Re: [HACKERS] SE-PostgreSQL Updated Revision (r1460)
Sorry, I attached incorrect patch file. It is the correct one. KaiGai Kohei wrote: Robert, The attached patch is a draft to replace RedHat/Fedora RPM centric expressions, to add a reference at Database Roles and Privileges chapter and a bit cleanups for the latest revision (r1467). In the previous revision, it noted users to check the version of RPM package, but the revised one notes actually required features. The version number is rewritten as a hint. What is your opinion? Thanks, KaiGai Kohei wrote: Robert Haas wrote: On Fri, Jan 23, 2009 at 12:30 AM, KaiGai Kohei kai...@ak.jp.nec.com wrote: The patch set of SE-PostgreSQL and related stuff were updated (r1460). [1/5] http://sepgsql.googlecode.com/files/sepostgresql-sepgsql-8.4devel-3-r1460.patch [2/5] http://sepgsql.googlecode.com/files/sepostgresql-utils-8.4devel-3-r1460.patch [3/5] http://sepgsql.googlecode.com/files/sepostgresql-policy-8.4devel-3-r1460.patch [4/5] http://sepgsql.googlecode.com/files/sepostgresql-docs-8.4devel-3-r1460.patch [5/5] http://sepgsql.googlecode.com/files/sepostgresql-tests-8.4devel-3-r1460.patch KaiGai - I read through your docs patch tonight and did some copy editing. Please see the attached patches, which I hope you will find helpful. I have attached my suggested changes both as a patch against v1460 (sepostgresql-docs-rmh-vs-1460.gz) and also as patch against CVS HEAD (sepostgresql-docs-rmh-vs-cvs-head), since I am not sure which is easier for you. I have a couple of general comments about the documentation: Thanks your feedbacks! I basically applied your fixes as is, expect for the following items: - You replaced ! Its providing access controls are not _bypassable_ for any clients ... by ! The access controls implemented by SE-PostgrSQL may not be _biased_ even ... I wanted to express it is unavoidable here, so I changed as: ! The access controls implemented by SE-PostgrSQL may not be _bypassed_ even ... - I found a typo: MAC is described as MAc. And, I have a question about documentation manner. - You represented getpeercon() function as a system call. But, it is actually a wrapper function of getsockopt(2) system call, so the getpeercon(3) is not a system call strictly. Is it necessary to represent these stuffs strictly correct? (Thus, I wrote it as API in the r1460.) 1. The docs as written are very Red Hat-centric, even to the point of making reference to specific versions of Red Hat RPMs. I think that the community will find this unacceptable, as Red Hat is certainly not the only SELinux-enabled distribution and I presume that we want to support all of them to an equal degree. I guess you pointed out about: 1. The Requirement section in Build and Installation assumes RedHat/Fedora's RPM package and its version number. 2. The security context and security policy used to explanation assumes specific security policy. 3. Labeled IPsec seciton points to RedHatEL4 Security Guide, and it assumes the racoon's configuration files are deployed as RPM package doing. About 1, is it necessary to rip the RPM specific version number and replace it as: selinux-policy which includes SE-PostgreSQL related stuffs. About 2, SELinux community provides its default security policy, and distributor's policy (including RedHat's one) is a derivative of the default policy. It is developed independent from distributor's cycle. http://oss.tresys.com/projects/refpolicy http://oss.tresys.com/repos/refpolicy/trunk/policy/modules/services/postgresql.te You can find some of sepgsql_ identifiers in postgresql.te. All the appeared identifiers are upstreamed, so these are not Red Hat specific. About 3, If it rips the link to Red Hat and does not assume specific path of racoon.conf, the explnation become neutral. 2. Some of the information that is documented here properly belongs in other sections of the documentation. For example, the information about GUCs clearly belongs somewhere in the section on server configuration where all of the other GUCs are documented, not in a separate sections about SE-PostgreSQL. These explanations are moved to Security and Authentication section in Chapter 18. Server Configuration. I suspect that all of the information about row-level ACLs should be ripped out of security.sgml and inserted into an appropriate portion of the Database Roles and Privileges chapter, leaving this file to talk just about SE-PostgreSQL. It is indeed an aspect of row-level ACLs. However, it is also a feature on PGACE framework, same as SE-PostgreSQL. An idea is to put a reference to indicate the row-level ACLs section on Database Roles and Privileges chapter, like: PostgreSQL has an enhancement of database roles and privileges mechanism which allows to database ACLs in row-level granuality. See, xref ... for more details. What do you think? 3. It seems to me that the analogy between SQL DAC and Unix user/group DAC is
Re: [HACKERS] SE-PostgreSQL Updated Revision (r1460)
I basically applied your fixes as is, expect for the following items: - You replaced ! Its providing access controls are not _bypassable_ for any clients ... by ! The access controls implemented by SE-PostgrSQL may not be _biased_ even I wanted to express it is unavoidable here, so I changed as: ! The access controls implemented by SE-PostgrSQL may not be _bypassed_ even ... Good catch, my mistake. - I found a typo: MAC is described as MAc. Also my mistake. And, I have a question about documentation manner. - You represented getpeercon() function as a system call. But, it is actually a wrapper function of getsockopt(2) system call, so the getpeercon(3) is not a system call strictly. Is it necessary to represent these stuffs strictly correct? (Thus, I wrote it as API in the r1460.) Oh, OK. It sounds a little awkward to me to refer to it as an API. Perhaps we could just refer to it as getpeercon(3) and not call it either an API or a system call. About 2, SELinux community provides its default security policy, and distributor's policy (including RedHat's one) is a derivative of the default policy. It is developed independent from distributor's cycle. http://oss.tresys.com/projects/refpolicy http://oss.tresys.com/repos/refpolicy/trunk/policy/modules/services/postgresql.te OK, I wasn't aware of that. I think perhaps you could spell this out a little more in the docs so people understand that there is an upstream version which includes SE-PostgreSQL support from version whatever. I suspect that all of the information about row-level ACLs should be ripped out of security.sgml and inserted into an appropriate portion of the Database Roles and Privileges chapter, leaving this file to talk just about SE-PostgreSQL. It is indeed an aspect of row-level ACLs. However, it is also a feature on PGACE framework, same as SE-PostgreSQL. An idea is to put a reference to indicate the row-level ACLs section on Database Roles and Privileges chapter, like: Actually, I think this should probably be broken up into three sections. All of the stuff about how PGACE is not very interesting to anyone who isn't a developer, so it should be moved to someplace under Internals. I would suggest just adding a new chapter to the end of that section, after How the Planner Uses Statistics. The database ACL stuff properly belongs in the Database Roles and Privileges section, and needs to be moved there, not just a cross-reference. The discussion of enhanced security and SE-PostgreSQL is another new chapter, probably immediately following Database Roles and Privileges. I would suggest calling it Enhanced Security and SE-PostgreSQL. For example, the section that defines MAC and DAC is a ways down in the document, but you use those terms a whole bunch of times before defining them. I'm not 100% sure that we even want to be defining MAC and DAC in our documentation, since those are general industry terms that are not PostgreSQL-specific. But if we are going to define them then we should try to do so in the clearest way possible. I can add the definitions of terms. However, it is unclear whether PostgreSQL documentation should include them, or not. For example, wikipedia has enough explanation for their generam meanings. http://en.wikipedia.org/wiki/Discretionary_Access_Control http://en.wikipedia.org/wiki/Mandatory_Access_Control It seems to me Discretionary Access Control (DAC) is an enough key to search its meaning. I agree. I think you should go through and rip out all of the definitions and explanations of what these terms mean, and just use them in the appropriate context. I think in general that the current documentation spends far too much time explaining what SE-PostgreSQL is and not enough time discussing the issues that are likely to come up when you're actually using it. For example, it seems to me that anyone who has any interest in using SE-PostgreSQL to control access to functions will need a much more complicated policy than what you are proposing here, and there doesn't seem to be much discussion of that issue. I'm not really looking for specific examples of how to build a policy so much as general considerations that you should keep in mind when trying to prevent information leakage via functions. I'm glad to see your help. I'll pay my efforts for documentations also. But English is not my mother language, so any suggestions are helpful for me. Well, your English is certainly better than my Japanese... ...Robert -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
Re: [HACKERS] SE-PostgreSQL Updated Revision (r1460)
Robert Haas wrote: On Fri, Jan 23, 2009 at 12:30 AM, KaiGai Kohei kai...@ak.jp.nec.com wrote: The patch set of SE-PostgreSQL and related stuff were updated (r1460). [1/5] http://sepgsql.googlecode.com/files/sepostgresql-sepgsql-8.4devel-3-r1460.patch [2/5] http://sepgsql.googlecode.com/files/sepostgresql-utils-8.4devel-3-r1460.patch [3/5] http://sepgsql.googlecode.com/files/sepostgresql-policy-8.4devel-3-r1460.patch [4/5] http://sepgsql.googlecode.com/files/sepostgresql-docs-8.4devel-3-r1460.patch [5/5] http://sepgsql.googlecode.com/files/sepostgresql-tests-8.4devel-3-r1460.patch KaiGai - I read through your docs patch tonight and did some copy editing. Please see the attached patches, which I hope you will find helpful. I have attached my suggested changes both as a patch against v1460 (sepostgresql-docs-rmh-vs-1460.gz) and also as patch against CVS HEAD (sepostgresql-docs-rmh-vs-cvs-head), since I am not sure which is easier for you. I have a couple of general comments about the documentation: Thanks your feedbacks! I basically applied your fixes as is, expect for the following items: - You replaced ! Its providing access controls are not _bypassable_ for any clients ... by ! The access controls implemented by SE-PostgrSQL may not be _biased_ even ... I wanted to express it is unavoidable here, so I changed as: ! The access controls implemented by SE-PostgrSQL may not be _bypassed_ even ... - I found a typo: MAC is described as MAc. And, I have a question about documentation manner. - You represented getpeercon() function as a system call. But, it is actually a wrapper function of getsockopt(2) system call, so the getpeercon(3) is not a system call strictly. Is it necessary to represent these stuffs strictly correct? (Thus, I wrote it as API in the r1460.) 1. The docs as written are very Red Hat-centric, even to the point of making reference to specific versions of Red Hat RPMs. I think that the community will find this unacceptable, as Red Hat is certainly not the only SELinux-enabled distribution and I presume that we want to support all of them to an equal degree. I guess you pointed out about: 1. The Requirement section in Build and Installation assumes RedHat/Fedora's RPM package and its version number. 2. The security context and security policy used to explanation assumes specific security policy. 3. Labeled IPsec seciton points to RedHatEL4 Security Guide, and it assumes the racoon's configuration files are deployed as RPM package doing. About 1, is it necessary to rip the RPM specific version number and replace it as: selinux-policy which includes SE-PostgreSQL related stuffs. About 2, SELinux community provides its default security policy, and distributor's policy (including RedHat's one) is a derivative of the default policy. It is developed independent from distributor's cycle. http://oss.tresys.com/projects/refpolicy http://oss.tresys.com/repos/refpolicy/trunk/policy/modules/services/postgresql.te You can find some of sepgsql_ identifiers in postgresql.te. All the appeared identifiers are upstreamed, so these are not Red Hat specific. About 3, If it rips the link to Red Hat and does not assume specific path of racoon.conf, the explnation become neutral. 2. Some of the information that is documented here properly belongs in other sections of the documentation. For example, the information about GUCs clearly belongs somewhere in the section on server configuration where all of the other GUCs are documented, not in a separate sections about SE-PostgreSQL. These explanations are moved to Security and Authentication section in Chapter 18. Server Configuration. I suspect that all of the information about row-level ACLs should be ripped out of security.sgml and inserted into an appropriate portion of the Database Roles and Privileges chapter, leaving this file to talk just about SE-PostgreSQL. It is indeed an aspect of row-level ACLs. However, it is also a feature on PGACE framework, same as SE-PostgreSQL. An idea is to put a reference to indicate the row-level ACLs section on Database Roles and Privileges chapter, like: PostgreSQL has an enhancement of database roles and privileges mechanism which allows to database ACLs in row-level granuality. See, xref ... for more details. What do you think? 3. It seems to me that the analogy between SQL DAC and Unix user/group DAC is mentioned far too many times, and there are other cases where information is repeated as well. I think it might help to reorganize the document a bit so that you introduce concepts in the right order. Indeed, it was redundant explanation. Thanks for youe edit. For example, the section that defines MAC and DAC is a ways down in the document, but you use those terms a whole bunch of times before defining them. I'm not 100% sure that we even want to be defining MAC and DAC in our documentation, since those are general industry
Re: [HACKERS] SE-PostgreSQL Updated Revision (r1460)
On Fri, Jan 23, 2009 at 12:30 AM, KaiGai Kohei kai...@ak.jp.nec.com wrote: The patch set of SE-PostgreSQL and related stuff were updated (r1460). [1/5] http://sepgsql.googlecode.com/files/sepostgresql-sepgsql-8.4devel-3-r1460.patch [2/5] http://sepgsql.googlecode.com/files/sepostgresql-utils-8.4devel-3-r1460.patch [3/5] http://sepgsql.googlecode.com/files/sepostgresql-policy-8.4devel-3-r1460.patch [4/5] http://sepgsql.googlecode.com/files/sepostgresql-docs-8.4devel-3-r1460.patch [5/5] http://sepgsql.googlecode.com/files/sepostgresql-tests-8.4devel-3-r1460.patch KaiGai - I read through your docs patch tonight and did some copy editing. Please see the attached patches, which I hope you will find helpful. I have attached my suggested changes both as a patch against v1460 (sepostgresql-docs-rmh-vs-1460.gz) and also as patch against CVS HEAD (sepostgresql-docs-rmh-vs-cvs-head), since I am not sure which is easier for you. I have a couple of general comments about the documentation: 1. The docs as written are very Red Hat-centric, even to the point of making reference to specific versions of Red Hat RPMs. I think that the community will find this unacceptable, as Red Hat is certainly not the only SELinux-enabled distribution and I presume that we want to support all of them to an equal degree. 2. Some of the information that is documented here properly belongs in other sections of the documentation. For example, the information about GUCs clearly belongs somewhere in the section on server configuration where all of the other GUCs are documented, not in a separate sections about SE-PostgreSQL. I suspect that all of the information about row-level ACLs should be ripped out of security.sgml and inserted into an appropriate portion of the Database Roles and Privileges chapter, leaving this file to talk just about SE-PostgreSQL. 3. It seems to me that the analogy between SQL DAC and Unix user/group DAC is mentioned far too many times, and there are other cases where information is repeated as well. I think it might help to reorganize the document a bit so that you introduce concepts in the right order. For example, the section that defines MAC and DAC is a ways down in the document, but you use those terms a whole bunch of times before defining them. I'm not 100% sure that we even want to be defining MAC and DAC in our documentation, since those are general industry terms that are not PostgreSQL-specific. But if we are going to define them then we should try to do so in the clearest way possible. Overall, I would say there is a fair amount of work left to be done to get this documentation up to par, but it's a good start and I hope that the attached patches and suggestions will be helpful. ...Robert sepostgresql-docs-rmh-vs-1460.patch.gz Description: GNU Zip compressed data sepostgresql-docs-rmh-vs-cvs-head.patch.gz Description: GNU Zip compressed data -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
[HACKERS] SE-PostgreSQL Updated Revision (r1460)
The patch set of SE-PostgreSQL and related stuff were updated (r1460).
[1/5]
http://sepgsql.googlecode.com/files/sepostgresql-sepgsql-8.4devel-3-r1460.patch
[2/5]
http://sepgsql.googlecode.com/files/sepostgresql-utils-8.4devel-3-r1460.patch
[3/5]
http://sepgsql.googlecode.com/files/sepostgresql-policy-8.4devel-3-r1460.patch
[4/5]
http://sepgsql.googlecode.com/files/sepostgresql-docs-8.4devel-3-r1460.patch
[5/5]
http://sepgsql.googlecode.com/files/sepostgresql-tests-8.4devel-3-r1460.patch
I reviewed the patch set by myself, and updated the following items.
However, I would like other hackers to review the code in honesty.
SE-PostgreSQL need any volunteers to review and comment the patch set.
Please give us your support!
List of updates:
- Rebased to the latest CVS HEAD, which includes the column-level
privileges based on the SQL-standard.
(The previous r1425 conflicts in some points.)
- Security policy (sepostgresql-devel.pp) was updated to fit both of
Fedora 10 and rawhide. Test cases are also modified to care the new
security policy.
- Cleanup: NUM_SELINUX_CATALOG was replaced by lengthof() macro to
avoid code duplications.
- Cleanup: sepgsqlCheckEmbeddedProcedure() is renamed to
sepgsqlCheckProcedureInstall() due to its confusable naming.
- Add a new permission: db_procedure:{install}
It enables to prevent malicious user-defined functions are installed
as a part of operators, conversions, types and so on.
The default policy allows to install functions labeled as sepgsql_proc_t
only, as an implementation of these facilities.
Meanwhile, functions defined by unprivileged users are labeled as
user_sepgsql_proc_t in default, and it is not allowed to install as
an operator and so on.
If DBA want to install user-defined functions for the purpose, he has to
confirm its harmless and relabel it to sepgsql_proc_t at first.
In the previous revision, it checked db_procedure:{execute} here,
but it is not enough actually, because unprivilged user is allowed to
execute self defined function.
- Code revising: The previous revision always denied required permissions,
when the kernel does not define them within its security policy.
But it can make unexpected behavior when we work SE-PostgreSQL on
a system with legacy security policy which lacks a part of newly
added permissions.
The revised one simply allows actions when these are undefined.
- Fixbug: It required superfluous permissions when we try to update
security_label system column but it does not change anything actually.
For example:
UPDATE t SET security_label = security_label;
This query does not change security_label, so we don't need to check
db_tuple:{relabelfrom} permission here.
It is obvious we cannot know what tuples are actually relabeled on
sepgsqlExecScan(), so any permission checks for write-operations are
moved to sepgsqlHeapTuple(Insert|Update|Delete) hooks.
- Fixbug: when we update pg_largeobject system catalog by hand, it has
a possibility to create/drop specific largeobject, so we add a check
on db_blob:{create drop} when pg_largeobject.loid is modified by
UPDATE statement.
For example:
UPDATE pg_largeobject SET loid = loid::int + 10 WHERE loid = 1234;
It is theoretically same as dropping a largeobject with loid:1234 and
creating a largeobject with loid:1244.
- Fixbug: Tome Lane pointed out a matter when a whole-row-reference on
the relation with RTE_JOIN makes crash at the Column-Level Privileges
thread. This revision added a special care for the situation.
It recursively walks on refered JoinExpr and picks up its sources to
check permission to them.
- Code revising: T_SEvalItemRelation and T_SEvalItemAttribute nodes are
integrated into T_SelinuxEvalItem node. In the previous revision,
it simply chains all appeared tables and columns as a list of obsoleted
node on Query-pgaceItem. But it has a trend the length of list grows long.
T_SelinuxEvalItem contains required permissions on a table and an array of
permissions for columns. It enables to keep the length of the list minimum.
Related stuffs in sepgsql/proxy.c is also revised.
- addEvalRelation() / addEvalAttribute() enhanced to handle
T_SelinuxEvalItem.
- Functions to handle inheritance tables and whole-row-reference are clearly
sorted out. expandEvalItemInheritance() handles inheritance tables, and
expandEvalItemWholeRowRefs() handles whole-row-reference.
- Add a hook: pgaceExecuteTruncate()
The previous revision checks permissions on truncated tables and tuples
on pgaceProcessUtility(), but this approach need to extract all the target
including cascaded ones, so it made code duplication.
The new hook is deployed on ExecuteTruncate() and delivers a list of already
opened relations with AccessExclusiveLock.
A new sepgsqlExecuteTruncate() checks needed permission on the hook.
- Cleanup: sepgsqlTupleName() always copied an identifier of tuple
