On Tue, Oct 21, 2014 at 03:42:13PM +0800, Qu Wenruo wrote:
> 'btrfs fi df' needs exactly one arguments as mount option,
> and due to the introduce of human readable options, some check is not
> valid now, the new optind can point to the ending NULL string.
>
> For example, you can run 'btrfs fi df' without any argument, and it will
> error as "ERROR: can't access '%s'" which means the argument number
> check is not valid.
Thanks for catching it.
> This patch will use the old check_argc_exact() but introduce opt_num to
> record how many options is provided, ensuring there will be and only be
> one mount point provided.
> --- a/cmds-filesystem.c
> +++ b/cmds-filesystem.c
> @@ -237,11 +237,11 @@ static int cmd_df(int argc, char **argv)
> struct btrfs_ioctl_space_args *sargs = NULL;
> int ret;
> int fd;
> + int opt_num = 0;
> char *path;
> DIR *dirstream = NULL;
> unsigned unit_mode = UNITS_DEFAULT;
>
> - optind = 1;
> while (1) {
> int long_index;
> static const struct option long_options[] = {
> @@ -257,6 +257,7 @@ static int cmd_df(int argc, char **argv)
> &long_index);
> if (c < 0)
> break;
> + opt_num++;
> switch (c) {
> case 'b':
> unit_mode = UNITS_RAW;
> @@ -290,7 +291,7 @@ static int cmd_df(int argc, char **argv)
> }
> }
>
> - if (check_argc_max(argc, optind + 1))
> + if (check_argc_exact(argc, 2 + opt_num))
Well, a simple change to check_argc_exact(argc, optind + 1) fixes it for
me. The optind is supposed to track the argument index and you're
duplicating it with opt_num.
I'll commit it with this simplified version (and changelog updated):
--- a/cmds-filesystem.c
+++ b/cmds-filesystem.c
@@ -248,7 +248,7 @@ static int cmd_filesystem_df(int argc, char **argv)
}
}
- if (check_argc_max(argc, optind + 1))
+ if (check_argc_exact(argc, optind + 1))
usage(cmd_filesystem_df_usage);
path = argv[optind];
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
