MFC r320498

  Clear the MAP_WIREFUTURE flag on the vm map in exec_new_vmspace() when it
  recycles the current vm space.  Otherwise, an mlockall(MCL_FUTURE) could
  still be in effect on the process after an execve(2), which violates the
  specification for mlockall(2).

  It's pointless for vm_map_stack() to check the MEMLOCK limit.  It will
  never be asked to wire the stack.  Moreover, it doesn't even implement
  wiring of the stack.

2 files changed, 10 insertions, 8 deletions

diff --git a/sys/kern/kern_exec.c b/sys/kern/kern_exec.c
index a1e673a..cc77489 100644
--- a/sys/kern/kern_exec.c
+++ b/sys/kern/kern_exec.c
@@ -1082,6 +1082,10 @@ exec_new_vmspace(imgp, sv)
 		shmexit(vmspace);
 		pmap_remove_pages(vmspace_pmap(vmspace));
 		vm_map_remove(map, vm_map_min(map), vm_map_max(map));
+		/* An exec terminates mlockall(MCL_FUTURE). */
+		vm_map_lock(map);
+		vm_map_modflags(map, 0, MAP_WIREFUTURE);
+		vm_map_unlock(map);
 	} else {
 		error = vmspace_exec(p, sv_minuser, sv->sv_maxuser);
 		if (error)
diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c
index cbaf18d..6913d22 100644
--- a/sys/vm/vm_map.c
+++ b/sys/vm/vm_map.c
@@ -3445,27 +3445,25 @@ vmspace_fork(struct vmspace *vm1, vm_ooffset_t *fork_charge)
 	return (vm2);
 }
 
+/*
+ * Create a process's stack for exec_new_vmspace().  This function is never
+ * asked to wire the newly created stack.
+ */
 int
 vm_map_stack(vm_map_t map, vm_offset_t addrbos, vm_size_t max_ssize,
     vm_prot_t prot, vm_prot_t max, int cow)
 {
 	vm_size_t growsize, init_ssize;
-	rlim_t lmemlim, vmemlim;
+	rlim_t vmemlim;
 	int rv;
 
+	MPASS((map->flags & MAP_WIREFUTURE) == 0);
 	growsize = sgrowsiz;
 	init_ssize = (max_ssize < growsize) ? max_ssize : growsize;
 	vm_map_lock(map);
 	PROC_LOCK(curproc);
-	lmemlim = lim_cur(curproc, RLIMIT_MEMLOCK);
 	vmemlim = lim_cur(curproc, RLIMIT_VMEM);
 	PROC_UNLOCK(curproc);
-	if (!old_mlock && map->flags & MAP_WIREFUTURE) {
-		if (ptoa(pmap_wired_count(map->pmap)) + init_ssize > lmemlim) {
-			rv = KERN_NO_SPACE;
-			goto out;
-		}
-	}
 	/* If we would blow our VMEM resource limit, no go */
 	if (map->size + init_ssize > vmemlim) {
 		rv = KERN_NO_SPACE;