powerpc/64s: Fix page table fragment refcount race vs speculative references

The page table fragment allocator uses the main page refcount racily
with respect to speculative references. A customer observed a BUG due
to page table page refcount underflow in the fragment allocator. This
can be caused by the fragment allocator set_page_count stomping on a
speculative reference, and then the speculative failure handler
decrements the new reference, and the underflow eventually pops when
the page tables are freed.

Fix this by using a dedicated field in the struct page for the page
table fragment allocator.

Fixes: 5c1f6ee9a31c ("powerpc: Reduce PTE table memory wastage")
Cc: stable@vger.kernel.org # v3.10+
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

1 files changed, 11 insertions, 6 deletions

diff --git a/arch/powerpc/mm/pgtable-book3s64.c b/arch/powerpc/mm/pgtable-book3s64.c
index 4afbfbb64bfd..78d0b3d5ebad 100644
--- a/arch/powerpc/mm/pgtable-book3s64.c
+++ b/arch/powerpc/mm/pgtable-book3s64.c
@@ -270,6 +270,8 @@ static pmd_t *__alloc_for_pmdcache(struct mm_struct *mm)
 		return NULL;
 	}
 
+	atomic_set(&page->pt_frag_refcount, 1);
+
 	ret = page_address(page);
 	/*
 	 * if we support only one fragment just return the
@@ -285,7 +287,7 @@ static pmd_t *__alloc_for_pmdcache(struct mm_struct *mm)
 	 * count.
 	 */
 	if (likely(!mm->context.pmd_frag)) {
-		set_page_count(page, PMD_FRAG_NR);
+		atomic_set(&page->pt_frag_refcount, PMD_FRAG_NR);
 		mm->context.pmd_frag = ret + PMD_FRAG_SIZE;
 	}
 	spin_unlock(&mm->page_table_lock);
@@ -308,9 +310,10 @@ void pmd_fragment_free(unsigned long *pmd)
 {
 	struct page *page = virt_to_page(pmd);
 
-	if (put_page_testzero(page)) {
+	BUG_ON(atomic_read(&page->pt_frag_refcount) <= 0);
+	if (atomic_dec_and_test(&page->pt_frag_refcount)) {
 		pgtable_pmd_page_dtor(page);
-		free_unref_page(page);
+		__free_page(page);
 	}
 }
 
@@ -352,6 +355,7 @@ static pte_t *__alloc_for_ptecache(struct mm_struct *mm, int kernel)
 			return NULL;
 	}
 
+	atomic_set(&page->pt_frag_refcount, 1);
 
 	ret = page_address(page);
 	/*
@@ -367,7 +371,7 @@ static pte_t *__alloc_for_ptecache(struct mm_struct *mm, int kernel)
 	 * count.
 	 */
 	if (likely(!mm->context.pte_frag)) {
-		set_page_count(page, PTE_FRAG_NR);
+		atomic_set(&page->pt_frag_refcount, PTE_FRAG_NR);
 		mm->context.pte_frag = ret + PTE_FRAG_SIZE;
 	}
 	spin_unlock(&mm->page_table_lock);
@@ -390,10 +394,11 @@ void pte_fragment_free(unsigned long *table, int kernel)
 {
 	struct page *page = virt_to_page(table);
 
-	if (put_page_testzero(page)) {
+	BUG_ON(atomic_read(&page->pt_frag_refcount) <= 0);
+	if (atomic_dec_and_test(&page->pt_frag_refcount)) {
 		if (!kernel)
 			pgtable_page_dtor(page);
-		free_unref_page(page);
+		__free_page(page);
 	}
 }