<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Inline below.<br>
</p>
<div class="moz-cite-prefix">On 26/06/2021 1:11 pm, Peter Firmstone
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:ed9503e3-0fcf-1f66-f147-4e0ed7e96f48@zeus.net.au">
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
<p>One more proposed change inline:<br>
</p>
<div class="moz-cite-prefix">On 26/06/2021 12:58 pm, Peter
Firmstone wrote:<br>
</div>
<blockquote type="cite"
cite="mid:9084f553-7440-de48-2e03-2447ebd280c5@zeus.net.au">
<meta http-equiv="Content-Type" content="text/html;
charset=UTF-8">
<p>Summary of Proposed Changes:</p>
<ol>
<li>GuardFactory & GuardFactorySpi to provide hooks for
authorization checks without SecurityManager or Policy.
(Note GuardFactory should never return null and instead
return a no-op Guard that hotspot can optimize out.<br>
</li>
<li>Existing Permission implementations to be obtained using
GuardFactorySpi implementations, until their removal. Note
that when SecurityManager is stubbed out and Permission
implementations are deprecated for removal, these should no
longer be provided by default, but instead need to be
enabled by entries in the java.security config file, in
preparation for their removal.<br>
</li>
<li>JDK code, no longer call Permission implementations
directly, instances obtained using GuardFactory, only when
enabled in the java.security configuration file.<br>
</li>
<li>Threads (system and virtual) updated to use a singleton
*unprivileged* AccessControlContext, instead of inherited
AccessControlContext, this is more appropriate for
Executors, the original inherited context was designed
before Executors were introduced.</li>
<li>Deprecation for removal of all Permission implementations
from the JDK platform. The existing implementations of
Permission introduce unnecessary complexity; they lack
sufficient flexibility resulting in a proliferation of
Permission grants required in policy files and some make
blocking network calls.<br>
</li>
<li>Introduce a system property to change AccessController
default behaviour, disable the stack walk by default, but
allow it to be re-enabled with a system property, replace
the stack walk array result of ProtectionDomains with an
*unprivileged* AccessControlContext, the
SubjectDomainCombiner can replace it with a,
AccessControlContext containing a single element array,
containing one ProtectionDomain with Principals. <br>
</li>
<li>AccessController::doPrivileged erases the DomainCombiner
by default, deprecate these methods, retain
doPrivilegedWithCombiner methods that preserve the
SubjectDomainCombiner. Developers should replace their
doPrivileged methods with doPrivilegedWithCombiner</li>
</ol>
</blockquote>
</blockquote>
<p><br>
</p>
<p>Just thinking out loud, it's possible someone might want to do
perform some task without privileges enabled, that is without the
Subject's principal's. In a system that grants privileges to
code and principals, this is generally unnecessary, as grants are
made to the combination of code and principals. However while
using the doPrivileged methods is possible, to remove privileges,
it would be better to provide an AccessController::doUnprivileged
method instead, which erase the DomainCombiner and use an
*unprivileged* AccessControlContext.<br>
</p>
<p>Since the doPrivileged methods are utilised by other methods in
AccessController, they should be made private when finally
deprecated for removal.</p>
<p>I have also just noticed a bug in
AccessController.AccHolder.innocuousAcc.</p>
<p>The innocuous AccessControlContext, is intended to have no
permission, hence it is constructed using the two argument
ProtectionDomain constructor, which causes ProtectionDomain to not
consult the Policy.</p>
<p>However, if a user obtains this ProtectionDomain and asks the
Policy for the ProtectionDomain's permission's by calling
Policy::getPermissions(ProtectionDomain), the Policy will return
AllPermission.<br>
</p>
<p>It is generally understood that a ProtectionDomain with a null
CodeSource is a system ProtectionDomain loaded by the bootstrap
ClassLoader.</p>
<p>I propose that innocuous AccessControlContext instead be given a
ProtectionDomain, with a non-null CodeSource, which has a null
URL. This is also considered by the Policy to be unprivileged.<br>
</p>
<br>
<blockquote type="cite"
cite="mid:ed9503e3-0fcf-1f66-f147-4e0ed7e96f48@zeus.net.au">
<blockquote type="cite"
cite="mid:9084f553-7440-de48-2e03-2447ebd280c5@zeus.net.au">
<ol>
<li>Deprecate for removal the CodeSource::implies method.</li>
<li>Give unique ProtectionDomain's with a meaninful CodeSource
to Java modules mapped to the boot loader, rather than using
a Shared ProtectionDomain with a null CodeSource.<br>
</li>
</ol>
</blockquote>
<p> 10. Deprecate for removal AccessController::checkPermission
and AccessControlContext::checkPermission methods.</p>
<p> 11. Undeprecate AccessController, AccessControlContext,
DomainCombiner, SubjectDomainCombiner and Subject::doAs methods,
while deprecating for removal methods stated in items above.<br>
</p>
<blockquote type="cite"
cite="mid:9084f553-7440-de48-2e03-2447ebd280c5@zeus.net.au">
<p>To clarify what an *unprivileged* AccessControlContext is:</p>
<blockquote>
<p>An instance of AccessControlContext, that contains a single
element array, containing a ProtectionDomain, with a non
null CodeSource, containing a null URL.</p>
</blockquote>
<p>Retention of AccessController, AccessControlContext,
DomainCombiner and SubjectDomainCombiner and Subject::doAs
methods.</p>
<p>Stubbing of SecurityManager and Policy, for runtime backward
compatibility. Update ProtectionDomain::implies method, to
*not* consult with the Policy. Note it's possible to get
access to the ProtectionDomain array contained within
AccessControlContext using a DomainCombiner.<br>
</p>
<p>This is backward compatible with existing usages of JAAS and
least painful method of transition for all concerned as well
as allowing complete flexibility of implementation.</p>
<p>Regards,</p>
<p>Peter Firmstone.<br>
</p>
<div class="moz-cite-prefix">On 25/06/2021 3:59 pm, Peter
Firmstone wrote:<br>
</div>
<blockquote type="cite"
cite="mid:2f315680-1cdb-1694-34a3-95312bf42ca7@zeus.net.au">Thanks
Dalibor, <br>
<br>
Would targeting Java 18 be practical? <br>
<br>
Also it won't take long to code a prototype, just not sure of
the process. <br>
<br>
Cheers, <br>
<br>
Peter. <br>
<br>
<br>
On 24/06/2021 9:30 pm, Dalibor Topic wrote: <br>
<blockquote type="cite">On 24.06.2021 04:24, Peter Firmstone
wrote: <br>
<blockquote type="cite">Thanks Andrew, <br>
<br>
For the simple case, of replacing the SecurityManager
stack walk, one could use reflection. <br>
<br>
Thank you for also confirming that is not possible (or at
least very unlikely) to add a GuardBuilder to Java 8, the
proposal is for JDK code to use a provider mechanism, to
intercept permission checks, so custom authentication
layers can be implemented, this could be accepted in
future versions of Java, but not existing. As it is said,
there is no harm in asking. <br>
</blockquote>
<br>
Generally speaking, adding any public APIs to a platform
release after its specification has been published, is
always going to be a very tall order involving the JCP,
among other things. It's not really worth it, when other
technical solutions, such as multi-release JARs, already
exist, that alleviate the necessity. <br>
<br>
cheers, <br>
dalibor topic <br>
<br>
</blockquote>
</blockquote>
</blockquote>
<pre class="moz-signature" cols="72">--
Regards,
Peter Firmstone
0498 286 363
Zeus Project Services Pty Ltd.</pre>
</blockquote>
<pre class="moz-signature" cols="72">
</pre>
</body>
</html>