-	return 1;
-}
-
-/*
- * The performance critical leaf functions are made noinline otherwise gcc
- * inlines everything into a single function which results in too much
- * register pressure.
- */
-static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct dev_pagemap *pgmap = NULL;
-	int nr_start = *nr, ret = 0;
-	pte_t *ptep, *ptem;
-
-	/*
-	 * Keep the original mapped PTE value (ptem) around since we
-	 * might increment ptep off the end of the page when finishing
-	 * our loop iteration.
-	 */
-	ptem = ptep = pte_offset_map(&pmd, addr);
-	do {
-		pte_t pte = gup_get_pte(ptep);
-		struct page *page;
-
-		/* Similar to the PMD case, NUMA hinting must take slow path */
-		if (pte_protnone(pte))
-			break;
-
-		if (!pte_allows_gup(pte_val(pte), write))
-			break;
-
-		if (pte_devmap(pte)) {
-			pgmap = get_dev_pagemap(pte_pfn(pte), pgmap);
-			if (unlikely(!pgmap)) {
-				undo_dev_pagemap(nr, nr_start, pages);
-				break;
-			}
-		} else if (pte_special(pte))
-			break;
-
-		VM_BUG_ON(!pfn_valid(pte_pfn(pte)));
-		page = pte_page(pte);
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		(*nr)++;
-
-	} while (ptep++, addr += PAGE_SIZE, addr != end);
-	if (addr == end)
-		ret = 1;
-	pte_unmap(ptem);
-
-	return ret;
-}
-
-static inline void get_head_page_multiple(struct page *page, int nr)
-{
-	VM_BUG_ON_PAGE(page != compound_head(page), page);
-	VM_BUG_ON_PAGE(page_count(page) == 0, page);
-	page_ref_add(page, nr);
-	SetPageReferenced(page);
-}
-
-static int __gup_device_huge(unsigned long pfn, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	int nr_start = *nr;
-	struct dev_pagemap *pgmap = NULL;
-
-	do {
-		struct page *page = pfn_to_page(pfn);
-
-		pgmap = get_dev_pagemap(pfn, pgmap);
-		if (unlikely(!pgmap)) {
-			undo_dev_pagemap(nr, nr_start, pages);
-			return 0;
-		}
-		SetPageReferenced(page);
-		pages[*nr] = page;
-		get_page(page);
-		put_dev_pagemap(pgmap);
-		(*nr)++;
-		pfn++;
-	} while (addr += PAGE_SIZE, addr != end);
-	return 1;
-}
-
-static int __gup_device_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pmd_pfn(pmd) + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static int __gup_device_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, struct page **pages, int *nr)
-{
-	unsigned long fault_pfn;
-
-	fault_pfn = pud_pfn(pud) + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	return __gup_device_huge(fault_pfn, addr, end, pages, nr);
-}
-
-static noinline int gup_huge_pmd(pmd_t pmd, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pmd_val(pmd), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pmd_pfn(pmd)));
-	if (pmd_devmap(pmd))
-		return __gup_device_huge_pmd(pmd, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pmd_flags(pmd) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pmd_page(pmd);
-	page = head + ((addr & ~PMD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pmd_range(pud_t pud, unsigned long addr, unsigned long end,
-		int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pmd_t *pmdp;
-
-	pmdp = pmd_offset(&pud, addr);
-	do {
-		pmd_t pmd = *pmdp;
-
-		next = pmd_addr_end(addr, end);
-		if (pmd_none(pmd))
-			return 0;
-		if (unlikely(pmd_large(pmd) || !pmd_present(pmd))) {
-			/*
-			 * NUMA hinting faults need to be handled in the GUP
-			 * slowpath for accounting purposes and so that they
-			 * can be serialised against THP migration.
-			 */
-			if (pmd_protnone(pmd))
-				return 0;
-			if (!gup_huge_pmd(pmd, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pte_range(pmd, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pmdp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static noinline int gup_huge_pud(pud_t pud, unsigned long addr,
-		unsigned long end, int write, struct page **pages, int *nr)
-{
-	struct page *head, *page;
-	int refs;
-
-	if (!pte_allows_gup(pud_val(pud), write))
-		return 0;
-
-	VM_BUG_ON(!pfn_valid(pud_pfn(pud)));
-	if (pud_devmap(pud))
-		return __gup_device_huge_pud(pud, addr, end, pages, nr);
-
-	/* hugepages are never "special" */
-	VM_BUG_ON(pud_flags(pud) & _PAGE_SPECIAL);
-
-	refs = 0;
-	head = pud_page(pud);
-	page = head + ((addr & ~PUD_MASK) >> PAGE_SHIFT);
-	do {
-		VM_BUG_ON_PAGE(compound_head(page) != head, page);
-		pages[*nr] = page;
-		(*nr)++;
-		page++;
-		refs++;
-	} while (addr += PAGE_SIZE, addr != end);
-	get_head_page_multiple(head, refs);
-
-	return 1;
-}
-
-static int gup_pud_range(p4d_t p4d, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	pud_t *pudp;
-
-	pudp = pud_offset(&p4d, addr);
-	do {
-		pud_t pud = *pudp;
-
-		next = pud_addr_end(addr, end);
-		if (pud_none(pud))
-			return 0;
-		if (unlikely(pud_large(pud))) {
-			if (!gup_huge_pud(pud, addr, next, write, pages, nr))
-				return 0;
-		} else {
-			if (!gup_pmd_range(pud, addr, next, write, pages, nr))
-				return 0;
-		}
-	} while (pudp++, addr = next, addr != end);
-
-	return 1;
-}
-
-static int gup_p4d_range(pgd_t pgd, unsigned long addr, unsigned long end,
-			int write, struct page **pages, int *nr)
-{
-	unsigned long next;
-	p4d_t *p4dp;
-
-	p4dp = p4d_offset(&pgd, addr);
-	do {
-		p4d_t p4d = *p4dp;
-
-		next = p4d_addr_end(addr, end);
-		if (p4d_none(p4d))
-			return 0;
-		BUILD_BUG_ON(p4d_large(p4d));
-		if (!gup_pud_range(p4d, addr, next, write, pages, nr))
-			return 0;
-	} while (p4dp++, addr = next, addr != end);
-
-	return 1;
-}
-
-/*
- * Like get_user_pages_fast() except its IRQ-safe in that it won't fall
- * back to the regular GUP.
- */
-int __get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			  struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	unsigned long flags;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-	end = start + len;
-	if (unlikely(!access_ok(write ? VERIFY_WRITE : VERIFY_READ,
-					(void __user *)start, len)))
-		return 0;
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_save(flags);
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			break;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			break;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_restore(flags);
-
-	return nr;
-}
-
-/**
- * get_user_pages_fast() - pin user pages in memory
- * @start:	starting user address
- * @nr_pages:	number of pages from start to pin
- * @write:	whether pages will be written to
- * @pages:	array that receives pointers to the pages pinned.
- * 		Should be at least nr_pages long.
- *
- * Attempt to pin user pages in memory without taking mm->mmap_sem.
- * If not successful, it will fall back to taking the lock and
- * calling get_user_pages().
- *
- * Returns number of pages pinned. This may be fewer than the number
- * requested. If nr_pages is 0 or negative, returns 0. If no pages
- * were pinned, returns -errno.
- */
-int get_user_pages_fast(unsigned long start, int nr_pages, int write,
-			struct page **pages)
-{
-	struct mm_struct *mm = current->mm;
-	unsigned long addr, len, end;
-	unsigned long next;
-	pgd_t *pgdp;
-	int nr = 0;
-
-	start &= PAGE_MASK;
-	addr = start;
-	len = (unsigned long) nr_pages << PAGE_SHIFT;
-
-	end = start + len;
-	if (end < start)
-		goto slow_irqon;
-
-#ifdef CONFIG_X86_64
-	if (end >> __VIRTUAL_MASK_SHIFT)
-		goto slow_irqon;
-#endif
-
-	/*
-	 * XXX: batch / limit 'nr', to avoid large irq off latency
-	 * needs some instrumenting to determine the common sizes used by
-	 * important workloads (eg. DB2), and whether limiting the batch size
-	 * will decrease performance.
-	 *
-	 * It seems like we're in the clear for the moment. Direct-IO is
-	 * the main guy that batches up lots of get_user_pages, and even
-	 * they are limited to 64-at-a-time which is not so many.
-	 */
-	/*
-	 * This doesn't prevent pagetable teardown, but does prevent
-	 * the pagetables and pages from being freed on x86.
-	 *
-	 * So long as we atomically load page table pointers versus teardown
-	 * (which we do on x86, with the above PAE exception), we can follow the
-	 * address down to the the page and take a ref on it.
-	 */
-	local_irq_disable();
-	pgdp = pgd_offset(mm, addr);
-	do {
-		pgd_t pgd = *pgdp;
-
-		next = pgd_addr_end(addr, end);
-		if (pgd_none(pgd))
-			goto slow;
-		if (!gup_p4d_range(pgd, addr, next, write, pages, &nr))
-			goto slow;
-	} while (pgdp++, addr = next, addr != end);
-	local_irq_enable();
-
-	VM_BUG_ON(nr != (end - start) >> PAGE_SHIFT);
-	return nr;
-
-	{
-		int ret;
-
-slow:
-		local_irq_enable();
-slow_irqon:
-		/* Try to get the remaining pages with get_user_pages */
-		start += nr << PAGE_SHIFT;
-		pages += nr;
-
-		ret = get_user_pages_unlocked(start,
-					      (end - start) >> PAGE_SHIFT,
-					      pages, write ? FOLL_WRITE : 0);
-
-		/* Have to be a bit careful with return values */
-		if (nr > 0) {
-			if (ret < 0)
-				ret = nr;
-			else
-				ret += nr;
-		}
-
-		return ret;
-	}
-}
diff --git a/mm/Kconfig b/mm/Kconfig
index 9b8fccb..c89f472 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -137,7 +137,7 @@ config HAVE_MEMBLOCK_NODE_MAP
 config HAVE_MEMBLOCK_PHYS_MAP
 	bool
 
-config HAVE_GENERIC_RCU_GUP
+config HAVE_GENERIC_GUP
 	bool
 
 config ARCH_DISCARD_MEMBLOCK
diff --git a/mm/gup.c b/mm/gup.c
index 527ec2c..2559a39 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -1155,7 +1155,7 @@ struct page *get_dump_page(unsigned long addr)
 #endif /* CONFIG_ELF_CORE */
 
 /*
- * Generic RCU Fast GUP
+ * Generic Fast GUP
  *
  * get_user_pages_fast attempts to pin user pages by walking the page
  * tables directly and avoids taking locks. Thus the walker needs to be
@@ -1176,8 +1176,8 @@ struct page *get_dump_page(unsigned long addr)
  * Before activating this code, please be aware that the following assumptions
  * are currently made:
  *
- *  *) HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table is used to free
- *      pages containing page tables.
+ *  *) Either HAVE_RCU_TABLE_FREE is enabled, and tlb_remove_table() is used to
+ *  free pages containing page tables or TLB flushing requires IPI broadcast.
  *
  *  *) ptes can be read atomically by the architecture.
  *
@@ -1187,7 +1187,7 @@ struct page *get_dump_page(unsigned long addr)
  *
  * This code is based heavily on the PowerPC implementation by Nick Piggin.
  */
-#ifdef CONFIG_HAVE_GENERIC_RCU_GUP
+#ifdef CONFIG_HAVE_GENERIC_GUP
 
 #ifndef gup_get_pte
 /*
@@ -1677,4 +1677,4 @@ int get_user_pages_fast(unsigned long start, int nr_pages, int write,
 	return ret;
 }
 
-#endif /* CONFIG_HAVE_GENERIC_RCU_GUP */
+#endif /* CONFIG_HAVE_GENERIC_GUP */