Re: [PATCH] udf: Fix to check the return value of load_nls

[Jan Kara]

On Mon 18-03-19 21:59:49, Aditya Pakki wrote:
> load_nls may fail and return an error message. The patch checks
> for such a scenario and passes the error upstream.
> 
> Signed-off-by: Aditya Pakki <pakki001@xxxxxxx>

Thanks for the patch! But there's other code handling nls_map in
udf_fill_super() which takes care about this situation:

        if ((uopt.flags & (1 << UDF_FLAG_NLS_MAP)) && !uopt.nls_map) {
                uopt.nls_map = load_nls_default();
                if (!uopt.nls_map)
                        uopt.flags &= ~(1 << UDF_FLAG_NLS_MAP);
                else
                        udf_debug("Using default NLS map\n");
        }

So if we failed to load nls map admin originally requested, we fall back to
a default nls map or just don't load anything. It is questionable whether
we shouldn't fail mount rather than trying to continue with a different nls
map but at this point I don't see a strong reason to change the current
behavior users can depend on...

I've added a comment to udf_parse_options() to explain the code flow.

								Honza

> ---
>  fs/udf/super.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/fs/udf/super.c b/fs/udf/super.c
> index ffd8038ff728..1a38271de6d9 100644
> --- a/fs/udf/super.c
> +++ b/fs/udf/super.c
> @@ -573,6 +573,8 @@ static int udf_parse_options(char *options, struct udf_options *uopt,
>  				if (uopt->nls_map)
>  					unload_nls(uopt->nls_map);
>  				uopt->nls_map = load_nls(args[0].from);
> +				if (!uopt->nls_map)
> +					return 0;
>  				uopt->flags |= (1 << UDF_FLAG_NLS_MAP);
>  			}
>  			break;
> -- 
> 2.17.1
> 
> 
-- 
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR