VM: don't bother with feeding upper limit to tlb_finish_mmu() in exit_mmap()
authorAl Viro <viro@zeniv.linux.org.uk>
Mon, 5 Mar 2012 19:03:47 +0000 (14:03 -0500)
committerAl Viro <viro@zeniv.linux.org.uk>
Wed, 21 Mar 2012 01:39:51 +0000 (21:39 -0400)
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

index da15a79b1441b665b0e4b962eea38a92407018fd..2b2b45eb816c26d11b477003c532818f9e6f316b 100644 (file)
--- a/mm/mmap.c
+++ b/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,