Instantiate payload modules so parameter validation occurs

[bcook-r7]

Calling .new on payload modules does not perform parameter validation, leading to a number cached sizes based on invalid parameters. Most notably, normalization does not occur either, which makes all OptBool params default to true. This caused the windows/adduser to be larger with its cached size parameter than it would be if you generate it with msfvenom or from framework directly.

Validation steps

• The following command works without validation errors:
  ./msfconsole -qx 'use exploit/windows/smb/ms08_067_netapi; set payload windows/adduser'
• tab completion with set payload using ms08_67_netapi should show 141 entries, not 119.
• specs should pass consistently

[jvazquez-r7]

:?

[jvazquez-r7]

Is it expected?

[bcook-r7]

Determining if a payload is :dynamic is a little precarious. It currently just generates the payload a few times and sees if the size changes, but this is not always reliable. We probably need a way to force a manual override. I probably got unlucky with the random seed.

[jvazquez-r7]

oook, expected so :) just curious, thanks!

[jvazquez-r7]

Looks like travis ask for ./spec/support/shared/examples/payload_cached_size_is_consistent update :)

[bcook-r7]

Ah, you're right - I need to move the default datastore options into the library code for this instead of coding it into the tool directly. Thanks.

[bcook-r7]

With the latest changes, some payloads marked as 'dynamic' go back to being statically sized. There were a few payloads that did something like cmd = datastore['CMD'] << "\x00", which has a side-effect of modifying the datastore option as well. This caused the payload to change size each time it was instantiated with the same datastore :P.

[jvazquez-r7]

It is an awesome work :) thanks @bcook-r7, handling it in a while!

[jvazquez-r7]

HEAD is now at 6b1e911... Instantiate payload modules so parameter validation occurs
aus-mac-0916:metasploit-framework jvazquez$ ./msfconsole -qx 'use exploit/windows/smb/ms08_067_netapi; set payload windows/adduser'
payload => windows/adduser
msf exploit(ms08_067_netapi) > set payload
Display all 141 possibilities? (y or n)

travis is green and also pass locally.

[jvazquez-r7]

I don't think it has sense as instance variable. Indeed, I don't think we can't create PayloadCachedSize even :). I'm going to switch to a constant before landing if you don't mind. Looks definitely like a constant. Feel free to fix me if I'm wrong, just reviewing changes :)

[jvazquez-r7]

Landed after minor fixes here: f25a5da

Thanks @bcook-r7, this is an awesome patch!

[bcook-r7]

Muchas gracias @jvazquez-r7