Re: API change advice: Passing plan invalidation info from the rewriter into the planner?

On 13 June 2014 01:13, Stephen Frost <sfrost(at)snowman(dot)net> wrote:
> Greg, all,
>
> I will reply to the emails in detail when I get a chance but am out of town
> at a funeral, so it'll likely be delayed. I did want to echo my agreement
> for the most part with Greg and in particular...
>
> On Thursday, June 12, 2014, Gregory Smith <gregsmithpgsql(at)gmail(dot)com> wrote:
>>
>> On 6/11/14, 10:26 AM, Robert Haas wrote:
>>>
>>> Now, as soon as we introduce the concept that selecting from a table
>>> might not really mean "read from the table" but "read from the table after
>>> applying this owner-specified qual", we're opening up a whole new set of
>>> attack surfaces. Every pg_dump is an opportunity to hack somebody else's
>>> account, or at least audit their activity.
>>
>>
>> I'm in full agreement we should clearly communicate the issues around
>> pg_dump in particular, because they can't necessarily be eliminated
>> altogether without some major work that's going to take a while to finish.
>> And if the work-around is some sort of GUC for killing RLS altogether,
>> that's ugly but not unacceptable to me as a short-term fix.
>
>
> A GUC which is enable / disable / error-instead may work quiet well, with
> error-instead for pg_dump default if people really want it (there would have
> to be a way to disable that though, imv).
>
> Note that enable is default in general, disable would be for superuser only
> (or on start-up) to disable everything, and error-instead anyone could use
> but it would error instead of implementing RLS when querying an RLS-enabled
> table.
>
> This approach was suggested by an existing user testing out this RLS
> approach, to be fair, but it looks pretty sane to me as a way to address
> some of these concerns. Certainly open to other ideas and thoughts though.
>

Yeah, I was thinking something like this could work, but I would go
further. Suppose you had separate GRANTable privileges for direct
access to individual tables, bypassing RLS, e.g.

GRANT DIRECT SELECT|INSERT|UPDATE|DELETE ON table_name TO role_name

Combined with the GUC (direct_table_access, say) to request direct
access to all tables. Then with direct_table_access = true/required, a
SELECT from a table would error if the user hadn't been granted the
DIRECT SELECT privilege on all the tables referenced in the query.
Tools like pg_dump would require direct_table_access, but there might
be other levels of access that didn't error out.

I think if I were using RLS, I would definitely want/expect this level
of fine-grained control over permissions on a per-table basis, rather
than the superuser/non-superuser level of control, or having
RLS-exempt users.

Actually, given the fact that the majority of users won't be using
RLS, I would be tempted to invert the above logic and have the new
privilege be for LIMITED access (via RLS quals). So a user granted the
normal SELECT privilege would be able to bypass RLS, but a user only
granted LIMITED SELECT wouldn't.

Regards,
Dean