[PATCH v3 0/2] Do some code cleanups related to mm

[PATCH v3 0/2] Do some code cleanups related to mm

[Qi Zheng]

Hi,

This patch series aims to do some code cleanups related to mm.

This series is based on next-20210827.

Comments and suggestions are welcome.

Thanks,
Qi.

Changelog in v2 -> v3:
 - Move the declaration of pmd_install() to mm/internal.h

Changelog in v1 -> v2:
 - Fix some typo and code style problems.

Qi Zheng (2):
  mm: introduce pmd_install() helper
  mm: remove redundant smp_wmb()

 mm/filemap.c        | 11 ++-----
 mm/internal.h       |  1 +
 mm/memory.c         | 86 ++++++++++++++++++++++++-----------------------------
 mm/sparse-vmemmap.c |  2 +-
 4 files changed, 43 insertions(+), 57 deletions(-)

-- 
2.11.0

[PATCH v3 1/2] mm: introduce pmd_install() helper

[Qi Zheng]

Currently we have three times the same few lines repeated in the
code. Deduplicate them by newly introduced pmd_install() helper.

Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/filemap.c  | 11 ++---------
 mm/internal.h |  1 +
 mm/memory.c   | 34 ++++++++++++++++------------------
 3 files changed, 19 insertions(+), 27 deletions(-)

diff --git a/mm/filemap.c b/mm/filemap.c
index c90b6e4984c9..923cbba1bf37 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -3209,15 +3209,8 @@ static bool filemap_map_pmd(struct vm_fault *vmf, struct page *page)
 	    }
 	}
 
-	if (pmd_none(*vmf->pmd)) {
-		vmf->ptl = pmd_lock(mm, vmf->pmd);
-		if (likely(pmd_none(*vmf->pmd))) {
-			mm_inc_nr_ptes(mm);
-			pmd_populate(mm, vmf->pmd, vmf->prealloc_pte);
-			vmf->prealloc_pte = NULL;
-		}
-		spin_unlock(vmf->ptl);
-	}
+	if (pmd_none(*vmf->pmd))
+		pmd_install(mm, vmf->pmd, &vmf->prealloc_pte);
 
 	/* See comment in handle_pte_fault() */
 	if (pmd_devmap_trans_unstable(vmf->pmd)) {
diff --git a/mm/internal.h b/mm/internal.h
index b1001ebeb286..18256e32a14c 100644
--- a/mm/internal.h
+++ b/mm/internal.h
@@ -47,6 +47,7 @@ bool __folio_end_writeback(struct folio *folio);
 
 void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *start_vma,
 		unsigned long floor, unsigned long ceiling);
+void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte);
 
 static inline bool can_madv_lru_vma(struct vm_area_struct *vma)
 {
diff --git a/mm/memory.c b/mm/memory.c
index 39e7a1495c3c..ef7b1762e996 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -433,9 +433,20 @@ void free_pgtables(struct mmu_gather *tlb, struct vm_area_struct *vma,
 	}
 }
 
+void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte)
+{
+	spinlock_t *ptl = pmd_lock(mm, pmd);
+
+	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
+		mm_inc_nr_ptes(mm);
+		pmd_populate(mm, pmd, *pte);
+		*pte = NULL;
+	}
+	spin_unlock(ptl);
+}
+
 int __pte_alloc(struct mm_struct *mm, pmd_t *pmd)
 {
-	spinlock_t *ptl;
 	pgtable_t new = pte_alloc_one(mm);
 	if (!new)
 		return -ENOMEM;
@@ -455,13 +466,7 @@ int __pte_alloc(struct mm_struct *mm, pmd_t *pmd)
 	 */
 	smp_wmb(); /* Could be smp_wmb__xxx(before|after)_spin_lock */
 
-	ptl = pmd_lock(mm, pmd);
-	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
-		mm_inc_nr_ptes(mm);
-		pmd_populate(mm, pmd, new);
-		new = NULL;
-	}
-	spin_unlock(ptl);
+	pmd_install(mm, pmd, &new);
 	if (new)
 		pte_free(mm, new);
 	return 0;
@@ -4027,17 +4032,10 @@ vm_fault_t finish_fault(struct vm_fault *vmf)
 				return ret;
 		}
 
-		if (vmf->prealloc_pte) {
-			vmf->ptl = pmd_lock(vma->vm_mm, vmf->pmd);
-			if (likely(pmd_none(*vmf->pmd))) {
-				mm_inc_nr_ptes(vma->vm_mm);
-				pmd_populate(vma->vm_mm, vmf->pmd, vmf->prealloc_pte);
-				vmf->prealloc_pte = NULL;
-			}
-			spin_unlock(vmf->ptl);
-		} else if (unlikely(pte_alloc(vma->vm_mm, vmf->pmd))) {
+		if (vmf->prealloc_pte)
+			pmd_install(vma->vm_mm, vmf->pmd, &vmf->prealloc_pte);
+		else if (unlikely(pte_alloc(vma->vm_mm, vmf->pmd)))
 			return VM_FAULT_OOM;
-		}
 	}
 
 	/* See comment in handle_pte_fault() */
-- 
2.11.0

[PATCH v3 2/2] mm: remove redundant smp_wmb()

[Qi Zheng]

The smp_wmb() which is in the __pte_alloc() is used to
ensure all ptes setup is visible before the pte is made
visible to other CPUs by being put into page tables. We
only need this when the pte is actually populated, so
move it to pmd_install(). __pte_alloc_kernel(),
__p4d_alloc(), __pud_alloc() and __pmd_alloc() are similar
to this case.

We can also defer smp_wmb() to the place where the pmd entry
is really populated by preallocated pte. There are two kinds
of user of preallocated pte, one is filemap & finish_fault(),
another is THP. The former does not need another smp_wmb()
because the smp_wmb() has been done by pmd_install().
Fortunately, the latter also does not need another smp_wmb()
because there is already a smp_wmb() before populating the
new pte when the THP uses a preallocated pte to split a huge
pmd.

Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
Reviewed-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: David Hildenbrand <david@redhat.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 mm/memory.c         | 52 +++++++++++++++++++++++-----------------------------
 mm/sparse-vmemmap.c |  2 +-
 2 files changed, 24 insertions(+), 30 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index ef7b1762e996..658d8df9c70f 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -439,6 +439,20 @@ void pmd_install(struct mm_struct *mm, pmd_t *pmd, pgtable_t *pte)
 
 	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
 		mm_inc_nr_ptes(mm);
+		/*
+		 * Ensure all pte setup (eg. pte page lock and page clearing) are
+		 * visible before the pte is made visible to other CPUs by being
+		 * put into page tables.
+		 *
+		 * The other side of the story is the pointer chasing in the page
+		 * table walking code (when walking the page table without locking;
+		 * ie. most of the time). Fortunately, these data accesses consist
+		 * of a chain of data-dependent loads, meaning most CPUs (alpha
+		 * being the notable exception) will already guarantee loads are
+		 * seen in-order. See the alpha page table accessors for the
+		 * smp_rmb() barriers in page table walking code.
+		 */
+		smp_wmb(); /* Could be smp_wmb__xxx(before|after)_spin_lock */
 		pmd_populate(mm, pmd, *pte);
 		*pte = NULL;
 	}
@@ -451,21 +465,6 @@ int __pte_alloc(struct mm_struct *mm, pmd_t *pmd)
 	if (!new)
 		return -ENOMEM;
 
-	/*
-	 * Ensure all pte setup (eg. pte page lock and page clearing) are
-	 * visible before the pte is made visible to other CPUs by being
-	 * put into page tables.
-	 *
-	 * The other side of the story is the pointer chasing in the page
-	 * table walking code (when walking the page table without locking;
-	 * ie. most of the time). Fortunately, these data accesses consist
-	 * of a chain of data-dependent loads, meaning most CPUs (alpha
-	 * being the notable exception) will already guarantee loads are
-	 * seen in-order. See the alpha page table accessors for the
-	 * smp_rmb() barriers in page table walking code.
-	 */
-	smp_wmb(); /* Could be smp_wmb__xxx(before|after)_spin_lock */
-
 	pmd_install(mm, pmd, &new);
 	if (new)
 		pte_free(mm, new);
@@ -478,10 +477,9 @@ int __pte_alloc_kernel(pmd_t *pmd)
 	if (!new)
 		return -ENOMEM;
 
-	smp_wmb(); /* See comment in __pte_alloc */
-
 	spin_lock(&init_mm.page_table_lock);
 	if (likely(pmd_none(*pmd))) {	/* Has another populated it ? */
+		smp_wmb(); /* See comment in pmd_install() */
 		pmd_populate_kernel(&init_mm, pmd, new);
 		new = NULL;
 	}
@@ -3857,7 +3855,6 @@ static vm_fault_t __do_fault(struct vm_fault *vmf)
 		vmf->prealloc_pte = pte_alloc_one(vma->vm_mm);
 		if (!vmf->prealloc_pte)
 			return VM_FAULT_OOM;
-		smp_wmb(); /* See comment in __pte_alloc() */
 	}
 
 	ret = vma->vm_ops->fault(vmf);
@@ -3919,7 +3916,6 @@ vm_fault_t do_set_pmd(struct vm_fault *vmf, struct page *page)
 		vmf->prealloc_pte = pte_alloc_one(vma->vm_mm);
 		if (!vmf->prealloc_pte)
 			return VM_FAULT_OOM;
-		smp_wmb(); /* See comment in __pte_alloc() */
 	}
 
 	vmf->ptl = pmd_lock(vma->vm_mm, vmf->pmd);
@@ -4144,7 +4140,6 @@ static vm_fault_t do_fault_around(struct vm_fault *vmf)
 		vmf->prealloc_pte = pte_alloc_one(vmf->vma->vm_mm);
 		if (!vmf->prealloc_pte)
 			return VM_FAULT_OOM;
-		smp_wmb(); /* See comment in __pte_alloc() */
 	}
 
 	return vmf->vma->vm_ops->map_pages(vmf, start_pgoff, end_pgoff);
@@ -4819,13 +4814,13 @@ int __p4d_alloc(struct mm_struct *mm, pgd_t *pgd, unsigned long address)
 	if (!new)
 		return -ENOMEM;
 
-	smp_wmb(); /* See comment in __pte_alloc */
-
 	spin_lock(&mm->page_table_lock);
-	if (pgd_present(*pgd))		/* Another has populated it */
+	if (pgd_present(*pgd)) {	/* Another has populated it */
 		p4d_free(mm, new);
-	else
+	} else {
+		smp_wmb(); /* See comment in pmd_install() */
 		pgd_populate(mm, pgd, new);
+	}
 	spin_unlock(&mm->page_table_lock);
 	return 0;
 }
@@ -4842,11 +4837,10 @@ int __pud_alloc(struct mm_struct *mm, p4d_t *p4d, unsigned long address)
 	if (!new)
 		return -ENOMEM;
 
-	smp_wmb(); /* See comment in __pte_alloc */
-
 	spin_lock(&mm->page_table_lock);
 	if (!p4d_present(*p4d)) {
 		mm_inc_nr_puds(mm);
+		smp_wmb(); /* See comment in pmd_install() */
 		p4d_populate(mm, p4d, new);
 	} else	/* Another has populated it */
 		pud_free(mm, new);
@@ -4867,14 +4861,14 @@ int __pmd_alloc(struct mm_struct *mm, pud_t *pud, unsigned long address)
 	if (!new)
 		return -ENOMEM;
 
-	smp_wmb(); /* See comment in __pte_alloc */
-
 	ptl = pud_lock(mm, pud);
 	if (!pud_present(*pud)) {
 		mm_inc_nr_pmds(mm);
+		smp_wmb(); /* See comment in pmd_install() */
 		pud_populate(mm, pud, new);
-	} else	/* Another has populated it */
+	} else {	/* Another has populated it */
 		pmd_free(mm, new);
+	}
 	spin_unlock(ptl);
 	return 0;
 }
diff --git a/mm/sparse-vmemmap.c b/mm/sparse-vmemmap.c
index bdce883f9286..db6df27c852a 100644
--- a/mm/sparse-vmemmap.c
+++ b/mm/sparse-vmemmap.c
@@ -76,7 +76,7 @@ static int split_vmemmap_huge_pmd(pmd_t *pmd, unsigned long start,
 		set_pte_at(&init_mm, addr, pte, entry);
 	}
 
-	/* Make pte visible before pmd. See comment in __pte_alloc(). */
+	/* Make pte visible before pmd. See comment in pmd_install(). */
 	smp_wmb();
 	pmd_populate_kernel(&init_mm, pmd, pgtable);
 
-- 
2.11.0

Re: [PATCH v3 0/2] Do some code cleanups related to mm

[Qi Zheng]

Hi Andrew,

This patch series seems to have no other problems.
Can it be merged into your mm tree?

Thanks,
Qi

On 9/1/21 6:27 PM, Qi Zheng wrote:
> Hi,
> 
> This patch series aims to do some code cleanups related to mm.
> 
> This series is based on next-20210827.
> 
> Comments and suggestions are welcome.
> 
> Thanks,
> Qi.
> 
> Changelog in v2 -> v3:
>   - Move the declaration of pmd_install() to mm/internal.h
> 
> Changelog in v1 -> v2:
>   - Fix some typo and code style problems.
> 
> Qi Zheng (2):
>    mm: introduce pmd_install() helper
>    mm: remove redundant smp_wmb()
> 
>   mm/filemap.c        | 11 ++-----
>   mm/internal.h       |  1 +
>   mm/memory.c         | 86 ++++++++++++++++++++++++-----------------------------
>   mm/sparse-vmemmap.c |  2 +-
>   4 files changed, 43 insertions(+), 57 deletions(-)
>