Re: [PATCH 1/4] selinux: add support for querying object classes and permissions from the running policy

[James Antill <jantill@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1432 bytes --]

On Wed, 2007-05-23 at 09:12 -0400, Christopher J. PeBenito wrote:
> 
> +static int get_permissions_callback(void *k, void *d, void *args)
> +{
> +       struct perm_datum *datum = d;
> +       char *name = k, **perms = args;
> +       int value = datum->value - 1;
> +
> +       perms[value] = kstrdup(name, GFP_ATOMIC);
> +       if (!perms[value])
> +               return -ENOMEM;
> +
> +       return 0;
> +}
> +
> +int security_get_permissions(char *class, char ***perms, int *nperms)
> +{
[...]
> +       *nperms = match->permissions.nprim;
> +       *perms = kcalloc(*nperms, sizeof(*perms), GFP_ATOMIC);
> +       if (!*perms)
> +               goto out;
> +
> +       if (match->comdatum) {
> +               rc = hashtab_map(match->comdatum->permissions.table,
> +                               get_permissions_callback, *perms);
> +               if (rc < 0)
> +                       goto err;
> +       }
> +
> +       rc = hashtab_map(match->permissions.table, get_permissions_callback,
> +                       *perms);


 So I'm probably just wasting your time out of ignorance here, but this
looks really suspicious to me. Why don't you need to add in
match->comdatum->permissions.nprim for the allocation? Also how is it
guaranteed that one "datum->value - 1" won't be the same in both
hashtabs? (if it isn't you've leaked memory).

-- 
James Antill <jantill@redhat.com>

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]