This fixes#4866, an issue with msfvenom not properly handling special
cases with generic payloads. So the story behind this fix is that
we have these two problems:
Problem 1: The current payload selection design relies on the payload
module in order to set the platform and arch. Almost all MSF payloads
contain a default platform and arch, however, the bind and reverse
generic payloads don't.
Problem 2: By default, Msf::Payload::Generic also explicitly sets the
PLATFORM and ARCH datastore options to nil. So there is no way the
payload generator can figure out what platform and arch to use.
As a result of these problems, msfvenom will actually end up getting
a Msf::Module::Platform as the default platform, which doesn't
actually represent any valid platform we can use (such as
Msf::Module::Platform::Windows). And the first item of ARCH_ALL for
the arch.
In addition, msfvenom has these two arguments that the user can use:
--platform and --arch. In most cases, these arguments are used more
like checks than actually setting anything. Because remember:
Framework's payload selector retreives the platform & arch from the
module (trusted), not the user input (untrusted). But from the user's
perspective it's impossible to know this.
After experimenting different ways to fix this, I came up with this
patch. It feels sort of more like a hack than a real fix, but as
far as I can tell, this is the best you can get unless you want to
redesign generic payload selection.
* Stageless payloads weren't adding brackets around IPv6 hosts.
* Staged HTTP handler was using an undefined function to check for IPv6
addresses when host header overriding was disabled.
Session expiry, comms timeout, retry total/wait are all now part of all
of the meterpreter payloads as these are going to be used for
maintaining access with resiliency and will aim for consistency across
the payload types.
This code adjusts the bind_tcp stager for x86 so that the listener
socket isn't close for meterpreter payloads. This means that meterpreter
can make an educated guess as to whether or not the payload was a bind
or tcp payload, and from there can attempt to establish communications
in the same way as before should something break along the way.
Some simple adjustments to the x64 meterpreter stage as well, but more
to come here.
Methods now require a hash. I went with the hash because 1) that's what
we seem to use everywhere else, and 2) I couldn't get the new keyword
arguments working nicely with the block syntax (I'm clearly stupid).
There is an old-looking bug where the deletekey command opens the key it tries
to delete, then deletes the same key name again. Basically, it uses the wrong
level of indirection.
If the session doesn't have a payload UUID we now generate one as best
we can. This code will probably go away when TCP related transports have
had the UUID stuf baked in.
joev