Re: [2/5] server: A new function "set_sd_defaults_from_token" that sets the security descriptor along with a token that will be used to gather defaults, instead of always using the primary token.

[Alexandre Julliard]

Joris van der Wel  writes:
> +/** Set the security descriptor using the current primary token for defaults. */
> +int default_set_sd( struct object *obj, const struct security_descriptor *sd,
> -
>                unsigned int set_info )
>
> +{
> - return set_sd_defaults_from_token( obj, sd, set_info, current ? current->process->token : NULL );

There shouldn't be any need to test the current pointer for null. If
it's needed, that probably indicates some other problem.

[2/5] server: A new function "set_sd_defaults_from_token" that sets the security descriptor along with a token that will be used to gather defaults, instead of always using the primary token.

[Joris van der Wel]

Alright, I will resubmit without that check.

I thought "current" being NULL is a valid condition because when I was
experiment with default security descriptors I noticed that current is
sometimes NULL when objects are created. I will address that in
another way when I get around to that.

On Tue, Jul 8, 2014 at 7:08 PM, Alexandre Julliard  wrote:
> Joris van der Wel  writes:
> > +/** Set the security descriptor using the current primary token for defaults. */
> > +int default_set_sd( struct object *obj, const struct security_descriptor *sd,
> > -
> >                unsigned int set_info )
> >
> > +{
> > - return set_sd_defaults_from_token( obj, sd, set_info, current ? current->process->token : NULL );
>
> There shouldn't be any need to test the current pointer for null. If
> it's needed, that probably indicates some other problem.
>
> --
> Alexandre Julliard
> [e-mail]