Re: [HACKERS] Simplify ACL handling for large objects and removal of superuser() checks

On Sat, Nov 11, 2017 at 9:01 AM, Michael Paquier
<michael(dot)paquier(at)gmail(dot)com> wrote:
> On Sat, Nov 11, 2017 at 12:50 AM, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>> On Fri, Nov 10, 2017 at 10:34 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
>>> I think as far as that goes, we can just change to "Therefore, by default
>>> their use is restricted ...". Then I suggest adding a <caution> para
>>> after that, with wording along the lines of
>>>
>>> It is possible to GRANT use of server-side lo_import and lo_export to
>>> non-superusers, but careful consideration of the security implications
>>> is required. A malicious user of such privileges could easily parlay
>>> them into becoming superuser (for example by rewriting server
>>> configuration files), or could attack the rest of the server's file
>>> system without bothering to obtain database superuser privileges as
>>> such. Access to roles having such privilege must therefore be guarded
>>> just as carefully as access to superuser roles. Nonetheless, if use
>>> of server-side lo_import or lo_export is needed for some routine task,
>>> it's safer to use a role of this sort than full superuser privilege,
>>> as that helps to reduce the risk of damage from accidental errors.
>>
>> +1. That seems like great language to me.
>
> +1. Not convinced that mentioning wrappers is worth the complication.
> Experienced admins likely already know such matters.

For archives' sake, doc improvements are committed as of 6d77652.
--
Michael