Re: [PATCH] selinux: remove duplicated initialization of 'i' for clean-up

[Ondrej Mosnacek]

Hi Austin,

On Tue, Aug 24, 2021 at 4:23 AM Austin Kim <austindh.kim@xxxxxxxxx> wrote:
>
> From: Austin Kim <austin.kim@xxxxxxx>
>
> The local variable 'i' is used to be incremented inside while loop
> within sidtab_convert_tree(). Before while loop, 'i' is set to 0
> inside if/else statement.
>
> Since there is no 'goto' statement within sidtab_convert_tree(),
> it had better initialize 'i' as 0 once before if/else statement.
>
> Signed-off-by: Austin Kim <austin.kim@xxxxxxx>
> ---
>  security/selinux/ss/sidtab.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/security/selinux/ss/sidtab.c b/security/selinux/ss/sidtab.c
> index 656d50b09f76..301620de63d3 100644
> --- a/security/selinux/ss/sidtab.c
> +++ b/security/selinux/ss/sidtab.c
> @@ -374,7 +374,7 @@ static int sidtab_convert_tree(union sidtab_entry_inner *edst,
>                                struct sidtab_convert_params *convert)
>  {
>         int rc;
> -       u32 i;
> +       u32 i = 0;
>
>         if (level != 0) {
>                 if (!edst->ptr_inner) {
> @@ -383,7 +383,6 @@ static int sidtab_convert_tree(union sidtab_entry_inner *edst,
>                         if (!edst->ptr_inner)
>                                 return -ENOMEM;
>                 }
> -               i = 0;
>                 while (i < SIDTAB_INNER_ENTRIES && *pos < count) {

I must say I prefer the original version more, because it makes it
clear when you look at the loop that it starts at 0. Once you move the
initialization to the declaration section, readers will have to scan
the code upwards to find it out. As is, it's also less prone to error
if e.g. someone adds another loop before the existing ones and reuses
the variable.

In case anyone is wondering why I didn't make these for loops when I
wrote this code: Since the loop condition a little more than the usual
"for(i = 0; i < n; i++)" pattern, my intention was to emphasize that
this is not a "regular" for loop and that readers should read the loop
carefully to not miss something. But perhaps that's not a good reason
and they would look more natural as for loops. If others think a for
loop would look better here, I'd be OK with a patch that makes these
into for loops instead.

>                         rc = sidtab_convert_tree(&edst->ptr_inner->entries[i],
>                                                  &esrc->ptr_inner->entries[i],
> @@ -400,7 +399,6 @@ static int sidtab_convert_tree(union sidtab_entry_inner *edst,
>                         if (!edst->ptr_leaf)
>                                 return -ENOMEM;
>                 }
> -               i = 0;
>                 while (i < SIDTAB_LEAF_ENTRIES && *pos < count) {
>                         rc = convert->func(&esrc->ptr_leaf->entries[i].context,
>                                            &edst->ptr_leaf->entries[i].context,
> --
> 2.20.1
>

--
Ondrej Mosnacek
Software Engineer, Linux Security - SELinux kernel
Red Hat, Inc.