[
https://issues.apache.org/jira/browse/DERBY-2109?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Martin Zaun updated DERBY-2109:
-------------------------------
Attachment: DERBY-2109-10.diff
DERBY-2109-10.stat
Please find attached patch update DERBY-2109-10 (replacing DERBY-2109-09):
a) Fixed AuthenticationTest (and other) failures; junit-all and derbyall run
fine with the patch on my machine.
b) Made the System Privileges checks only dependent upon the presence of a
SecurityManager, not upon Authentication, per derby-dev discussion "JMX meeting
system authorization (DERBY-2109 & 1387)".
c) Moved class SecurityUtil from o.a.d.security to o.a.d.iapi.security, per
suggestion by Dan.
d) Investigated an issue with non-normalized names in policy files where a
grant to user edWard would not give permissions to the authenticated users
edward or EDWARD but edWard only.
Bottomline: When evaluating permissions, the (Sun) Java Security Runtime uses
the principal names as found literally in the policy file and not as returned
by SystemPrincipal.getName() (where we could return normalized names).
As a workaround, our class SystemPrincipal could implement the non-standard
interface com.sun.security.auth.PrincipalComparator, which declares a method
implies(Subject) that allows for Principals to match Subjects based on
normalized Authorization Identifiers, e.g., a policy grant to edWard would
give permissions to all users edWard, edward, and EDWARD. But then we'd be
relying upon a non-standard Security Runtime class and behaviour. (Yet another
approach would be to construct the Subject with a Principal list having all
lower/upper case combinations of an identifier, but that's clearly not
feasible.)
I changed classes SecurityUtil and SystemPrincipal and added javadoc comments
reflecting that
- SystemPrincipal names are not Authorization Identifiers (i.e., non-delimited
identifiers handled case-insensitively by magic of SystemPrincipal);
- SecurityUtil constructs a Subject with two SystemPrincipals: one with the
literal user name and another with the normalized name.
For users who want to put specific principal grants into customized policy
files this means:
- a grant to a principal named EDWARD gives permissions to all users edward,
edWard, and EDWARD (and all other lower/upper case spellings);
- a grant to principal edWard only gives permissions to the user edWard.
In my view that's still an acceptable and documentable behaviour; but as said,
I haven't found any other feasible workaround lifting the Java Security
Runtime's literal handling of principal names.
e) Added comments to SystemPrivilegesPermissionTest.java on the use of
Subject.doAsPrivileged() versus doAs() after further experiments with the Java
Security Runtime. Bottomline: doAsPrivileged() is the only implementation
option for our purposes.
f) The patch does not reflect the latest suggestions by Dan for the shutdown
permission following a new naming scheme (wanted to get this patch out and
tested first). But will look into it and comment.
Martin
> System privileges
> -----------------
>
> Key: DERBY-2109
> URL: https://issues.apache.org/jira/browse/DERBY-2109
> Project: Derby
> Issue Type: New Feature
> Components: Security
> Affects Versions: 10.3.1.4
> Reporter: Rick Hillegas
> Assignee: Martin Zaun
> Attachments: DERBY-2109-02.diff, DERBY-2109-02.stat,
> derby-2109-03-javadoc-see-tags.diff, DERBY-2109-04.diff, DERBY-2109-04.stat,
> DERBY-2109-05and06.diff, DERBY-2109-05and06.stat, DERBY-2109-07.diff,
> DERBY-2109-07.stat, DERBY-2109-08.diff, DERBY-2109-08.stat,
> DERBY-2109-08_addendum.diff, DERBY-2109-08_addendum.stat, DERBY-2109-09.diff,
> DERBY-2109-09.stat, DERBY-2109-10.diff, DERBY-2109-10.stat,
> SystemPrivilegesBehaviour.html, systemPrivs.html, systemPrivs.html,
> systemPrivs.html, systemPrivs.html
>
>
> Add mechanisms for controlling system-level privileges in Derby. See the
> related email discussion at
> http://article.gmane.org/gmane.comp.apache.db.derby.devel/33151.
> The 10.2 GRANT/REVOKE work was a big step forward in making Derby more
> secure in a client/server configuration. I'd like to plug more client/server
> security holes in 10.3. In particular, I'd like to focus on authorization
> issues which the ANSI spec doesn't address.
> Here are the important issues which came out of the email discussion.
> Missing privileges that are above the level of a single database:
> - Create Database
> - Shutdown all databases
> - Shutdown System
> Missing privileges specific to a particular database:
> - Shutdown that Database
> - Encrypt that database
> - Upgrade database
> - Create (in that Database) Java Plugins (currently Functions/Procedures,
> but someday Aggregates and VTIs)
> Note that 10.2 gave us GRANT/REVOKE control over the following
> database-specific issues, via granting execute privilege to system
> procedures:
> Jar Handling
> Backup Routines
> Admin Routines
> Import/Export
> Property Handling
> Check Table
> In addition, since 10.0, the privilege of connecting to a database has been
> controlled by two properties (derby.database.fullAccessUsers and
> derby.database.defaultConnectionMode) as described in the security section of
> the Developer's Guide (see
> http://db.apache.org/derby/docs/10.2/devguide/cdevcsecure865818.html).
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
