Status Report on SE-PostgreSQL

From: KaiGai Kohei <kaigai(at)kaigai(dot)gr(dot)jp>
To: "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Status Report on SE-PostgreSQL
Date: 2009-01-17 04:21:46
Message-ID: 49715CDA.5090506@kaigai.gr.jp
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

I also think it is a good idea to summarize current status of
SE-PostgreSQL, as Simon Riggs doing on his works.

The current revision of SE-PostgreSQL is 1425, available here:

[1/5] http://sepgsql.googlecode.com/files/sepostgresql-sepgsql-8.4devel-3-r1425.patch
[2/5] http://sepgsql.googlecode.com/files/sepostgresql-utils-8.4devel-3-r1425.patch
[3/5] http://sepgsql.googlecode.com/files/sepostgresql-policy-8.4devel-3-r1425.patch
[4/5] http://sepgsql.googlecode.com/files/sepostgresql-docs-8.4devel-3-r1425.patch
[5/5] http://sepgsql.googlecode.com/files/sepostgresql-tests-8.4devel-3-r1425.patch

We had various kind of comments, feature requests and discussions during
previous/current commit fest, then whole of them are already included.

Currently, we have no open issues here.

As I summarized as follows, we had many discussions about its design
issues mainly, so my patch set has been updated to support them.
I believe we should move to detailed-reviews to merge the feature any
time now, since we should aware of v8.4 schedule.

I really would like folks to help/volunteer reviewing the patches, please!

* CommitFest:Nov
- Simon Riggs requires a new GUC option to turn on/off row-level security
labeling to reduce storage comsumption, then updated as follows:
http://archives.postgresql.org/message-id/492691A8.8030103@ak.jp.nec.com
- Bruce Momjian suggested Row-level database ACLs to be compiled in default.
- Discussions for default compile options: PostgreSQL doesn't prefer compile
time option to turn on/off features, except for platform specific one.
SE-PostgreSQL is indeed platform specific feature. But, it makes other
issue that need mutually-exclusive enhanced security feature.
We concluded it as follows:
- All configurable features should be compiled within a single binary.
- Both of DAC and MAC should be available simultaneously in row-level also.
- DAC is hardwired, and we allow users to choose an enhanced security feature.
- I updated the patch set to support both of Row-level database ACLs and
an enhanced security feature (SELinux) simultaneously. ('08/12/17)
http://archives.postgresql.org/message-id/4948B6BD.1050402@ak.jp.nec.com
- Robert Haas concerned about Stephen Frost's column-level privileges has
a trouble, so it's unclear whether it can get merged into v8.4.
- I also worked for his patch, then it got being ready for commit:
http://archives.postgresql.org/message-id/20090116045825.GY4656@tamriel.snowman.net
- Alvaro Herrera suggested "static inline" is not preferable.

* CommitFest:Sep
- Peter Eisentraut commented about its design specifications:
http://archives.postgresql.org/message-id/48D03953.6000308@gmx.net
- The hot issues were lack of fine-grained access controls in SQL-level,
and covert channels with row-level controls.
- We finally made agreement to provide platform independent row-level controls,
and explicit documentation about covert channels in PK/FK constraints.
No one didn't want to apply polyinstantiation idea.
- Simon Riggs requires wiki article to introduce SE-PostgreSQL.
http://wiki.postgresql.org/wiki/SEPostgreSQL
- Patch set was updated to support Row-level database ACLs
http://archives.postgresql.org/message-id/48F46606.4080207@ak.jp.nec.com

* CommitFest:Jul
- The patch set got documentation/testcases.
- Peter Eisentraut commented about some of items:
http://archives.postgresql.org/message-id/200807071739.58428.peter_e@gmx.net
- Then, these items are updated:
http://archives.postgresql.org/message-id/48773188.6000809@ak.jp.nec.com

* CommitFest:May
- First patch set for v8.4 were proposed.
- Tom Lane gave us various items to be improved.
http://archives.postgresql.org/message-id/3275.1210019965@sss.pgh.pa.us
- I had a presentation at PGcon2008 ottawa.
http://sepgsql.googlecode.com/files/PGCON20080523.pdf

* Prior phase
- First proposal of PGACE security framework, but I didn't know it was
just after the date of feature freeze in v8.3. So, it was suggested
to wait for v8.4 development cycle. ('07/04/17)
- 8.2.x based SE-PostgreSQL announced. ('07/09/04)
- SE-PostgreSQL package got merged into Fedora Project. ('07/11/08)
- 8.3.x based SE-PostgreSQL announced. ('08/03/08)

Thanks,
--
KaiGai Kohei <kaigai(at)kaigai(dot)gr(dot)jp>

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Simon Riggs 2009-01-17 08:54:38 MemoryContextSwitchTo (Re: [GENERAL] Autovacuum daemon terminated by signal 11)
Previous Message Alvaro Herrera 2009-01-17 04:14:33 Re: Autovacuum daemon terminated by signal 11