[S390] noexec protection

This provides a noexec protection on s390 hardware. Our hardware does
not have any bits left in the pte for a hw noexec bit, so this is a
different approach using shadow page tables and a special addressing
mode that allows separate address spaces for code and data.

As a special feature of our "secondary-space" addressing mode, separate
page tables can be specified for the translation of data addresses
(storage operands) and instruction addresses. The shadow page table is
used for the instruction addresses and the standard page table for the
data addresses.
The shadow page table is linked to the standard page table by a pointer
in page->lru.next of the struct page corresponding to the page that
contains the standard page table (since page->private is not really
private with the pte_lock and the page table pages are not in the LRU
list).
Depending on the software bits of a pte, it is either inserted into
both page tables or just into the standard (data) page table. Pages of
a vma that does not have the VM_EXEC bit set get mapped only in the
data address space. Any try to execute code on such a page will cause a
page translation exception. The standard reaction to this is a SIGSEGV
with two exceptions: the two system call opcodes 0x0a77 (sys_sigreturn)
and 0x0aad (sys_rt_sigreturn) are allowed. They are stored by the
kernel to the signal stack frame. Unfortunately, the signal return
mechanism cannot be modified to use an SA_RESTORER because the
exception unwinding code depends on the system call opcode stored
behind the signal stack frame.

This feature requires that user space is executed in secondary-space
mode and the kernel in home-space mode, which means that the addressing
modes need to be switched and that the noexec protection only works
for user space.
After switching the addressing modes, we cannot use the mvcp/mvcs
instructions anymore to copy between kernel and user space. A new
mvcos instruction has been added to the z9 EC/BC hardware which allows
to copy between arbitrary address spaces, but on older hardware the
page tables need to be walked manually.

Signed-off-by: Gerald Schaefer <geraldsc@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

3 files changed, 96 insertions, 12 deletions

diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 3382e29..9ff143e 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -137,7 +137,9 @@ static int __check_access_register(struct pt_regs *regs, int error_code)
 
 /*
  * Check which address space the address belongs to.
- * Returns 1 for user space and 0 for kernel space.
+ * May return 1 or 2 for user space and 0 for kernel space.
+ * Returns 2 for user space in primary addressing mode with
+ * CONFIG_S390_EXEC_PROTECT on and kernel parameter noexec=on.
  */
 static inline int check_user_space(struct pt_regs *regs, int error_code)
 {
@@ -154,7 +156,7 @@ static inline int check_user_space(struct pt_regs *regs, int error_code)
 		return __check_access_register(regs, error_code);
 	if (descriptor == 2)
 		return current->thread.mm_segment.ar4;
-	return descriptor != 0;
+	return ((descriptor != 0) ^ (switch_amode)) << s390_noexec;
 }
 
 /*
@@ -183,6 +185,77 @@ static void do_sigsegv(struct pt_regs *regs, unsigned long error_code,
 	force_sig_info(SIGSEGV, &si, current);
 }
 
+#ifdef CONFIG_S390_EXEC_PROTECT
+extern long sys_sigreturn(struct pt_regs *regs);
+extern long sys_rt_sigreturn(struct pt_regs *regs);
+extern long sys32_sigreturn(struct pt_regs *regs);
+extern long sys32_rt_sigreturn(struct pt_regs *regs);
+
+static inline void do_sigreturn(struct mm_struct *mm, struct pt_regs *regs,
+				int rt)
+{
+	up_read(&mm->mmap_sem);
+	clear_tsk_thread_flag(current, TIF_SINGLE_STEP);
+#ifdef CONFIG_COMPAT
+	if (test_tsk_thread_flag(current, TIF_31BIT)) {
+		if (rt)
+			sys32_rt_sigreturn(regs);
+		else
+			sys32_sigreturn(regs);
+		return;
+	}
+#endif /* CONFIG_COMPAT */
+	if (rt)
+		sys_rt_sigreturn(regs);
+	else
+		sys_sigreturn(regs);
+	return;
+}
+
+static int signal_return(struct mm_struct *mm, struct pt_regs *regs,
+			 unsigned long address, unsigned long error_code)
+{
+	pgd_t *pgd;
+	pmd_t *pmd;
+	pte_t *pte;
+	u16 *instruction;
+	unsigned long pfn, uaddr = regs->psw.addr;
+
+	spin_lock(&mm->page_table_lock);
+	pgd = pgd_offset(mm, uaddr);
+	if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd)))
+		goto out_fault;
+	pmd = pmd_offset(pgd, uaddr);
+	if (pmd_none(*pmd) || unlikely(pmd_bad(*pmd)))
+		goto out_fault;
+	pte = pte_offset_map(pmd_offset(pgd_offset(mm, uaddr), uaddr), uaddr);
+	if (!pte || !pte_present(*pte))
+		goto out_fault;
+	pfn = pte_pfn(*pte);
+	if (!pfn_valid(pfn))
+		goto out_fault;
+	spin_unlock(&mm->page_table_lock);
+
+	instruction = (u16 *) ((pfn << PAGE_SHIFT) + (uaddr & (PAGE_SIZE-1)));
+	if (*instruction == 0x0a77)
+		do_sigreturn(mm, regs, 0);
+	else if (*instruction == 0x0aad)
+		do_sigreturn(mm, regs, 1);
+	else {
+		printk("- XXX - do_exception: task = %s, primary, NO EXEC "
+		       "-> SIGSEGV\n", current->comm);
+		up_read(&mm->mmap_sem);
+		current->thread.prot_addr = address;
+		current->thread.trap_no = error_code;
+		do_sigsegv(regs, error_code, SEGV_MAPERR, address);
+	}
+	return 0;
+out_fault:
+	spin_unlock(&mm->page_table_lock);
+	return -EFAULT;
+}
+#endif /* CONFIG_S390_EXEC_PROTECT */
+
 /*
  * This routine handles page faults.  It determines the address,
  * and the problem, and then passes it off to one of the appropriate
@@ -260,6 +333,17 @@ do_exception(struct pt_regs *regs, unsigned long error_code, int is_protection)
         vma = find_vma(mm, address);
         if (!vma)
                 goto bad_area;
+
+#ifdef CONFIG_S390_EXEC_PROTECT
+	if (unlikely((user_address == 2) && !(vma->vm_flags & VM_EXEC)))
+		if (!signal_return(mm, regs, address, error_code))
+			/*
+			 * signal_return() has done an up_read(&mm->mmap_sem)
+			 * if it returns 0.
+			 */
+			return;
+#endif
+
         if (vma->vm_start <= address) 
                 goto good_area;
         if (!(vma->vm_flags & VM_GROWSDOWN))
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index 0e7e9ac..162a338 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -104,7 +104,7 @@ static void __init setup_ro_region(void)
 		pmd = pmd_offset(pgd, address);
 		pte = pte_offset_kernel(pmd, address);
 		new_pte = mk_pte_phys(address, __pgprot(_PAGE_RO));
-		set_pte(pte, new_pte);
+		*pte = new_pte;
 	}
 }
 
@@ -124,11 +124,11 @@ void __init paging_init(void)
 #ifdef CONFIG_64BIT
 	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERN_REGION_TABLE;
 	for (i = 0; i < PTRS_PER_PGD; i++)
-		pgd_clear(pg_dir + i);
+		pgd_clear_kernel(pg_dir + i);
 #else
 	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERNSEG_TABLE;
 	for (i = 0; i < PTRS_PER_PGD; i++)
-		pmd_clear((pmd_t *)(pg_dir + i));
+		pmd_clear_kernel((pmd_t *)(pg_dir + i));
 #endif
 	vmem_map_init();
 	setup_ro_region();
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index cd3d93e..92a5651 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -82,7 +82,7 @@ static inline pmd_t *vmem_pmd_alloc(void)
 	if (!pmd)
 		return NULL;
 	for (i = 0; i < PTRS_PER_PMD; i++)
-		pmd_clear(pmd + i);
+		pmd_clear_kernel(pmd + i);
 	return pmd;
 }
 
@@ -97,7 +97,7 @@ static inline pte_t *vmem_pte_alloc(void)
 		return NULL;
 	pte_val(empty_pte) = _PAGE_TYPE_EMPTY;
 	for (i = 0; i < PTRS_PER_PTE; i++)
-		set_pte(pte + i, empty_pte);
+		pte[i] = empty_pte;
 	return pte;
 }
 
@@ -119,7 +119,7 @@ static int vmem_add_range(unsigned long start, unsigned long size)
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pgd_populate(&init_mm, pg_dir, pm_dir);
+			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
 		}
 
 		pm_dir = pmd_offset(pg_dir, address);
@@ -132,7 +132,7 @@ static int vmem_add_range(unsigned long start, unsigned long size)
 
 		pt_dir = pte_offset_kernel(pm_dir, address);
 		pte = pfn_pte(address >> PAGE_SHIFT, PAGE_KERNEL);
-		set_pte(pt_dir, pte);
+		*pt_dir = pte;
 	}
 	ret = 0;
 out:
@@ -161,7 +161,7 @@ static void vmem_remove_range(unsigned long start, unsigned long size)
 		if (pmd_none(*pm_dir))
 			continue;
 		pt_dir = pte_offset_kernel(pm_dir, address);
-		set_pte(pt_dir, pte);
+		*pt_dir = pte;
 	}
 	flush_tlb_kernel_range(start, start + size);
 }
@@ -191,7 +191,7 @@ static int vmem_add_mem_map(unsigned long start, unsigned long size)
 			pm_dir = vmem_pmd_alloc();
 			if (!pm_dir)
 				goto out;
-			pgd_populate(&init_mm, pg_dir, pm_dir);
+			pgd_populate_kernel(&init_mm, pg_dir, pm_dir);
 		}
 
 		pm_dir = pmd_offset(pg_dir, address);
@@ -210,7 +210,7 @@ static int vmem_add_mem_map(unsigned long start, unsigned long size)
 			if (!new_page)
 				goto out;
 			pte = pfn_pte(new_page >> PAGE_SHIFT, PAGE_KERNEL);
-			set_pte(pt_dir, pte);
+			*pt_dir = pte;
 		}
 	}
 	ret = 0;