encodelib.py is a Python library which implements some handy SSH-2
encoding primitives; samplekex.py uses that to fabricate the start of
an SSH connection, up to the point where key exchange totally fails
its crypto.
The idea is that you adapt samplekex.py to construct initial-kex
sequences with particular properties, in order to test robustness and
security fixes that affect the initial-kex sequence. For example, I
used an adaptation of this to test the Diffie-Hellman range check
that's just gone into 0.64.
(cherry picked from commit 12d5b00d62)
To understand the handle leak bug that I fixed in git commit
7549f2da40, I had to think fairly hard
to remind myself what all this code was doing, which means the
comments weren't good enough. Expanded and rewritten some of them in
the hope that things will be clearer next time.
(cherry picked from commit a87a14ae0f)
Cherry-picker's notes: this apparently pointless commit is required on
this branch because it's a dependency of the rather less pointless
9fec2e7738.
The absence of these could have prevented sensitive private key
information from being properly cleared out of memory that PuTTY tools
had finished with.
Thanks to Patrick Coleman for spotting this and sending a patch.
We incremented buf by a few bytes, so we must decrement the
corresponding length by the same amount, or else makekey() could
overrun.
Thanks to Patrick Coleman for the patch.
Florent Daigniere of Matta points out that RFC 4253 actually
_requires_ us to refuse to accept out-of-range values, though it isn't
completely clear to me why this should be a MUST on the receiving end.
Matta considers this to be a security vulnerability, on the grounds
that if a server should accidentally send an obviously useless value
such as 1 then we will fail to reject it and agree a key that an
eavesdropper could also figure out. Their id for this vulnerability is
MATTA-2015-002.
Bare string exceptions aren't supported any more.
Patch by Will Aoki, plus a backward compatibility tweak from Colin Watson.
Seen working with Python 2.4.3 and 2.7.6.
If (say) a read handle returns EOF, and its gotdata function responds
by calling handle_free(), then we want the handle to have already had
its defunct flag set so that the handle can be destroyed. Otherwise
handle_free will set the 'done' flag to ask the subthread to
terminate, and then sit and wait for it to say it's done so -
forgetting that it signalled termination already by returning EOF, and
hence will not be responding to that signal.
Ditto for write errors on write handles, though that should happen
less often.
The code for cleaning up handle structures works by the main thread
asking the per-handle subthread to shut down by means of setting its
'done' flag, and then once the subthread signals back through its
event object that it's done so, the main thread frees all its
resources and removes the event object from the list of things being
checked in the program's event loop.
But read threads were not sending back that final event acknowledging
a request to shut down, so their event objects were never being
cleaned up.
Bug spotted by Ronald Weiss.
It would be rare to have a host keypair in .ppk format or on a client
machine to load into PuTTYgen, and it might confuse people into thinking
they are required to do so.
It tries to use the local username as the remote username if it has no
better ideas, but the presence of Default Settings would defeat this,
even if it had no username set. Reported by Jonathan Amery.
We were checking the return value of CreateThread for validity, but
not keeping it to free afterwards if it _was_ valid. Also, we weren't
closing ctx->event in the valid case either. Patch due to Tim Kosse.
If we use getaddrinfo to translate the source IP address into a
sockaddr, then we need to freeaddrinfo the returned data later. Patch
due to Tim Kosse.
I don't think anyone has ever actually called it that, colloquially
_or_ formally, and if anyone ever did (in a bug report, say) I'd
probably have to stop and think to work out what they meant. It's
universally called Plink, and should be officially so as well :-)
This one spotted in the old-fashioned way, by actually attempting a
Plink raw connection and wondering why it didn't seem to be reading
from standard input! Turns out 'bufsize' is uninitialised until the
first send, which can inhibit any stdin reading if it gets a large
enough nonsense value.
mkfiles.pl was giving a couple of annoying perl warnings, because some
makefile_extra strings were never set by Recipe. We already have the
&def function to convert undefs into "" for this reason, but weren't
using it everywhere. Now I think we are.
If a sharing downstream asks for an auth method we don't understand,
we should send them CHANNEL_FAILURE *and then stop processing*. Ahem.
(Spotted while examining this code in the course of Coverity-related
fixes, but not itself a Coverity-found problem.)
If (Msg)WaitForMultipleObjects returns WAIT_TIMEOUT, we expect 'next'
to have been initialised. This can occur without having called
run_timers(), if a toplevel callback was pending, so we can't expect
run_timers to have reliably initialised 'next'.
I'm not actually convinced this could have come up in either of the
affected programs (Windows PSFTP and Plink), due to the list of things
toplevel callbacks are currently used for, but it certainly wants
fixing anyway for the future.
Spotted by Coverity.
Namely, any ldisc that you send actual data through should have a
terminal attached, because the ldisc editing/echoing system is
designed entirely for use with a terminal. The only time you can have
an ldisc with no terminal is when it's only ever used by the backend
to report changes to the front end in edit/echo status, e.g. by Unix
Plink.
Coverity spotted an oddity in ldisc_send which after a while I decided
would never have actually caused a problem, but OTOH I agree that it
was confusing, so now hopefully it's less so.
'p += strcspn' returns p always non-NULL and sometimes pointing at \0,
as opposed to 'p = strchr' which returns p sometimes non-NULL and
never pointing at \0. Test the pointer after the call accordingly.
Thanks, Coverity.
This ought to happen in ssh_do_close alongside the code that shuts
down other local listening things like port forwardings, for the same
obvious reason. In particular, we should get through this _before_ we
put up a modal dialog box telling the user what just went wrong with
the SSH connection, so that further sessions started while that box is
active don't try futilely to connect to the not-really-listening
zombie upstream.
Changing it can't have any useful effect, since we have strictly
enforced that the host key used for rekeys is the same as the first key
exchange since b8e668c.
We should NULL out mac after freeing it, so that the cleanup code
doesn't try to free it again; also if the final key creation fails, we
should avoid freeing ret->comment when we're going to go to that same
cleanup code which will free 'comment' which contains the same pointer.
Thanks to Christopher Staite for pointing these out.
