Similarly to my recent addition of NEON-accelerated AES, these new
implementations drop in alongside the SHA-NI ones, under a different
set of ifdefs. All the details of selection and detection are
essentially the same as they were for the AES code.
This tears out the entire previous random-pool system in sshrand.c. In
its place is a system pretty close to Ferguson and Schneier's
'Fortuna' generator, with the main difference being that I use SHA-256
instead of AES for the generation side of the system (rationale given
in comment).
The PRNG implementation lives in sshprng.c, and defines a self-
contained data type with no state stored outside the object, so you
can instantiate however many of them you like. The old sshrand.c still
exists, but in place of the previous random pool system, it's just
become a client of sshprng.c, whose job is to hold a single global
instance of the PRNG type, and manage its reference count, save file,
noise-collection timers and similar administrative business.
Advantages of this change include:
- Fortuna is designed with a more varied threat model in mind than my
old home-grown random pool. For example, after any request for
random numbers, it automatically re-seeds itself, so that if the
state of the PRNG should be leaked, it won't give enough
information to find out what past outputs _were_.
- The PRNG type can be instantiated with any hash function; the
instance used by the main tools is based on SHA-256, an improvement
on the old pool's use of SHA-1.
- The new PRNG only uses the completely standard interface to the
hash function API, instead of having to have privileged access to
the internal SHA-1 block transform function. This will make it
easier to revamp the hash code in general, and also it means that
hardware-accelerated versions of SHA-256 will automatically be used
for the PRNG as well as for everything else.
- The new PRNG can be _tested_! Because it has an actual (if not
quite explicit) specification for exactly what the output numbers
_ought_ to be derived from the hashes of, I can (and have) put
tests in cryptsuite that ensure the output really is being derived
in the way I think it is. The old pool could have been returning
any old nonsense and it would have been very hard to tell for sure.
The upcoming PRNG revamp will want to tell noise sources apart, so
that it can treat them all fairly. So I've added an extra parameter to
noise_ultralight and random_add_noise, which takes values in an
enumeration covering all the vague classes of entropy source I'm
collecting. In this commit, though, it's simply ignored.
This is in preparation for a PRNG revamp which will want to have a
well defined boundary for any given request-for-randomness, so that it
can destroy the evidence afterwards. So no more looping round calling
random_byte() and then stopping when we feel like it: now you say up
front how many random bytes you want, and call random_read() which
gives you that many in one go.
Most of the call sites that had to be fixed are fairly mechanical, and
quite a few ended up more concise afterwards. A few became more
cumbersome, such as mp_random_bits, in which the new API doesn't let
me load the random bytes directly into the target integer without
triggering undefined behaviour, so instead I have to allocate a
separate temporary buffer.
The _most_ interesting call site was in the PKCS#1 v1.5 padding code
in sshrsa.c (used in SSH-1), in which you need a stream of _nonzero_
random bytes. The previous code just looped on random_byte, retrying
if it got a zero. Now I'm doing a much more interesting thing with an
mpint, essentially scaling a binary fraction repeatedly to extract a
number in the range [0,255) and then adding 1 to it.
Mostly on the Unix side: there are lots of places the Windows code was
collecting noise that the corresponding Unix/GTK code wasn't bothering
to, such as mouse movements, keystrokes and various network events.
Also, both platforms had forgotten to collect noise when reading data
from a pipe to a local proxy process, even though in that
configuration that's morally equivalent to the network packet timings
that we'd normally be collecting from.
Another thing pointed out by ASan: new_unix_listener takes ownership
of the SockAddr you give it, so I shouldn't have been freeing it at
the end of platform_make_x11_server().
All the hash-specific state structures, and the functions that
directly accessed them, are now local to the source files implementing
the hashes themselves. Everywhere we previously used those types or
functions, we're now using the standard ssh_hash or ssh2_mac API.
The 'simple' functions (hmacmd5_simple, SHA_Simple etc) are now a pair
of wrappers in sshauxcrypt.c, each of which takes an algorithm
structure and can do the same conceptual thing regardless of what it
is.
The SSH wire protocol for tty modes corresponding to control
characters (e.g. configuring what Ctrl-Foo you can press to generate
SIGINT or SIGQUIT) specifies (RFC 4254 section 8, under VINTR, saying
'similarly for the other characters') that you should send the value
255 on the wire if you want _no_ character code to map to the action
in question.
But in the <termios.h> API, that's indicated by setting the
appropriate field of 'struct termios' to _POSIX_VDISABLE, which is a
platform-dependent value and varies between (at least) Linux and *BSD.
On the client side, Unix Plink has always known this: when it copies
the local termios settings into a struct ssh_ttymodes to be sent on
the wire, it checks for _POSIX_VDISABLE and replaces it with 255. But
uxpty.c, mapping ssh_ttymodes back to termios for Uppity's pty
sessions, wasn't making the reverse transformation.
The refactored sshaes.c gives me a convenient slot to drop in a second
hardware-accelerated AES implementation, similar to the existing one
but using Arm NEON intrinsics in place of the x86 AES-NI ones.
This needed a minor structural change, because Arm systems are often
heterogeneous, containing more than one type of CPU which won't
necessarily all support the same set of architecture features. So you
can't test at run time for the presence of AES acceleration by
querying the CPU you're running on - even if you found a way to do it,
the answer wouldn't be reliable once the OS started migrating your
process between CPUs. Instead, you have to ask the OS itself, because
only that knows about _all_ the CPUs on the system. So that means the
aes_hw_available() mechanism has to extend a tentacle into each
platform subdirectory.
The trickiest part was the nest of ifdefs that tries to detect whether
the compiler can support the necessary parts. I had successful
test-compiles on several compilers, and was able to run the code
directly on an AArch64 tablet (so I know it passes cryptsuite), but
it's likely that at least some Arm platforms won't be able to build it
because of some path through the ifdefs that I haven't been able to
test yet.
When the user clicks 'yes' to a 'weak crypto primitive' warning, and
another such warning is pending next in line, we were failing an
assertion when ssh2transport called register_dialog() for the second
warning box, because the result callback in gtkdlg.c had not called
unregister_dialog() for the previous one yet. Now that's done before
rather than after delivering the result to the dialog's client.
This is the commit that f3295e0fb _should_ have been. Yesterday I just
added some typedefs so that I didn't have to wear out my fingers
typing 'struct' in new code, but what I ought to have done is to move
all the typedefs into defs.h with the rest, and then go through
cleaning up the legacy 'struct's all through the existing code.
But I was mostly trying to concentrate on getting the test suite
finished, so I just did the minimum. Now it's time to come back and do
it better.
Taking a leaf out of the LLVM code base: this macro still includes an
assert(false) so that the message will show up in a typical build, but
it follows it up with a call to a function explicitly marked as no-
return.
So this ought to do a better job of convincing compilers that once a
code path hits this function it _really doesn't_ have to still faff
about with making up a bogus return value or filling in a variable
that 'might be used uninitialised' in the following code that won't be
reached anyway.
I've gone through the existing code looking for the assert(false) /
assert(0) idiom and replaced all the ones I found with the new macro,
which also meant I could remove a few pointless return statements and
variable initialisations that I'd already had to put in to placate
compiler front ends.
The old 'Bignum' data type is gone completely, and so is sshbn.c. In
its place is a new thing called 'mp_int', handled by an entirely new
library module mpint.c, with API differences both large and small.
The main aim of this change is that the new library should be free of
timing- and cache-related side channels. I've written the code so that
it _should_ - assuming I haven't made any mistakes - do all of its
work without either control flow or memory addressing depending on the
data words of the input numbers. (Though, being an _arbitrary_
precision library, it does have to at least depend on the sizes of the
numbers - but there's a 'formal' size that can vary separately from
the actual magnitude of the represented integer, so if you want to
keep it secret that your number is actually small, it should work fine
to have a very long mp_int and just happen to store 23 in it.) So I've
done all my conditionalisation by means of computing both answers and
doing bit-masking to swap the right one into place, and all loops over
the words of an mp_int go up to the formal size rather than the actual
size.
I haven't actually tested the constant-time property in any rigorous
way yet (I'm still considering the best way to do it). But this code
is surely at the very least a big improvement on the old version, even
if I later find a few more things to fix.
I've also completely rewritten the low-level elliptic curve arithmetic
from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c
than it is to the SSH end of the code. The new elliptic curve code
keeps all coordinates in Montgomery-multiplication transformed form to
speed up all the multiplications mod the same prime, and only converts
them back when you ask for the affine coordinates. Also, I adopted
extended coordinates for the Edwards curve implementation.
sshecc.c has also had a near-total rewrite in the course of switching
it over to the new system. While I was there, I've separated ECDSA and
EdDSA more completely - they now have separate vtables, instead of a
single vtable in which nearly every function had a big if statement in
it - and also made the externally exposed types for an ECDSA key and
an ECDH context different.
A minor new feature: since the new arithmetic code includes a modular
square root function, we can now support the compressed point
representation for the NIST curves. We seem to have been getting along
fine without that so far, but it seemed a shame not to put it in,
since it was suddenly easy.
In sshrsa.c, one major change is that I've removed the RSA blinding
step in rsa_privkey_op, in which we randomise the ciphertext before
doing the decryption. The purpose of that was to avoid timing leaks
giving away the plaintext - but the new arithmetic code should take
that in its stride in the course of also being careful enough to avoid
leaking the _private key_, which RSA blinding had no way to do
anything about in any case.
Apart from those specific points, most of the rest of the changes are
more or less mechanical, just changing type names and translating code
into the new API.
A user points out that the call to close_directory() in pscp.c's
rsource() function should have been inside rather than outside the if
statement that checks whether the directory handle is NULL. As a
result, any failed attempt to open a directory during a 'pscp -r'
recursive upload leads to a null-pointer dereference.
Moved the close_directory call to where it should be, and also
arranged to print the OS error code if the directory-open fails, by
also changing the API of open_directory to return an error string on
failure.
In the previous commit I happened to notice a %.150s in a ppl_logevent
call, which was probably an important safety precaution a couple of
decades ago when that format string was being used for an sprintf into
a fixed-size buffer, but now it's just pointless cruft.
This commit removes all printf string formatting directives with a
compile-time fixed size, with the one exception of a %.3s used to cut
out a 3-letter month name in scpserver.c. In cases where the format
string in question was already going to an arbitrary-length function
like dupprintf or ppl_logevent, that's all I've done; in cases where
there was still a fixed-size buffer, I've replaced it with a dynamic
buffer and dupprintf.
In the past, I've had a lot of macros which you call with double
parentheses, along the lines of debug(("format string", params)), so
that the inner parens protect the commas and permit the macro to treat
the whole printf-style argument list as one macro argument.
That's all very well, but it's a bit inconvenient (it doesn't leave
you any way to implement such a macro by prepending another argument
to the list), and now this code base's rules allow C99isms, I can
switch all those macros to using a single pair of parens, using the
C99 ability to say '...' in the parameter list of the #define and get
at the corresponding suffix of the arguments as __VA_ARGS__.
So I'm doing it. I've made the following printf-style macros variadic:
bpp_logevent, ppl_logevent, ppl_printf and debug.
While I'm here, I've also fixed up a collection of conditioned-out
calls to debug() in the Windows front end which were clearly expecting
a macro with a different calling syntax, because they had an integer
parameter first. If I ever have a need to condition those back in,
they should actually work now.
A long time ago, in commit 4d77b6567, I moved the generation of the
arrow-key escape sequences into a function format_arrow_key(). Mostly
the reason for that was a special purpose I had in mind at the time
which involved auto-generating the same sequences in response to
things other than a keypress, but I always thought it would be nice to
centralise a lot more of PuTTY's complicated keyboard handling in the
same way - at least the handling of the function keys and their
numerous static and dynamic config options.
In this year's general spirit of tidying up and refactoring, I think
it's finally time. So here I introduce three more centralised
functions for dealing with the numbered function keys, the small
keypad (Ins, Home, PgUp etc) and the numeric keypad. Lots of horrible
and duplicated code from the key handling functions in window.c and
gtkwin.c is now more sensibly centralised: each platform keyboard
handler concerns itself with the local format of a keyboard event and
platform-specific enumeration of key codes, and once it's decided what
the logical key press actually _is_, it hands off to the new functions
in terminal.c to generate the appropriate escape code.
Mostly this is intended to be a refactoring without functional change,
leaving the keyboard handling how it's always been. But in cases where
the Windows and GTK handlers were accidentally inconsistent, I've
fixed the inconsistency rather than carefully keeping both sides how
they were. Known consistency fixes:
- swapping the arrow keys between normal (ESC [ A) and application
(ESC O A) is now done by pressing Ctrl with them, and _not_ by
pressing Shift. That was how it was always supposed to work, and
how it's worked on GTK all along, but on Windows it's been done by
Shift as well since 2010, due to a bug at the call site of
format_arrow_key() introduced when I originally wrote that function.
- in Xterm function key mode plus application keypad mode, the /*-
keys on the numeric keypad now send ESC O {o,j,m} in place of ESC O
{Q,R,S}. That's how the Windows keyboard handler has worked all
along (it was a deliberate behaviour tweak for the Xterm-like
function key mode, because in that mode ESC O {Q,R,S} are generated
by F2-F4). But the GTK keyboard handler omitted that particular
special case and was still sending ESC O {Q,R,S} for those keys in
all application keypad modes.
- also in Xterm function key mode plus app keypad mode, we only
generates the app-keypad escape sequences if Num Lock is on; with
Num Lock off, the numeric keypad becomes arrow keys and
Home/End/etc, just as it would in non-app-keypad mode. Windows has
done this all along, but again, GTK lacked that special case.
This fixes a batch of clang-analyzer warnings of the form 'you
declared / assigned this variable and then never use it'. It doesn't
fix _all_ of them - some are there so that when I add code in the
future _it_ can use the variable without me having to remember to
start setting it - but these are the ones I thought it would make the
code better instead of worse to fix.
uxnet.c's sk_namelookup and the sorting-key construction in
pangofont_enum_fonts() were both using s[n]printf and strncpy into
buffers that had no real need to be fixed-size; format_telnet_command
and the GTK Event Log selection-data builder were doing their own
sresize loops, but now we have strbuf they can just use that and save
redoing the same work.
This is another cleanup I felt a need for while I was doing
boolification. If you define a function or variable in one .c file and
declare it extern in another, then nothing will check you haven't got
the types of the two declarations mismatched - so when you're
_changing_ the type, it's a pain to make sure you've caught all the
copies of it.
It's better to put all those extern declarations in header files, so
that the declaration in the header is also in scope for the
definition. Then the compiler will complain if they don't match, which
is what I want.
sk_startup and sk_nextaddr are entirely internal to winnet.c; nearly
all of import.c and minibidi.c's internal routines should have been
static and weren't; {read,write}_utf8 are internal to charset/utf8.c
(and didn't even need separate declarations at all); do_sftp_cleanup
is internal to psftp.c, and get_listitemheight to gtkdlg.c.
While I was editing those prototypes anyway, I've also added missing
'const' to the 'char *' passphrase parameters in import,c.
For a start, they now have different names on Windows and Unix,
reflecting their different roles: on Windows they apply escaping to
any string that's going to be used as a registry key (be it a session
name, or a host name for host key storage), whereas on Unix they're
for constructing saved-session file names in particular (and also
handle the special case of filling in "Default Settings" for NULL).
Also, they now produce output by writing to a strbuf, which simplifies
a lot of the call sites. In particular, the strbuf output idiom is
passed on to enum_settings_next, which is especially nice because its
only actual caller was doing an ad-hoc realloc loop that I can now get
rid of completely.
Thirdly, on Windows they're centralised into winmisc.c instead of
living in winstore.c, because that way Pageant can use the unescape
function too. (It was spotting the duplication there that made me
think of doing this in the first place, but once I'd started, I had to
keep unravelling the thread...)
I came across this unexplained static variable in my boolification
trawl. It seems clearly unintentional that it has only one instance
instead of one per terminal window - the code in question closely
resembles the Windows front end, and I think this must just be a
variable that I never swept up into 'inst' in the very early days when
I was making gtkwin.c out of a cloned-and-hacked window.c in the first
place.
These days it's even a bug, now that the OS X port actually does run
multiple terminal windows in the same process: if one goes into mouse
reporting mode, I'm pretty sure this would have done confusing things
to the effects of mouse actions in the other.
My normal habit these days, in new code, is to treat int and bool as
_almost_ completely separate types. I'm still willing to use C's
implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine,
no need to spell it out as blob.len != 0), but generally, if a
variable is going to be conceptually a boolean, I like to declare it
bool and assign to it using 'true' or 'false' rather than 0 or 1.
PuTTY is an exception, because it predates the C99 bool, and I've
stuck to its existing coding style even when adding new code to it.
But it's been annoying me more and more, so now that I've decided C99
bool is an acceptable thing to require from our toolchain in the first
place, here's a quite thorough trawl through the source doing
'boolification'. Many variables and function parameters are now typed
as bool rather than int; many assignments of 0 or 1 to those variables
are now spelled 'true' or 'false'.
I managed this thorough conversion with the help of a custom clang
plugin that I wrote to trawl the AST and apply heuristics to point out
where things might want changing. So I've even managed to do a decent
job on parts of the code I haven't looked at in years!
To make the plugin's work easier, I pushed platform front ends
generally in the direction of using standard 'bool' in preference to
platform-specific boolean types like Windows BOOL or GTK's gboolean;
I've left the platform booleans in places they _have_ to be for the
platform APIs to work right, but variables only used by my own code
have been converted wherever I found them.
In a few places there are int values that look very like booleans in
_most_ of the places they're used, but have a rarely-used third value,
or a distinction between different nonzero values that most users
don't care about. In these cases, I've _removed_ uses of 'true' and
'false' for the return values, to emphasise that there's something
more subtle going on than a simple boolean answer:
- the 'multisel' field in dialog.h's list box structure, for which
the GTK front end in particular recognises a difference between 1
and 2 but nearly everything else treats as boolean
- the 'urgent' parameter to plug_receive, where 1 vs 2 tells you
something about the specific location of the urgent pointer, but
most clients only care about 0 vs 'something nonzero'
- the return value of wc_match, where -1 indicates a syntax error in
the wildcard.
- the return values from SSH-1 RSA-key loading functions, which use
-1 for 'wrong passphrase' and 0 for all other failures (so any
caller which already knows it's not loading an _encrypted private_
key can treat them as boolean)
- term->esc_query, and the 'query' parameter in toggle_mode in
terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h,
but can also hold -1 for some other intervening character that we
don't support.
In a few places there's an integer that I haven't turned into a bool
even though it really _can_ only take values 0 or 1 (and, as above,
tried to make the call sites consistent in not calling those values
true and false), on the grounds that I thought it would make it more
confusing to imply that the 0 value was in some sense 'negative' or
bad and the 1 positive or good:
- the return value of plug_accepting uses the POSIXish convention of
0=success and nonzero=error; I think if I made it bool then I'd
also want to reverse its sense, and that's a job for a separate
piece of work.
- the 'screen' parameter to lineptr() in terminal.c, where 0 and 1
represent the default and alternate screens. There's no obvious
reason why one of those should be considered 'true' or 'positive'
or 'success' - they're just indices - so I've left it as int.
ssh_scp_recv had particularly confusing semantics for its previous int
return value: its call sites used '<= 0' to check for error, but it
never actually returned a negative number, just 0 or 1. Now the
function and its call sites agree that it's a bool.
In a couple of places I've renamed variables called 'ret', because I
don't like that name any more - it's unclear whether it means the
return value (in preparation) for the _containing_ function or the
return value received from a subroutine call, and occasionally I've
accidentally used the same variable for both and introduced a bug. So
where one of those got in my way, I've renamed it to 'toret' or 'retd'
(the latter short for 'returned') in line with my usual modern
practice, but I haven't done a thorough job of finding all of them.
Finally, one amusing side effect of doing this is that I've had to
separate quite a few chained assignments. It used to be perfectly fine
to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a
the 'true' defined by stdbool.h, that idiom provokes a warning from
gcc: 'suggest parentheses around assignment used as truth value'!
I think this is the full set of things that ought logically to be
boolean.
One annoyance is that quite a few radio-button controls in config.c
address Conf fields that are now bool rather than int, which means
that the shared handler function can't just access them all with
conf_{get,set}_int. Rather than back out the rigorous separation of
int and bool in conf.c itself, I've just added a similar alternative
handler function for the bool-typed ones.
This commit includes <stdbool.h> from defs.h and deletes my
traditional definitions of TRUE and FALSE, but other than that, it's a
100% mechanical search-and-replace transforming all uses of TRUE and
FALSE into the C99-standardised lowercase spellings.
No actual types are changed in this commit; that will come next. This
is just getting the noise out of the way, so that subsequent commits
can have a higher proportion of signal.
The annoying int64.h is completely retired, since C99 guarantees a
64-bit integer type that you can actually treat like an ordinary
integer. Also, I've replaced the local typedefs uint32 and word32
(scattered through different parts of the crypto code) with the
standard uint32_t.
If values are boolean, it's confusing to use & and | in place of &&
and ||. In two of these three cases it was simply a typo and I've used
the other one; in the third, it was a deliberate avoidance of short-
circuit evaluation (and commented as such), but having seen how easy
it is to make the same typo by accident, I've decided it's clearer to
just move the LHS and RHS evaluations outside the expression.
x11_char_struct returns a pointer or NULL, so while returning FALSE
would have ended up doing the right thing after macro expansion and
integer->pointer conversion, it wasn't actually how I _meant_ to spell
a failure return.
In GTK 2, this function was a new and convenient way to override the
order in which the Tab key cycled through the sub-widgets of a
container, replacing the more verbose mechanism in GTK 1 where you had
to provide a custom implementation of the "focus" method in
GtkContainerClass.
GTK 3.24 has now deprecated gtk_container_set_focus_chain(),
apparently on the grounds that that old system is what they think you
_ought_ to be doing. So I've abandoned set_focus_chain completely, and
switched back to doing it by a custom focus method for _all_ versions
of GTK, with the only slight wrinkle being that between GTK 1 and 2
the method in question moved from GtkContainer to GtkWidget (my guess
is so that an individual non-container widget can still have multiple
separately focusable UI components).
After the recent Seat and LogContext revamps, _nearly_ all the
remaining uses of the type 'Frontend' were in terminal.c, which needs
all sorts of interactions with the GUI window the terminal lives in,
from the obvious (actually drawing text on the window, reading and
writing the clipboard) to the obscure (minimising, maximising and
moving the window in response to particular escape sequences).
All of those functions are now provided by an abstraction called
TermWin. The few remaining uses of Frontend after _that_ are internal
to a particular platform directory, so as to spread the implementation
of that particular kind of Frontend between multiple source files; so
I've renamed all of those so that they take a more specifically named
type that refers to the particular implementation rather than the
general abstraction.
So now the name 'Frontend' no longer exists in the code base at all,
and everywhere one used to be used, it's completely clear whether it
was operating in one of Frontend's three abstract roles (and if so,
which), or whether it was specific to a particular implementation.
Another type that's disappeared is 'Context', which used to be a
typedef defined to something different on each platform, describing
whatever short-lived resources were necessary to draw on the terminal
window: the front end would provide a ready-made one when calling
term_paint, and the terminal could request one with get_ctx/free_ctx
if it wanted to do proactive window updates. Now that drawing context
lives inside the TermWin itself, because there was never any need to
have two of those contexts live at the same time.
(Another minor API change is that the window-title functions - both
reading and writing - have had a missing 'const' added to their char *
parameters / return values.)
I don't expect this change to enable any particularly interesting new
functionality (in particular, I have no plans that need more than one
implementation of TermWin in the same application). But it completes
the tidying-up that began with the Seat and LogContext rework.
This adds the server side of the SSH-2 keyboard-interactive protocol,
and the pair of very similar SSH-1 methods AUTH_TIS and AUTH_CCARD
(which basically differ only in message numbers, and each involve a
single challenge from the server and a response from the user).
We now show the --help output if invoked with no arguments, and the
help text also includes a big safety warning in the hope of stopping
anyone from mistaking this for a _secure_ SSH server implementation.
While I'm here, the errors now all use appname[] in place of
constantly repeating the program name. (Not because I anticipate a
change right now, but if nothing else, it makes things easier moving
errors out into shared source files or between applications.)
Naturally, there's one really glaring goof I find out instants after
'git push'! If Pageant starts a watchdog subprocess which will wait
until the main process terminates and then clean up the socket, then
it had better not have that subprocess keep the standard I/O handles
open, or else commands like eval $(pageant -X) won't terminate.
I've applied the same fix in the X11 socket creation, though I think
it's less critical there.
Unlike the traditional Unix SSH server organisation, the SFTP server
is built into the same process as all the rest of the code. sesschan.c
spots a subsystem request for "sftp", and responds to it by
instantiating an SftpServer object and swapping out its own vtable for
one that talks to it.
(I rather like the idea of an object swapping its own vtable for a
different one in the middle of its lifetime! This is one of those
tricks that would be absurdly hard to implement in a 'proper' OO
language, but when you're doing vtables by hand in C, it's no more
difficult than any other piece of ordinary pointer manipulation. As
long as the methods in both vtables expect the same physical structure
layout, it doesn't cause a problem.)
The SftpServer object doesn't deal directly with SFTP packet formats;
it implements the SFTP server logic in a more abstract way, by having
a vtable method for each SFTP request type with an appropriate
parameter list. It sends its replies by calling methods in another
vtable called SftpReplyBuilder, which in the normal case will write an
SFTP reply packet to send back to the client. So SftpServer can focus
more or less completely on the details of a particular filesystem API
- and hence, the implementation I've got lives in the unix source
directory, and works directly with file descriptors and struct stat
and the like.
(One purpose of this abstraction layer is that I may well want to
write a second dummy implementation, for test-suite purposes, with
completely controllable behaviour, and now I have a handy place to
plug it in in place of the live filesystem.)
In between sesschan's parsing of the byte stream into SFTP packets and
the SftpServer object, there's a layer in the new file sftpserver.c
which does the actual packet decoding and encoding: each request
packet is passed to that, which pulls the fields out of the request
packet and calls the appropriate method of SftpServer. It also
provides the default SftpReplyBuilder which makes the output packet.
I've moved some code out of the previous SFTP client implementation -
basic packet construction code, and in particular the BinarySink/
BinarySource marshalling fuinction for fxp_attrs - into sftpcommon.c,
so that the two directions can share as much as possible.
This server is NOT SECURE! If anyone is reading this commit message,
DO NOT DEPLOY IT IN A HOSTILE-FACING ENVIRONMENT! Its purpose is to
speak the server end of everything PuTTY speaks on the client side, so
that I can test that I haven't broken PuTTY when I reorganise its
code, even things like RSA key exchange or chained auth methods which
it's hard to find a server that speaks at all.
(For this reason, it's declared with [UT] in the Recipe file, so that
it falls into the same category as programs like testbn, which won't
be installed by 'make install'.)
Working title is 'Uppity', partly for 'Universal PuTTY Protocol
Interaction Test Yoke', but mostly because it looks quite like the
word 'PuTTY' with part of it reversed. (Apparently 'test yoke' is a
very rarely used term meaning something not altogether unlike 'test
harness', which is a bit of a stretch, but it'll do.)
It doesn't actually _support_ everything I want yet. At the moment,
it's a proof of concept only. But it has most of the machinery
present, and the parts it's missing - such as chained auth methods -
should be easy enough to add because I've built in the required
flexibility, in the form of an AuthPolicy object which can request
them if it wants to. However, the current AuthPolicy object is
entirely trivial, and will let in any user with the password "weasel".
(Another way in which this is not a production-ready server is that it
also has no interaction with the OS's authentication system. In
particular, it will not only let in any user with the same password,
but it won't even change uid - it will open shells and forwardings
under whatever user id you started it up as.)
Currently, the program can only speak the SSH protocol on its standard
I/O channels (using the new FdSocket facility), so if you want it to
listen on a network port, you'll have to run it from some kind of
separate listening program similar to inetd. For my own tests, I'm not
even doing that: I'm just having PuTTY spawn it as a local proxy
process, which also conveniently eliminates the risk of anyone hostile
connecting to it.
The bulk of the actual code reorganisation is already done by previous
commits, so this change is _mostly_ just dropping in a new set of
server-specific source files alongside the client-specific ones I
created recently. The remaining changes in the shared SSH code are
numerous, but all minor:
- a few extra parameters to BPP and PPL constructors (e.g. 'are you
in server mode?'), and pass both sets of SSH-1 protocol flags from
the login to the connection layer
- in server mode, unconditionally send our version string _before_
waiting for the remote one
- a new hook in the SSH-1 BPP to handle enabling compression in
server mode, where the message exchange works the other way round
- new code in the SSH-2 BPP to do _deferred_ compression the other
way round (the non-deferred version is still nicely symmetric)
- in the SSH-2 transport layer, some adjustments to do key derivation
either way round (swapping round the identifying letters in the
various hash preimages, and making sure to list the KEXINITs in the
right order)
- also in the SSH-2 transport layer, an if statement that controls
whether we send SERVICE_REQUEST and wait for SERVICE_ACCEPT, or
vice versa
- new ConnectionLayer methods for opening outgoing channels for X and
agent forwardings
- new functions in portfwd.c to establish listening sockets suitable
for remote-to-local port forwarding (i.e. not under the direction
of a Conf the way it's done on the client side).
If the child process's standard input is provided by a pipe that's
separate from its output channels, we can - and should - honour a
request to cause that process to receive input EOF, by closing the
output end of that pipe.
As usual, we do this by setting a pending-EOF flag and calling
try_write, to ensure that any buffered output data is sent before the
pipe actually closes.
Not every "session" channel in SSH allocates a pty at all, of course,
and so I'll need a way to run a subprocess without doing so. The
simplest approach seems to be to expand uxpty's remit so that the pty
is optional: now it can open either a pty or a set of pipes for
stdin/out/err, according to an option provided to pty_backend_create.
(It amuses me that without this option I'd have an SSH server which is
incapable of _not_ honouring the "pty-req" channel request. That's
normally the easy part!)
This breaks the previous one-to-one coupling between pty backend
instances and file descriptors passed to uxsel, which I was using to
look up the Pty structure in a tree234 indexed by fd when an uxsel
notification came back. So now each Pty structure contains a
collection of subobjects of a new type PtyFd, and _those_ are what's
stored in the fd-indexed tree.
Another awkward part is that uxsel_set is not incremental: the rwx
flags you pass to it completely supersede the previous set for that
file descriptor, so I had to set up the logic that decides whether
we're trying to read or write each fd in a way that can cope equally
well with the fd aliasing another one (if it's the pty master) or not
(if there are three completely separate pipes).
The SS_SIGFOO family are implemented by sending a signal directly to
the pid of the immediate child process.
I had had the vague idea that it might be more desirable to send the
specified signal to the foreground process group in the tty. That way,
you'd be able to SIGINT (say) the foreground job in a shell session,
and return to the shell _prompt_ without terminating the whole
session, and you could do this in an emergency even if the job was a
full-screen application which had configured termios so that no
keystroke generated SIGINT.
But as far as I can see there's no actual way to do that. I wasn't
able to find any ioctl or termios call to send a signal to a pty's
foreground pgrp, and you can't even do it manually via kill(2) because
first you'd have to find out what the pgrp id _is_, and according to
the man pages, you can only call tcgetpgrp on the slave end of the pty
and even then only if it's your controlling terminal.
So SS_SIGFOO goes to the child process, because that's the only place
I can find that I _can_ send it to sensibly.
SS_BRK translates to tcsendbreak, of course (though I haven't actually
seen any effect of calling this on a pty master, not even if I set
PARMRK on the slave end which by my understanding _ought_ to show me
when break events occur).
This will be applied to the pty's termios settings at creation time,
superseding the default settings uxpty has always used. It works by
including the new sshttymodes.h with TTYMODES_LOCAL_ONLY defined, so
that modes not supported by a particular Unix system are automatically
quietly ignored.
Of course, a struct ssh_ttymodes always has the option of representing
"please make no change to the defaults", and of course, that's
precisely what is done by the one that pty_init constructs for clients
that aren't calling pty_backend_create directly.
The function that does the main pty setup is now called
pty_backend_create(), and has an API better suited to uxpty in
particular than the standard backend_init() virtual constructor. It
leaves off a load of standard parameters to backend_init() which
aren't really relevant to this backend, and it adds the 'argv'
parameter to pass in a split-up command line, which is unique to it.
The old creation function still exists, as a tiny wrapper that calls
the new pty_backend_create. And that version still gets the argv
parameter from the process-global variable pty_argv[], so the call
sites in pterm haven't had to change for this.
This will make it possible to instantiate a pty backend directly from
the SSH server code, without having to do anything really excessively
cumbersome to pass in a subcommand in the form of pre-split argv. (And
I'll add a few more specialist parameters to the new function shortly.)
Simon Tatham
Pavel I. Kryukov