[3003] in java-interest
implementing a security manager
daemon@ATHENA.MIT.EDU (Anselm Baird_smith)
Thu Oct 26 05:20:49 1995
Date: Wed, 25 Oct 1995 18:43:15 +0100
From: Anselm Baird_smith <Anselm.Baird_Smith@inria.fr>
To: java-interest@java.Eng.Sun.COM
Hi,
I have two questions related to this problem:
- SecurityManager has a `void checkWrite (int fd)' method that seems
to be called when something is written to a socket. How can I
retreive any appropriate informations (eg host, port, target host,
target port) from this fd, in order to implement a reasonable
security policy. What's the idea behind this since I can already
reject most socket (and probably File) calls (e.g through
checkListen, checkConnect, etc)
- How can I access state information from the classes on the
stack. For example, if I want to implement a policy that says:
'remote code is only able to socket with the host it comes from'. In
this case, I need to retreive (probably from the exceution context)
an *instance* of the agent that is executing, in order to get the
host it came from.
Both problems could be overcome if I enforced remote code to use my
own API (e.g agentConnect()), but this is probably not the way to go.
Any docs on these (and other) problems related to implementing a
SecurityManager ?
Thanks,
Anselm.
-
This message was sent to the java-interest mailing list
Info: send 'help' to java-interest-request@java.sun.com
