[PATCH 0/2] mm, smaps: MMUPageSize for device-dax

[PATCH 0/2] mm, smaps: MMUPageSize for device-dax

[Dan Williams]

Andrew,

Here is another occasion where we want special-case hugetlbfs enabling
to also apply to device-dax. I started to wonder what other hstate
conversions we might do beyond ->split() and ->pagesize(), but this
appears to be the last of the usages of hstate_vma() in
generic/non-hugetlbfs specific code paths.

This is 4.16 material.

---

Dan Williams (2):
      mm, hugetlbfs: introduce ->pagesize() to vm_operations_struct
      device-dax: implement ->pagesize() for smaps to report MMUPageSize


 arch/powerpc/mm/hugetlbpage.c |    5 +----
 drivers/dax/device.c          |   10 ++++++++++
 include/linux/hugetlb.h       |   30 ++++++++++++++++++++++++------
 include/linux/mm.h            |    1 +
 mm/hugetlb.c                  |   38 ++++++++------------------------------
 5 files changed, 44 insertions(+), 40 deletions(-)

--
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 1/2] mm, hugetlbfs: introduce ->pagesize() to vm_operations_struct

[Dan Williams]

When device-dax is operating in huge-page mode we want it to behave like
hugetlbfs and report the MMU page mapping size that is being enforced by
the vma. Similar to commit 31383c6865a5 "mm, hugetlbfs: introduce
->split() to vm_operations_struct" it would be messy to teach
vma_mmu_pagesize() about device-dax page mapping sizes in the same
(hstate) way that hugetlbfs communicates this attribute.  Instead, these
patches introduce a new ->pagesize() vm operation.

Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Reported-by: Jane Chu <jane.chu@oracle.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 arch/powerpc/mm/hugetlbpage.c |    5 +----
 include/linux/hugetlb.h       |   30 ++++++++++++++++++++++++------
 include/linux/mm.h            |    1 +
 mm/hugetlb.c                  |   38 ++++++++------------------------------
 4 files changed, 34 insertions(+), 40 deletions(-)

diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c
index a9b9083c5e49..c6a2e577e842 100644
--- a/arch/powerpc/mm/hugetlbpage.c
+++ b/arch/powerpc/mm/hugetlbpage.c
@@ -568,10 +568,7 @@ unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
 	if (!radix_enabled())
 		return 1UL << mmu_psize_to_shift(psize);
 #endif
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	return huge_page_size(hstate_vma(vma));
+	return vma_kernel_pagesize(vma);
 }
 
 static inline bool is_power_of_4(unsigned long x)
diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
index 82a25880714a..716ccf14ff7b 100644
--- a/include/linux/hugetlb.h
+++ b/include/linux/hugetlb.h
@@ -254,6 +254,30 @@ enum {
 	HUGETLB_ANONHUGE_INODE  = 2,
 };
 
+/*
+ * Return the size of the pages allocated when backing a VMA. In the majority
+ * cases this will be same size as used by the page table entries.
+ */
+static inline unsigned long vma_kernel_pagesize(struct vm_area_struct *vma)
+{
+	if (vma->vm_ops && vma->vm_ops->pagesize)
+		return vma->vm_ops->pagesize(vma);
+	return PAGE_SIZE;
+}
+
+/*
+ * Return the page size being used by the MMU to back a VMA. In the majority
+ * of cases, the page size used by the kernel matches the MMU size. On
+ * architectures where it differs, an architecture-specific version of this
+ * function is required.
+ */
+#ifndef vma_mmu_pagesize
+static inline unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
+{
+	return vma_kernel_pagesize(vma);
+}
+#endif
+
 #ifdef CONFIG_HUGETLBFS
 struct hugetlbfs_sb_info {
 	long	max_inodes;   /* inodes allowed */
@@ -395,10 +419,6 @@ static inline unsigned long huge_page_size(struct hstate *h)
 	return (unsigned long)PAGE_SIZE << h->order;
 }
 
-extern unsigned long vma_kernel_pagesize(struct vm_area_struct *vma);
-
-extern unsigned long vma_mmu_pagesize(struct vm_area_struct *vma);
-
 static inline unsigned long huge_page_mask(struct hstate *h)
 {
 	return h->mask;
@@ -533,8 +553,6 @@ struct hstate {};
 #define page_hstate(page) NULL
 #define huge_page_size(h) PAGE_SIZE
 #define huge_page_mask(h) PAGE_MASK
-#define vma_kernel_pagesize(v) PAGE_SIZE
-#define vma_mmu_pagesize(v) PAGE_SIZE
 #define huge_page_order(h) 0
 #define huge_page_shift(h) PAGE_SHIFT
 static inline bool hstate_is_gigantic(struct hstate *h)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index ea818ff739cd..37b9aef91ec7 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -383,6 +383,7 @@ struct vm_operations_struct {
 	int (*huge_fault)(struct vm_fault *vmf, enum page_entry_size pe_size);
 	void (*map_pages)(struct vm_fault *vmf,
 			pgoff_t start_pgoff, pgoff_t end_pgoff);
+	unsigned long (*pagesize)(struct vm_area_struct * area);
 
 	/* notification that a previously read-only page is about to become
 	 * writable, if an error is returned it will cause a SIGBUS */
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 9a334f5fb730..f45ded1f978a 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -632,36 +632,6 @@ pgoff_t linear_hugepage_index(struct vm_area_struct *vma,
 EXPORT_SYMBOL_GPL(linear_hugepage_index);
 
 /*
- * Return the size of the pages allocated when backing a VMA. In the majority
- * cases this will be same size as used by the page table entries.
- */
-unsigned long vma_kernel_pagesize(struct vm_area_struct *vma)
-{
-	struct hstate *hstate;
-
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	hstate = hstate_vma(vma);
-
-	return 1UL << huge_page_shift(hstate);
-}
-EXPORT_SYMBOL_GPL(vma_kernel_pagesize);
-
-/*
- * Return the page size being used by the MMU to back a VMA. In the majority
- * of cases, the page size used by the kernel matches the MMU size. On
- * architectures where it differs, an architecture-specific version of this
- * function is required.
- */
-#ifndef vma_mmu_pagesize
-unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
-{
-	return vma_kernel_pagesize(vma);
-}
-#endif
-
-/*
  * Flags for MAP_PRIVATE reservations.  These are stored in the bottom
  * bits of the reservation map pointer, which are always clear due to
  * alignment.
@@ -3132,6 +3102,13 @@ static int hugetlb_vm_op_split(struct vm_area_struct *vma, unsigned long addr)
 	return 0;
 }
 
+static unsigned long hugetlb_vm_op_pagesize(struct vm_area_struct *vma)
+{
+	struct hstate *hstate = hstate_vma(vma);
+
+	return 1UL << huge_page_shift(hstate);
+}
+
 /*
  * We cannot handle pagefaults against hugetlb pages at all.  They cause
  * handle_mm_fault() to try to instantiate regular-sized pages in the
@@ -3149,6 +3126,7 @@ const struct vm_operations_struct hugetlb_vm_ops = {
 	.open = hugetlb_vm_op_open,
 	.close = hugetlb_vm_op_close,
 	.split = hugetlb_vm_op_split,
+	.pagesize = hugetlb_vm_op_pagesize,
 };
 
 static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page,

--
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 1/2] mm, hugetlbfs: introduce ->pagesize() to vm_operations_struct

[Michal Hocko]

On Thu 07-12-17 19:30:55, Dan Williams wrote:
> When device-dax is operating in huge-page mode we want it to behave like
> hugetlbfs and report the MMU page mapping size that is being enforced by
> the vma. Similar to commit 31383c6865a5 "mm, hugetlbfs: introduce
> ->split() to vm_operations_struct" it would be messy to teach
> vma_mmu_pagesize() about device-dax page mapping sizes in the same
> (hstate) way that hugetlbfs communicates this attribute.  Instead, these
> patches introduce a new ->pagesize() vm operation.
> 
> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Cc: Paul Mackerras <paulus@samba.org>
> Cc: Michael Ellerman <mpe@ellerman.id.au>
> Reported-by: Jane Chu <jane.chu@oracle.com>
> Signed-off-by: Dan Williams <dan.j.williams@intel.com>

My build battery choked on the following
In file included from drivers/infiniband/core/umem_odp.c:41:0:
./include/linux/hugetlb.h: In function 'vma_kernel_pagesize':
./include/linux/hugetlb.h:262:32: error: dereferencing pointer to incomplete type
  if (vma->vm_ops && vma->vm_ops->pagesize)
                                ^
./include/linux/hugetlb.h:263:21: error: dereferencing pointer to incomplete type
   return vma->vm_ops->pagesize(vma);

I thought that adding #include <linux/mm.h> into linux/hugetlb.h would
be sufficient but then it failed for powerpc defconfig which overrides
vma_kernel_pagesize
In file included from ./include/linux/hugetlb.h:452:0,
                 from arch/powerpc/mm/hugetlbpage.c:14:
./arch/powerpc/include/asm/hugetlb.h:131:26: error: redefinition of 'vma_mmu_pagesize'
 #define vma_mmu_pagesize vma_mmu_pagesize
                          ^
arch/powerpc/mm/hugetlbpage.c:563:15: note: in expansion of macro 'vma_mmu_pagesize'
 unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
               ^
In file included from arch/powerpc/mm/hugetlbpage.c:14:0:
./include/linux/hugetlb.h:275:29: note: previous definition of 'vma_mmu_pagesize' was here
 static inline unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)

So it looks this needs something more laborous.
-- 
Michal Hocko
SUSE Labs

--
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 1/2] mm, hugetlbfs: introduce ->pagesize() to vm_operations_struct

[Michael Ellerman]

Michal Hocko <mhocko@kernel.org> writes:

> On Thu 07-12-17 19:30:55, Dan Williams wrote:
>> When device-dax is operating in huge-page mode we want it to behave like
>> hugetlbfs and report the MMU page mapping size that is being enforced by
>> the vma. Similar to commit 31383c6865a5 "mm, hugetlbfs: introduce
>> ->split() to vm_operations_struct" it would be messy to teach
>> vma_mmu_pagesize() about device-dax page mapping sizes in the same
>> (hstate) way that hugetlbfs communicates this attribute.  Instead, these
>> patches introduce a new ->pagesize() vm operation.
>> 
>> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>> Cc: Paul Mackerras <paulus@samba.org>
>> Cc: Michael Ellerman <mpe@ellerman.id.au>
>> Reported-by: Jane Chu <jane.chu@oracle.com>
>> Signed-off-by: Dan Williams <dan.j.williams@intel.com>
>
> My build battery choked on the following
> In file included from drivers/infiniband/core/umem_odp.c:41:0:
> ./include/linux/hugetlb.h: In function 'vma_kernel_pagesize':
> ./include/linux/hugetlb.h:262:32: error: dereferencing pointer to incomplete type
>   if (vma->vm_ops && vma->vm_ops->pagesize)
>                                 ^
> ./include/linux/hugetlb.h:263:21: error: dereferencing pointer to incomplete type
>    return vma->vm_ops->pagesize(vma);
>
> I thought that adding #include <linux/mm.h> into linux/hugetlb.h would
> be sufficient but then it failed for powerpc defconfig which overrides
> vma_kernel_pagesize
> In file included from ./include/linux/hugetlb.h:452:0,
>                  from arch/powerpc/mm/hugetlbpage.c:14:
> ./arch/powerpc/include/asm/hugetlb.h:131:26: error: redefinition of 'vma_mmu_pagesize'
>  #define vma_mmu_pagesize vma_mmu_pagesize
>                           ^
> arch/powerpc/mm/hugetlbpage.c:563:15: note: in expansion of macro 'vma_mmu_pagesize'
>  unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
>                ^
> In file included from arch/powerpc/mm/hugetlbpage.c:14:0:
> ./include/linux/hugetlb.h:275:29: note: previous definition of 'vma_mmu_pagesize' was here
>  static inline unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
>
> So it looks this needs something more laborous.

This builds for me.

cheers


diff --git a/arch/powerpc/include/asm/hugetlb.h b/arch/powerpc/include/asm/hugetlb.h
index 14c9d44f355b..3cc6ca1bdaf2 100644
--- a/arch/powerpc/include/asm/hugetlb.h
+++ b/arch/powerpc/include/asm/hugetlb.h
@@ -123,6 +123,7 @@ void hugetlb_free_pgd_range(struct mmu_gather *tlb, unsigned long addr,
  * to override the version in mm/hugetlb.c
  */
 #define vma_mmu_pagesize vma_mmu_pagesize
+unsigned long vma_mmu_pagesize(struct vm_area_struct *vma);
 
 /*
  * If the arch doesn't supply something else, assume that hugepage
diff --git a/arch/powerpc/mm/hugetlbpage.c b/arch/powerpc/mm/hugetlbpage.c
index a9b9083c5e49..c6a2e577e842 100644
--- a/arch/powerpc/mm/hugetlbpage.c
+++ b/arch/powerpc/mm/hugetlbpage.c
@@ -568,10 +568,7 @@ unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
 	if (!radix_enabled())
 		return 1UL << mmu_psize_to_shift(psize);
 #endif
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	return huge_page_size(hstate_vma(vma));
+	return vma_kernel_pagesize(vma);
 }
 
 static inline bool is_power_of_4(unsigned long x)
diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
index 6e3696c7b35a..fe7b74325856 100644
--- a/include/linux/hugetlb.h
+++ b/include/linux/hugetlb.h
@@ -394,10 +394,6 @@ static inline unsigned long huge_page_size(struct hstate *h)
 	return (unsigned long)PAGE_SIZE << h->order;
 }
 
-extern unsigned long vma_kernel_pagesize(struct vm_area_struct *vma);
-
-extern unsigned long vma_mmu_pagesize(struct vm_area_struct *vma);
-
 static inline unsigned long huge_page_mask(struct hstate *h)
 {
 	return h->mask;
@@ -430,6 +426,30 @@ static inline unsigned int blocks_per_huge_page(struct hstate *h)
 
 #include <asm/hugetlb.h>
 
+/*
+ * Return the size of the pages allocated when backing a VMA. In the majority
+ * cases this will be same size as used by the page table entries.
+ */
+static inline unsigned long vma_kernel_pagesize(struct vm_area_struct *vma)
+{
+	if (vma->vm_ops && vma->vm_ops->pagesize)
+		return vma->vm_ops->pagesize(vma);
+	return PAGE_SIZE;
+}
+
+/*
+ * Return the page size being used by the MMU to back a VMA. In the majority
+ * of cases, the page size used by the kernel matches the MMU size. On
+ * architectures where it differs, an architecture-specific version of this
+ * function is required.
+ */
+#ifndef vma_mmu_pagesize
+static inline unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
+{
+	return vma_kernel_pagesize(vma);
+}
+#endif
+
 #ifndef arch_make_huge_pte
 static inline pte_t arch_make_huge_pte(pte_t entry, struct vm_area_struct *vma,
 				       struct page *page, int writable)
@@ -533,8 +553,6 @@ struct hstate {};
 #define page_hstate(page) NULL
 #define huge_page_size(h) PAGE_SIZE
 #define huge_page_mask(h) PAGE_MASK
-#define vma_kernel_pagesize(v) PAGE_SIZE
-#define vma_mmu_pagesize(v) PAGE_SIZE
 #define huge_page_order(h) 0
 #define huge_page_shift(h) PAGE_SHIFT
 static inline bool hstate_is_gigantic(struct hstate *h)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 7661156552d3..1933499f896d 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -383,6 +383,7 @@ struct vm_operations_struct {
 	int (*huge_fault)(struct vm_fault *vmf, enum page_entry_size pe_size);
 	void (*map_pages)(struct vm_fault *vmf,
 			pgoff_t start_pgoff, pgoff_t end_pgoff);
+	unsigned long (*pagesize)(struct vm_area_struct * area);
 
 	/* notification that a previously read-only page is about to become
 	 * writable, if an error is returned it will cause a SIGBUS */
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 4137fb67cd79..7c1c45bb3d08 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -629,36 +629,6 @@ pgoff_t linear_hugepage_index(struct vm_area_struct *vma,
 }
 EXPORT_SYMBOL_GPL(linear_hugepage_index);
 
-/*
- * Return the size of the pages allocated when backing a VMA. In the majority
- * cases this will be same size as used by the page table entries.
- */
-unsigned long vma_kernel_pagesize(struct vm_area_struct *vma)
-{
-	struct hstate *hstate;
-
-	if (!is_vm_hugetlb_page(vma))
-		return PAGE_SIZE;
-
-	hstate = hstate_vma(vma);
-
-	return 1UL << huge_page_shift(hstate);
-}
-EXPORT_SYMBOL_GPL(vma_kernel_pagesize);
-
-/*
- * Return the page size being used by the MMU to back a VMA. In the majority
- * of cases, the page size used by the kernel matches the MMU size. On
- * architectures where it differs, an architecture-specific version of this
- * function is required.
- */
-#ifndef vma_mmu_pagesize
-unsigned long vma_mmu_pagesize(struct vm_area_struct *vma)
-{
-	return vma_kernel_pagesize(vma);
-}
-#endif
-
 /*
  * Flags for MAP_PRIVATE reservations.  These are stored in the bottom
  * bits of the reservation map pointer, which are always clear due to
@@ -3142,6 +3112,13 @@ static int hugetlb_vm_op_split(struct vm_area_struct *vma, unsigned long addr)
 	return 0;
 }
 
+static unsigned long hugetlb_vm_op_pagesize(struct vm_area_struct *vma)
+{
+	struct hstate *hstate = hstate_vma(vma);
+
+	return 1UL << huge_page_shift(hstate);
+}
+
 /*
  * We cannot handle pagefaults against hugetlb pages at all.  They cause
  * handle_mm_fault() to try to instantiate regular-sized pages in the
@@ -3159,6 +3136,7 @@ const struct vm_operations_struct hugetlb_vm_ops = {
 	.open = hugetlb_vm_op_open,
 	.close = hugetlb_vm_op_close,
 	.split = hugetlb_vm_op_split,
+	.pagesize = hugetlb_vm_op_pagesize,
 };
 
 static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page,

--
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 2/2] device-dax: implement ->pagesize() for smaps to report MMUPageSize

[Dan Williams]

Given that device-dax is making similar page mapping size guarantees as
hugetlbfs, emit the size in smaps and any other kernel path that
requests the mapping size of a vma.

Reported-by: Jane Chu <jane.chu@oracle.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
---
 drivers/dax/device.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/drivers/dax/device.c b/drivers/dax/device.c
index 7b0bf825c4e7..b57cd5a7b0bd 100644
--- a/drivers/dax/device.c
+++ b/drivers/dax/device.c
@@ -439,10 +439,20 @@ static int dev_dax_split(struct vm_area_struct *vma, unsigned long addr)
 	return 0;
 }
 
+static unsigned long dev_dax_pagesize(struct vm_area_struct *vma)
+{
+	struct file *filp = vma->vm_file;
+	struct dev_dax *dev_dax = filp->private_data;
+	struct dax_region *dax_region = dev_dax->region;
+
+	return dax_region->align;
+}
+
 static const struct vm_operations_struct dax_vm_ops = {
 	.fault = dev_dax_fault,
 	.huge_fault = dev_dax_huge_fault,
 	.split = dev_dax_split,
+	.pagesize = dev_dax_pagesize,
 };
 
 static int dax_mmap(struct file *filp, struct vm_area_struct *vma)

--
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>