VM: don't bother with feeding upper limit to tlb_finish_mmu() in exit…

…_mmap()

no point, really - the only instance that cares about those arguments of
tlb_finish_mmu() is itanic and there we explicitly check if that's called
from exit_mmap() (i.e. that ->fullmm is set), in which case we ignore those
arguments completely.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

mm/mmap.c

@@ -2253,7 +2253,7 @@ void exit_mmap(struct mm_struct *mm)
 	vm_unacct_memory(nr_accounted);
 
 	free_pgtables(&tlb, vma, FIRST_USER_ADDRESS, 0);
-	tlb_finish_mmu(&tlb, 0, end);
+	tlb_finish_mmu(&tlb, 0, -1);
 
 	/*
 	 * Walk the list again, actually closing and freeing it,