[S390] pgtables: Fix race in enable_sie vs. page table ops

The current enable_sie code sets the mm->context.pgstes bit to tell
dup_mm that the new mm should have extended page tables. This bit is also
used by the s390 specific page table primitives to decide about the page
table layout - which means context.pgstes has two meanings. This can cause
any kind of bugs. For example  - e.g. shrink_zone can call
ptep_clear_flush_young while enable_sie is running. ptep_clear_flush_young
will test for context.pgstes. Since enable_sie changed that value of the old
struct mm without changing the page table layout ptep_clear_flush_young will
do the wrong thing.
The solution is to split pgstes into two bits
- one for the allocation
- one for the current state

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

3 files changed, 22 insertions, 8 deletions

diff --git a/arch/s390/include/asm/mmu.h b/arch/s390/include/asm/mmu.h
index 5dd5e7b3476..d2b4ff83147 100644
--- a/arch/s390/include/asm/mmu.h
+++ b/arch/s390/include/asm/mmu.h
@@ -7,7 +7,8 @@ typedef struct {
 	unsigned long asce_bits;
 	unsigned long asce_limit;
 	int noexec;
-	int pgstes;
+	int has_pgste;	 /* The mmu context has extended page tables */
+	int alloc_pgste; /* cloned contexts will have extended page tables */
 } mm_context_t;
 
 #endif
diff --git a/arch/s390/include/asm/mmu_context.h b/arch/s390/include/asm/mmu_context.h
index 4c2fbf48c9c..28ec870655a 100644
--- a/arch/s390/include/asm/mmu_context.h
+++ b/arch/s390/include/asm/mmu_context.h
@@ -20,12 +20,25 @@ static inline int init_new_context(struct task_struct *tsk,
 #ifdef CONFIG_64BIT
 	mm->context.asce_bits |= _ASCE_TYPE_REGION3;
 #endif
-	if (current->mm->context.pgstes) {
+	if (current->mm->context.alloc_pgste) {
+		/*
+		 * alloc_pgste indicates, that any NEW context will be created
+		 * with extended page tables. The old context is unchanged. The
+		 * page table allocation and the page table operations will
+		 * look at has_pgste to distinguish normal and extended page
+		 * tables. The only way to create extended page tables is to
+		 * set alloc_pgste and then create a new context (e.g. dup_mm).
+		 * The page table allocation is called after init_new_context
+		 * and if has_pgste is set, it will create extended page
+		 * tables.
+		 */
 		mm->context.noexec = 0;
-		mm->context.pgstes = 1;
+		mm->context.has_pgste = 1;
+		mm->context.alloc_pgste = 1;
 	} else {
 		mm->context.noexec = s390_noexec;
-		mm->context.pgstes = 0;
+		mm->context.has_pgste = 0;
+		mm->context.alloc_pgste = 0;
 	}
 	mm->context.asce_limit = STACK_TOP_MAX;
 	crst_table_init((unsigned long *) mm->pgd, pgd_entry_type(mm));
diff --git a/arch/s390/include/asm/pgtable.h b/arch/s390/include/asm/pgtable.h
index 1a928f84afd..7fc76133b3e 100644
--- a/arch/s390/include/asm/pgtable.h
+++ b/arch/s390/include/asm/pgtable.h
@@ -679,7 +679,7 @@ static inline void pmd_clear(pmd_t *pmd)
 
 static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
 {
-	if (mm->context.pgstes)
+	if (mm->context.has_pgste)
 		ptep_rcp_copy(ptep);
 	pte_val(*ptep) = _PAGE_TYPE_EMPTY;
 	if (mm->context.noexec)
@@ -763,7 +763,7 @@ static inline int kvm_s390_test_and_clear_page_dirty(struct mm_struct *mm,
 	struct page *page;
 	unsigned int skey;
 
-	if (!mm->context.pgstes)
+	if (!mm->context.has_pgste)
 		return -EINVAL;
 	rcp_lock(ptep);
 	pgste = (unsigned long *) (ptep + PTRS_PER_PTE);
@@ -794,7 +794,7 @@ static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
 	int young;
 	unsigned long *pgste;
 
-	if (!vma->vm_mm->context.pgstes)
+	if (!vma->vm_mm->context.has_pgste)
 		return 0;
 	physpage = pte_val(*ptep) & PAGE_MASK;
 	pgste = (unsigned long *) (ptep + PTRS_PER_PTE);
@@ -844,7 +844,7 @@ static inline void __ptep_ipte(unsigned long address, pte_t *ptep)
 static inline void ptep_invalidate(struct mm_struct *mm,
 				   unsigned long address, pte_t *ptep)
 {
-	if (mm->context.pgstes) {
+	if (mm->context.has_pgste) {
 		rcp_lock(ptep);
 		__ptep_ipte(address, ptep);
 		ptep_rcp_copy(ptep);