Potential XSS on mark_safe function

[hanchl]

When using make_safe function, it is easy to cause XSS attacks. However, there are a lot of make_safe function uses in the horizon code, such as using the dashboard interface to obtain instance information, using the render function for server-side rendering etc.. Should we consider adding keyword filtering to prevent attacks?
Examples for related code:

File: horizon\horizon\forms\fields.py
235 output.append('</select>')
236 return mark_safe('\n'.join(output))
237

File: horizon\openstack_dashboard\dashboards\project\instances\tables.py
1185 '</span>').format(help_tooltip, icon_classes)
1186 return mark_safe(locked_status)
1187