9 files changed, 92 insertions, 10 deletions

diff --git a/arch/x86/kernel/vm86_32.c b/arch/x86/kernel/vm86_32.c
index b466cab..328cb37 100644
--- a/arch/x86/kernel/vm86_32.c
+++ b/arch/x86/kernel/vm86_32.c
@@ -172,6 +172,7 @@ static void mark_screen_rdonly(struct mm_struct *mm)
 	spinlock_t *ptl;
 	int i;
 
+	down_write(&mm->mmap_sem);
 	pgd = pgd_offset(mm, 0xA0000);
 	if (pgd_none_or_clear_bad(pgd))
 		goto out;
@@ -190,6 +191,7 @@ static void mark_screen_rdonly(struct mm_struct *mm)
 	}
 	pte_unmap_unlock(pte, ptl);
 out:
+	up_write(&mm->mmap_sem);
 	flush_tlb();
 }
 
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 7dcd2a2..3efa725 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -409,6 +409,9 @@ static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 	} else {
 		spin_unlock(&walk->mm->page_table_lock);
 	}
+
+	if (pmd_trans_unstable(pmd))
+		return 0;
 	/*
 	 * The mmap_sem held all the way back in m_start() is what
 	 * keeps khugepaged out of here and from collapsing things
@@ -507,6 +510,8 @@ static int clear_refs_pte_range(pmd_t *pmd, unsigned long addr,
 	struct page *page;
 
 	split_huge_page_pmd(walk->mm, pmd);
+	if (pmd_trans_unstable(pmd))
+		return 0;
 
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
 	for (; addr != end; pte++, addr += PAGE_SIZE) {
@@ -670,6 +675,8 @@ static int pagemap_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end,
 	int err = 0;
 
 	split_huge_page_pmd(walk->mm, pmd);
+	if (pmd_trans_unstable(pmd))
+		return 0;
 
 	/* find the first VMA at or above 'addr' */
 	vma = find_vma(walk->mm, addr);
@@ -961,6 +968,8 @@ static int gather_pte_stats(pmd_t *pmd, unsigned long addr,
 		spin_unlock(&walk->mm->page_table_lock);
 	}
 
+	if (pmd_trans_unstable(pmd))
+		return 0;
 	orig_pte = pte = pte_offset_map_lock(walk->mm, pmd, addr, &ptl);
 	do {
 		struct page *page = can_gather_numa_stats(*pte, md->vma, addr);
diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h
index 76bff2b..a03c098 100644
--- a/include/asm-generic/pgtable.h
+++ b/include/asm-generic/pgtable.h
@@ -425,6 +425,8 @@ extern void untrack_pfn_vma(struct vm_area_struct *vma, unsigned long pfn,
 				unsigned long size);
 #endif
 
+#ifdef CONFIG_MMU
+
 #ifndef CONFIG_TRANSPARENT_HUGEPAGE
 static inline int pmd_trans_huge(pmd_t pmd)
 {
@@ -441,7 +443,66 @@ static inline int pmd_write(pmd_t pmd)
 	return 0;
 }
 #endif /* __HAVE_ARCH_PMD_WRITE */
+#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
+
+/*
+ * This function is meant to be used by sites walking pagetables with
+ * the mmap_sem hold in read mode to protect against MADV_DONTNEED and
+ * transhuge page faults. MADV_DONTNEED can convert a transhuge pmd
+ * into a null pmd and the transhuge page fault can convert a null pmd
+ * into an hugepmd or into a regular pmd (if the hugepage allocation
+ * fails). While holding the mmap_sem in read mode the pmd becomes
+ * stable and stops changing under us only if it's not null and not a
+ * transhuge pmd. When those races occurs and this function makes a
+ * difference vs the standard pmd_none_or_clear_bad, the result is
+ * undefined so behaving like if the pmd was none is safe (because it
+ * can return none anyway). The compiler level barrier() is critically
+ * important to compute the two checks atomically on the same pmdval.
+ */
+static inline int pmd_none_or_trans_huge_or_clear_bad(pmd_t *pmd)
+{
+	/* depend on compiler for an atomic pmd read */
+	pmd_t pmdval = *pmd;
+	/*
+	 * The barrier will stabilize the pmdval in a register or on
+	 * the stack so that it will stop changing under the code.
+	 */
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	barrier();
+#endif
+	if (pmd_none(pmdval))
+		return 1;
+	if (unlikely(pmd_bad(pmdval))) {
+		if (!pmd_trans_huge(pmdval))
+			pmd_clear_bad(pmd);
+		return 1;
+	}
+	return 0;
+}
+
+/*
+ * This is a noop if Transparent Hugepage Support is not built into
+ * the kernel. Otherwise it is equivalent to
+ * pmd_none_or_trans_huge_or_clear_bad(), and shall only be called in
+ * places that already verified the pmd is not none and they want to
+ * walk ptes while holding the mmap sem in read mode (write mode don't
+ * need this). If THP is not enabled, the pmd can't go away under the
+ * code even if MADV_DONTNEED runs, but if THP is enabled we need to
+ * run a pmd_trans_unstable before walking the ptes after
+ * split_huge_page_pmd returns (because it may have run when the pmd
+ * become null, but then a page fault can map in a THP and not a
+ * regular page).
+ */
+static inline int pmd_trans_unstable(pmd_t *pmd)
+{
+#ifdef CONFIG_TRANSPARENT_HUGEPAGE
+	return pmd_none_or_trans_huge_or_clear_bad(pmd);
+#else
+	return 0;
 #endif
+}
+
+#endif /* CONFIG_MMU */
 
 #endif /* !__ASSEMBLY__ */
 
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 26c6f4e..3728181 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -5230,6 +5230,8 @@ static int mem_cgroup_count_precharge_pte_range(pmd_t *pmd,
 	spinlock_t *ptl;
 
 	split_huge_page_pmd(walk->mm, pmd);
+	if (pmd_trans_unstable(pmd))
+		return 0;
 
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
 	for (; addr != end; pte++, addr += PAGE_SIZE)
@@ -5390,6 +5392,8 @@ static int mem_cgroup_move_charge_pte_range(pmd_t *pmd,
 	spinlock_t *ptl;
 
 	split_huge_page_pmd(walk->mm, pmd);
+	if (pmd_trans_unstable(pmd))
+		return 0;
 retry:
 	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
 	for (; addr != end; addr += PAGE_SIZE) {
diff --git a/mm/memory.c b/mm/memory.c
index 347e5fa..e01abb9 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -1247,16 +1247,24 @@ static inline unsigned long zap_pmd_range(struct mmu_gather *tlb,
 	do {
 		next = pmd_addr_end(addr, end);
 		if (pmd_trans_huge(*pmd)) {
-			if (next-addr != HPAGE_PMD_SIZE) {
+			if (next - addr != HPAGE_PMD_SIZE) {
 				VM_BUG_ON(!rwsem_is_locked(&tlb->mm->mmap_sem));
 				split_huge_page_pmd(vma->vm_mm, pmd);
 			} else if (zap_huge_pmd(tlb, vma, pmd, addr))
-				continue;
+				goto next;
 			/* fall through */
 		}
-		if (pmd_none_or_clear_bad(pmd))
-			continue;
+		/*
+		 * Here there can be other concurrent MADV_DONTNEED or
+		 * trans huge page faults running, and if the pmd is
+		 * none or trans huge it can change under us. This is
+		 * because MADV_DONTNEED holds the mmap_sem in read
+		 * mode.
+		 */
+		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
+			goto next;
 		next = zap_pte_range(tlb, vma, pmd, addr, next, details);
+next:
 		cond_resched();
 	} while (pmd++, addr = next, addr != end);
 
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 47296fe..0a37570 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -512,7 +512,7 @@ static inline int check_pmd_range(struct vm_area_struct *vma, pud_t *pud,
 	do {
 		next = pmd_addr_end(addr, end);
 		split_huge_page_pmd(vma->vm_mm, pmd);
-		if (pmd_none_or_clear_bad(pmd))
+		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
 			continue;
 		if (check_pte_range(vma, pmd, addr, next, nodes,
 				    flags, private))
diff --git a/mm/mincore.c b/mm/mincore.c
index 636a868..936b4ce 100644
--- a/mm/mincore.c
+++ b/mm/mincore.c
@@ -164,7 +164,7 @@ static void mincore_pmd_range(struct vm_area_struct *vma, pud_t *pud,
 			}
 			/* fall through */
 		}
-		if (pmd_none_or_clear_bad(pmd))
+		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
 			mincore_unmapped_range(vma, addr, next, vec);
 		else
 			mincore_pte_range(vma, pmd, addr, next, vec);
diff --git a/mm/pagewalk.c b/mm/pagewalk.c
index 2f5cf10..aa9701e 100644
--- a/mm/pagewalk.c
+++ b/mm/pagewalk.c
@@ -59,7 +59,7 @@ again:
 			continue;
 
 		split_huge_page_pmd(walk->mm, pmd);
-		if (pmd_none_or_clear_bad(pmd))
+		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
 			goto again;
 		err = walk_pte_range(pmd, addr, next, walk);
 		if (err)
diff --git a/mm/swapfile.c b/mm/swapfile.c
index 00a962c..44595a3 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -932,9 +932,7 @@ static inline int unuse_pmd_range(struct vm_area_struct *vma, pud_t *pud,
 	pmd = pmd_offset(pud, addr);
 	do {
 		next = pmd_addr_end(addr, end);
-		if (unlikely(pmd_trans_huge(*pmd)))
-			continue;
-		if (pmd_none_or_clear_bad(pmd))
+		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
 			continue;
 		ret = unuse_pte_range(vma, pmd, addr, next, entry, page);
 		if (ret)