2001-01-24 14:08:20 +00:00
|
|
|
/*
|
|
|
|
|
* ldisc.c: PuTTY line discipline. Sits between the input coming
|
|
|
|
|
* from keypresses in the window, and the output channel leading to
|
|
|
|
|
* the back end. Implements echo and/or local line editing,
|
|
|
|
|
* depending on what's currently configured.
|
|
|
|
|
*/
|
|
|
|
|
|
1999-11-09 12:05:34 +00:00
|
|
|
#include <stdio.h>
|
1999-11-22 10:07:24 +00:00
|
|
|
#include <ctype.h>
|
2014-11-22 10:37:14 +00:00
|
|
|
#include <assert.h>
|
1999-11-09 12:05:34 +00:00
|
|
|
|
|
|
|
|
#include "putty.h"
|
2002-10-22 16:11:33 +00:00
|
|
|
#include "terminal.h"
|
2021-09-13 09:14:33 +00:00
|
|
|
|
|
|
|
|
struct Ldisc_tag {
|
|
|
|
|
Terminal *term;
|
|
|
|
|
Backend *backend;
|
|
|
|
|
Seat *seat;
|
|
|
|
|
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
/*
|
|
|
|
|
* When the backend is not reporting true from sendok(), terminal
|
|
|
|
|
* input that comes here is stored in this bufchain instead. When
|
|
|
|
|
* the backend later decides it wants session input, we empty the
|
|
|
|
|
* queue in ldisc_check_sendok_callback(), passing its contents on
|
|
|
|
|
* to the backend. Before then, we also provide data from this
|
|
|
|
|
* queue to term_get_userpass_input() via ldisc_get_input_token(),
|
|
|
|
|
* to be interpreted as user responses to username and password
|
|
|
|
|
* prompts during authentication.
|
|
|
|
|
*
|
|
|
|
|
* Unfortunately, the data stored in this queue is not all of the
|
|
|
|
|
* same type: our output to the backend consists of both raw bytes
|
|
|
|
|
* sent to backend_send(), and also session specials such as
|
|
|
|
|
* SS_EOL and SS_EC. So we have to encode our queued data in a way
|
|
|
|
|
* that can represent both.
|
|
|
|
|
*
|
|
|
|
|
* The encoding is private to this source file, so we can change
|
|
|
|
|
* it if necessary and only have to worry about the encode and
|
|
|
|
|
* decode functions here. Currently, it is:
|
|
|
|
|
*
|
|
|
|
|
* - Bytes other than 0xFF are stored literally.
|
|
|
|
|
* - The byte 0xFF itself is stored as 0xFF 0xFF.
|
|
|
|
|
* - A session special (code, arg) is stored as 0xFF, followed by
|
|
|
|
|
* a big-endian 4-byte integer containing code, followed by
|
|
|
|
|
* another big-endian 4-byte integer containing arg.
|
|
|
|
|
*
|
|
|
|
|
* (This representation relies on session special codes being at
|
|
|
|
|
* most 0xFEFFFFFF when represented in 32 bits, so that the first
|
|
|
|
|
* byte of the 'code' integer can't be confused with the 0xFF
|
|
|
|
|
* followup byte indicating a literal 0xFF, But since session
|
|
|
|
|
* special codes are defined by an enum counting up from zero, and
|
|
|
|
|
* there are only a couple of dozen of them, that shouldn't be a
|
|
|
|
|
* problem! Even so, just in case, an assertion checks that at
|
|
|
|
|
* encode time.)
|
|
|
|
|
*/
|
|
|
|
|
bufchain input_queue;
|
|
|
|
|
|
|
|
|
|
IdempotentCallback input_queue_callback;
|
|
|
|
|
prompts_t *prompts;
|
|
|
|
|
|
2021-09-13 09:14:33 +00:00
|
|
|
/*
|
|
|
|
|
* Values cached out of conf.
|
|
|
|
|
*/
|
|
|
|
|
bool telnet_keyboard, telnet_newline;
|
|
|
|
|
int protocol, localecho, localedit;
|
|
|
|
|
|
|
|
|
|
char *buf;
|
|
|
|
|
size_t buflen, bufsiz;
|
|
|
|
|
bool quotenext;
|
|
|
|
|
};
|
2002-10-26 10:16:19 +00:00
|
|
|
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
#define ECHOING (ldisc->localecho == FORCE_ON || \
|
|
|
|
|
(ldisc->localecho == AUTO && \
|
2019-06-17 19:49:46 +00:00
|
|
|
(backend_ldisc_option_state(ldisc->backend, LD_ECHO))))
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
#define EDITING (ldisc->localedit == FORCE_ON || \
|
|
|
|
|
(ldisc->localedit == AUTO && \
|
2019-06-17 19:49:46 +00:00
|
|
|
(backend_ldisc_option_state(ldisc->backend, LD_EDIT))))
|
1999-11-09 12:05:34 +00:00
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
static void c_write(Ldisc *ldisc, const void *buf, int len)
|
2001-05-06 14:35:20 +00:00
|
|
|
{
|
New abstraction 'Seat', to pass to backends.
This is a new vtable-based abstraction which is passed to a backend in
place of Frontend, and it implements only the subset of the Frontend
functions needed by a backend. (Many other Frontend functions still
exist, notably the wide range of things called by terminal.c providing
platform-independent operations on the GUI terminal window.)
The purpose of making it a vtable is that this opens up the
possibility of creating a backend as an internal implementation detail
of some other activity, by providing just that one backend with a
custom Seat that implements the methods differently.
For example, this refactoring should make it feasible to directly
implement an SSH proxy type, aka the 'jump host' feature supported by
OpenSSH, aka 'open a secondary SSH session in MAINCHAN_DIRECT_TCP
mode, and then expose the main channel of that as the Socket for the
primary connection'. (Which of course you can already do by spawning
'plink -nc' as a separate proxy process, but this would permit it in
the _same_ process without anything getting confused.)
I've centralised a full set of stub methods in misc.c for the new
abstraction, which allows me to get rid of several annoying stubs in
the previous code. Also, while I'm here, I've moved a lot of
duplicated modalfatalbox() type functions from application main
program files into wincons.c / uxcons.c, which I think saves
duplication overall. (A minor visible effect is that the prefixes on
those console-based fatal error messages will now be more consistent
between applications.)
2018-10-11 18:58:42 +00:00
|
|
|
seat_stdout(ldisc->seat, buf, len);
|
2000-10-20 13:51:46 +00:00
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
static int plen(Ldisc *ldisc, unsigned char c)
|
2001-05-06 14:35:20 +00:00
|
|
|
{
|
2002-10-26 10:16:19 +00:00
|
|
|
if ((c >= 32 && c <= 126) || (c >= 160 && !in_utf(ldisc->term)))
|
2019-09-08 19:29:00 +00:00
|
|
|
return 1;
|
1999-11-09 12:05:34 +00:00
|
|
|
else if (c < 128)
|
2019-09-08 19:29:00 +00:00
|
|
|
return 2; /* ^x for some x */
|
2004-05-22 10:36:50 +00:00
|
|
|
else if (in_utf(ldisc->term) && c >= 0xC0)
|
2019-09-08 19:29:00 +00:00
|
|
|
return 1; /* UTF-8 introducer character
|
|
|
|
|
* (FIXME: combining / wide chars) */
|
2004-05-22 10:36:50 +00:00
|
|
|
else if (in_utf(ldisc->term) && c >= 0x80 && c < 0xC0)
|
2019-09-08 19:29:00 +00:00
|
|
|
return 0; /* UTF-8 followup character */
|
1999-11-09 12:05:34 +00:00
|
|
|
else
|
2019-09-08 19:29:00 +00:00
|
|
|
return 4; /* <XY> hex representation */
|
1999-11-09 12:05:34 +00:00
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
static void pwrite(Ldisc *ldisc, unsigned char c)
|
2001-05-06 14:35:20 +00:00
|
|
|
{
|
2004-05-22 10:36:50 +00:00
|
|
|
if ((c >= 32 && c <= 126) ||
|
2019-09-08 19:29:00 +00:00
|
|
|
(!in_utf(ldisc->term) && c >= 0xA0) ||
|
|
|
|
|
(in_utf(ldisc->term) && c >= 0x80)) {
|
|
|
|
|
c_write(ldisc, &c, 1);
|
1999-11-09 12:05:34 +00:00
|
|
|
} else if (c < 128) {
|
2019-09-08 19:29:00 +00:00
|
|
|
char cc[2];
|
|
|
|
|
cc[1] = (c == 127 ? '?' : c + 0x40);
|
|
|
|
|
cc[0] = '^';
|
|
|
|
|
c_write(ldisc, cc, 2);
|
1999-11-09 12:05:34 +00:00
|
|
|
} else {
|
2019-09-08 19:29:00 +00:00
|
|
|
char cc[5];
|
|
|
|
|
sprintf(cc, "<%02X>", c);
|
|
|
|
|
c_write(ldisc, cc, 4);
|
1999-11-09 12:05:34 +00:00
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
Convert a lot of 'int' variables to 'bool'.
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'!
2018-11-02 19:23:19 +00:00
|
|
|
static bool char_start(Ldisc *ldisc, unsigned char c)
|
2004-05-22 10:36:50 +00:00
|
|
|
{
|
|
|
|
|
if (in_utf(ldisc->term))
|
2019-09-08 19:29:00 +00:00
|
|
|
return (c < 0x80 || c >= 0xC0);
|
2004-05-22 10:36:50 +00:00
|
|
|
else
|
2019-09-08 19:29:00 +00:00
|
|
|
return true;
|
2004-05-22 10:36:50 +00:00
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
static void bsb(Ldisc *ldisc, int n)
|
2001-05-06 14:35:20 +00:00
|
|
|
{
|
1999-11-09 12:05:34 +00:00
|
|
|
while (n--)
|
2019-09-08 19:29:00 +00:00
|
|
|
c_write(ldisc, "\010 \010", 3);
|
1999-11-09 12:05:34 +00:00
|
|
|
}
|
|
|
|
|
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
static void ldisc_input_queue_callback(void *ctx);
|
|
|
|
|
|
2000-03-11 14:03:04 +00:00
|
|
|
#define CTRL(x) (x^'@')
|
2001-05-09 15:12:26 +00:00
|
|
|
#define KCTRL(x) ((x^'@') | 0x100)
|
2000-03-11 14:03:04 +00:00
|
|
|
|
New abstraction 'Seat', to pass to backends.
This is a new vtable-based abstraction which is passed to a backend in
place of Frontend, and it implements only the subset of the Frontend
functions needed by a backend. (Many other Frontend functions still
exist, notably the wide range of things called by terminal.c providing
platform-independent operations on the GUI terminal window.)
The purpose of making it a vtable is that this opens up the
possibility of creating a backend as an internal implementation detail
of some other activity, by providing just that one backend with a
custom Seat that implements the methods differently.
For example, this refactoring should make it feasible to directly
implement an SSH proxy type, aka the 'jump host' feature supported by
OpenSSH, aka 'open a secondary SSH session in MAINCHAN_DIRECT_TCP
mode, and then expose the main channel of that as the Socket for the
primary connection'. (Which of course you can already do by spawning
'plink -nc' as a separate proxy process, but this would permit it in
the _same_ process without anything getting confused.)
I've centralised a full set of stub methods in misc.c for the new
abstraction, which allows me to get rid of several annoying stubs in
the previous code. Also, while I'm here, I've moved a lot of
duplicated modalfatalbox() type functions from application main
program files into wincons.c / uxcons.c, which I think saves
duplication overall. (A minor visible effect is that the prefixes on
those console-based fatal error messages will now be more consistent
between applications.)
2018-10-11 18:58:42 +00:00
|
|
|
Ldisc *ldisc_create(Conf *conf, Terminal *term, Backend *backend, Seat *seat)
|
2001-05-06 14:35:20 +00:00
|
|
|
{
|
2018-09-11 14:02:59 +00:00
|
|
|
Ldisc *ldisc = snew(Ldisc);
|
2002-10-26 10:16:19 +00:00
|
|
|
|
|
|
|
|
ldisc->buf = NULL;
|
|
|
|
|
ldisc->buflen = 0;
|
|
|
|
|
ldisc->bufsiz = 0;
|
Convert a lot of 'int' variables to 'bool'.
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'!
2018-11-02 19:23:19 +00:00
|
|
|
ldisc->quotenext = false;
|
2002-10-26 10:16:19 +00:00
|
|
|
|
2018-09-11 15:23:38 +00:00
|
|
|
ldisc->backend = backend;
|
2002-10-26 10:16:19 +00:00
|
|
|
ldisc->term = term;
|
New abstraction 'Seat', to pass to backends.
This is a new vtable-based abstraction which is passed to a backend in
place of Frontend, and it implements only the subset of the Frontend
functions needed by a backend. (Many other Frontend functions still
exist, notably the wide range of things called by terminal.c providing
platform-independent operations on the GUI terminal window.)
The purpose of making it a vtable is that this opens up the
possibility of creating a backend as an internal implementation detail
of some other activity, by providing just that one backend with a
custom Seat that implements the methods differently.
For example, this refactoring should make it feasible to directly
implement an SSH proxy type, aka the 'jump host' feature supported by
OpenSSH, aka 'open a secondary SSH session in MAINCHAN_DIRECT_TCP
mode, and then expose the main channel of that as the Socket for the
primary connection'. (Which of course you can already do by spawning
'plink -nc' as a separate proxy process, but this would permit it in
the _same_ process without anything getting confused.)
I've centralised a full set of stub methods in misc.c for the new
abstraction, which allows me to get rid of several annoying stubs in
the previous code. Also, while I'm here, I've moved a lot of
duplicated modalfatalbox() type functions from application main
program files into wincons.c / uxcons.c, which I think saves
duplication overall. (A minor visible effect is that the prefixes on
those console-based fatal error messages will now be more consistent
between applications.)
2018-10-11 18:58:42 +00:00
|
|
|
ldisc->seat = seat;
|
2002-10-26 10:16:19 +00:00
|
|
|
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
bufchain_init(&ldisc->input_queue);
|
|
|
|
|
|
|
|
|
|
ldisc->prompts = NULL;
|
|
|
|
|
ldisc->input_queue_callback.fn = ldisc_input_queue_callback;
|
|
|
|
|
ldisc->input_queue_callback.ctx = ldisc;
|
|
|
|
|
ldisc->input_queue_callback.queued = false;
|
|
|
|
|
bufchain_set_callback(&ldisc->input_queue, &ldisc->input_queue_callback);
|
|
|
|
|
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
ldisc_configure(ldisc, conf);
|
|
|
|
|
|
2002-10-26 10:16:19 +00:00
|
|
|
/* Link ourselves into the backend and the terminal */
|
|
|
|
|
if (term)
|
2019-09-08 19:29:00 +00:00
|
|
|
term->ldisc = ldisc;
|
2018-09-11 15:23:38 +00:00
|
|
|
if (backend)
|
|
|
|
|
backend_provide_ldisc(backend, ldisc);
|
2002-10-26 10:16:19 +00:00
|
|
|
|
|
|
|
|
return ldisc;
|
|
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
void ldisc_configure(Ldisc *ldisc, Conf *conf)
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
{
|
2018-10-29 19:57:31 +00:00
|
|
|
ldisc->telnet_keyboard = conf_get_bool(conf, CONF_telnet_keyboard);
|
|
|
|
|
ldisc->telnet_newline = conf_get_bool(conf, CONF_telnet_newline);
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
ldisc->protocol = conf_get_int(conf, CONF_protocol);
|
|
|
|
|
ldisc->localecho = conf_get_int(conf, CONF_localecho);
|
|
|
|
|
ldisc->localedit = conf_get_int(conf, CONF_localedit);
|
|
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
void ldisc_free(Ldisc *ldisc)
|
2003-01-15 23:30:21 +00:00
|
|
|
{
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
bufchain_clear(&ldisc->input_queue);
|
2003-01-15 23:30:21 +00:00
|
|
|
if (ldisc->term)
|
2019-09-08 19:29:00 +00:00
|
|
|
ldisc->term->ldisc = NULL;
|
2018-09-11 15:23:38 +00:00
|
|
|
if (ldisc->backend)
|
|
|
|
|
backend_provide_ldisc(ldisc->backend, NULL);
|
2003-01-15 23:30:21 +00:00
|
|
|
if (ldisc->buf)
|
2019-09-08 19:29:00 +00:00
|
|
|
sfree(ldisc->buf);
|
free_prompts: deal with a reference from an Ldisc.
In a GUI app, when interactive userpass input begins, the Ldisc
acquires a reference to a prompts_t. If something bad happens to the
SSH connection (e.g. unexpected server-side closure), then all the SSH
layers will be destroyed, including freeing that prompts_t. So the
Ldisc will have a stale reference to it, which it might potentially
use.
To fix that, I've arranged a back-pointer so that prompts_t itself can
find the Ldisc's reference to it, and NULL it out on free. So now,
whichever of a prompts_t and an Ldisc is freed first, the link between
them should be cleanly broken.
(I'm not 100% sure this is absolutely necessary, in the sense of
whether a sequence of events can _actually_ happen that causes a stale
pointer dereference. But I don't want to take the chance!)
2021-09-16 08:41:03 +00:00
|
|
|
if (ldisc->prompts && ldisc->prompts->ldisc_ptr_to_us == &ldisc->prompts)
|
|
|
|
|
ldisc->prompts->ldisc_ptr_to_us = NULL;
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
delete_callbacks_for_context(ldisc);
|
2003-01-15 23:30:21 +00:00
|
|
|
sfree(ldisc);
|
|
|
|
|
}
|
|
|
|
|
|
2018-09-11 14:02:59 +00:00
|
|
|
void ldisc_echoedit_update(Ldisc *ldisc)
|
Move echo/edit state change functionality out of ldisc_send.
I'm not actually sure why we've always had back ends notify ldisc of
changes to echo/edit settings by giving ldisc_send(ldisc,NULL,0,0) a
special meaning, instead of by having a separate dedicated notify
function with its own prototype and parameter set. Coverity's recent
observation that the two kinds of call don't even have the same
requirements on the ldisc (particularly, whether ldisc->term can be
NULL) makes me realise that it's really high time I separated the two
conceptually different operations into actually different functions.
While I'm here, I've renamed the confusing ldisc_update() function
which that special operation ends up feeding to, because it's not
actually a function applying to an ldisc - it applies to a front end.
So ldisc_send(ldisc,NULL,0,0) is now ldisc_echoedit_update(ldisc), and
that in turn figures out the current echo/edit settings before passing
them on to frontend_echoedit_update(). I think that should be clearer.
2014-11-22 16:12:47 +00:00
|
|
|
{
|
New abstraction 'Seat', to pass to backends.
This is a new vtable-based abstraction which is passed to a backend in
place of Frontend, and it implements only the subset of the Frontend
functions needed by a backend. (Many other Frontend functions still
exist, notably the wide range of things called by terminal.c providing
platform-independent operations on the GUI terminal window.)
The purpose of making it a vtable is that this opens up the
possibility of creating a backend as an internal implementation detail
of some other activity, by providing just that one backend with a
custom Seat that implements the methods differently.
For example, this refactoring should make it feasible to directly
implement an SSH proxy type, aka the 'jump host' feature supported by
OpenSSH, aka 'open a secondary SSH session in MAINCHAN_DIRECT_TCP
mode, and then expose the main channel of that as the Socket for the
primary connection'. (Which of course you can already do by spawning
'plink -nc' as a separate proxy process, but this would permit it in
the _same_ process without anything getting confused.)
I've centralised a full set of stub methods in misc.c for the new
abstraction, which allows me to get rid of several annoying stubs in
the previous code. Also, while I'm here, I've moved a lot of
duplicated modalfatalbox() type functions from application main
program files into wincons.c / uxcons.c, which I think saves
duplication overall. (A minor visible effect is that the prefixes on
those console-based fatal error messages will now be more consistent
between applications.)
2018-10-11 18:58:42 +00:00
|
|
|
seat_echoedit_update(ldisc->seat, ECHOING, EDITING);
|
Move echo/edit state change functionality out of ldisc_send.
I'm not actually sure why we've always had back ends notify ldisc of
changes to echo/edit settings by giving ldisc_send(ldisc,NULL,0,0) a
special meaning, instead of by having a separate dedicated notify
function with its own prototype and parameter set. Coverity's recent
observation that the two kinds of call don't even have the same
requirements on the ldisc (particularly, whether ldisc->term can be
NULL) makes me realise that it's really high time I separated the two
conceptually different operations into actually different functions.
While I'm here, I've renamed the confusing ldisc_update() function
which that special operation ends up feeding to, because it's not
actually a function applying to an ldisc - it applies to a front end.
So ldisc_send(ldisc,NULL,0,0) is now ldisc_echoedit_update(ldisc), and
that in turn figures out the current echo/edit settings before passing
them on to frontend_echoedit_update(). I think that should be clearer.
2014-11-22 16:12:47 +00:00
|
|
|
}
|
|
|
|
|
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
void ldisc_enable_prompt_callback(Ldisc *ldisc, prompts_t *prompts)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* Called by the terminal to indicate that there's a prompts_t
|
|
|
|
|
* currently in flight, or to indicate that one has just finished
|
|
|
|
|
* (by passing NULL). When ldisc->prompts is not null, we notify
|
|
|
|
|
* the terminal whenever new data arrives in our input queue, so
|
|
|
|
|
* that it can continue the interactive prompting process.
|
|
|
|
|
*/
|
|
|
|
|
ldisc->prompts = prompts;
|
free_prompts: deal with a reference from an Ldisc.
In a GUI app, when interactive userpass input begins, the Ldisc
acquires a reference to a prompts_t. If something bad happens to the
SSH connection (e.g. unexpected server-side closure), then all the SSH
layers will be destroyed, including freeing that prompts_t. So the
Ldisc will have a stale reference to it, which it might potentially
use.
To fix that, I've arranged a back-pointer so that prompts_t itself can
find the Ldisc's reference to it, and NULL it out on free. So now,
whichever of a prompts_t and an Ldisc is freed first, the link between
them should be cleanly broken.
(I'm not 100% sure this is absolutely necessary, in the sense of
whether a sequence of events can _actually_ happen that causes a stale
pointer dereference. But I don't want to take the chance!)
2021-09-16 08:41:03 +00:00
|
|
|
if (prompts)
|
|
|
|
|
ldisc->prompts->ldisc_ptr_to_us = &ldisc->prompts;
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void ldisc_input_queue_callback(void *ctx)
|
|
|
|
|
{
|
|
|
|
|
/*
|
|
|
|
|
* Toplevel callback that is triggered whenever the input queue
|
|
|
|
|
* lengthens. If we're currently processing an interactive prompt,
|
|
|
|
|
* we call back the Terminal to tell it to do some more stuff with
|
|
|
|
|
* that prompt based on the new input.
|
|
|
|
|
*/
|
|
|
|
|
Ldisc *ldisc = (Ldisc *)ctx;
|
|
|
|
|
if (ldisc->term && ldisc->prompts) {
|
|
|
|
|
/*
|
|
|
|
|
* The integer return value from this call is discarded,
|
|
|
|
|
* because we have no channel to pass it on to the backend
|
|
|
|
|
* that originally wanted it. But that's OK, because if the
|
|
|
|
|
* return value is >= 0 (that is, the prompts are either
|
|
|
|
|
* completely filled in, or aborted by the user), then the
|
|
|
|
|
* terminal will notify the callback in the prompts_t, and
|
|
|
|
|
* when that calls term_get_userpass_input again, it will
|
|
|
|
|
* return the same answer again.
|
|
|
|
|
*/
|
|
|
|
|
term_get_userpass_input(ldisc->term, ldisc->prompts);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void ldisc_to_backend_raw(
|
|
|
|
|
Ldisc *ldisc, const void *vbuf, size_t len)
|
|
|
|
|
{
|
|
|
|
|
if (backend_sendok(ldisc->backend)) {
|
|
|
|
|
backend_send(ldisc->backend, vbuf, len);
|
|
|
|
|
} else {
|
|
|
|
|
const char *buf = (const char *)vbuf;
|
|
|
|
|
while (len > 0) {
|
|
|
|
|
/*
|
|
|
|
|
* Encode raw data in input_queue, by storing large chunks
|
|
|
|
|
* as long as they don't include 0xFF, and pausing every
|
|
|
|
|
* time they do to escape it.
|
|
|
|
|
*/
|
|
|
|
|
const char *ff = memchr(buf, '\xFF', len);
|
|
|
|
|
size_t this_len = ff ? ff - buf : len;
|
|
|
|
|
if (this_len > 0) {
|
|
|
|
|
bufchain_add(&ldisc->input_queue, buf, len);
|
|
|
|
|
} else {
|
|
|
|
|
bufchain_add(&ldisc->input_queue, "\xFF\xFF", 2);
|
|
|
|
|
this_len = 1;
|
|
|
|
|
}
|
|
|
|
|
buf += this_len;
|
|
|
|
|
len -= this_len;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void ldisc_to_backend_special(
|
|
|
|
|
Ldisc *ldisc, SessionSpecialCode code, int arg)
|
|
|
|
|
{
|
|
|
|
|
if (backend_sendok(ldisc->backend)) {
|
|
|
|
|
backend_special(ldisc->backend, code, arg);
|
|
|
|
|
} else {
|
|
|
|
|
/*
|
|
|
|
|
* Encode a session special in input_queue.
|
|
|
|
|
*/
|
|
|
|
|
unsigned char data[9];
|
|
|
|
|
data[0] = 0xFF;
|
|
|
|
|
PUT_32BIT_MSB_FIRST(data+1, code);
|
|
|
|
|
PUT_32BIT_MSB_FIRST(data+5, arg);
|
|
|
|
|
assert(data[1] != 0xFF &&
|
|
|
|
|
"SessionSpecialCode encoding collides with FF FF escape");
|
|
|
|
|
bufchain_add(&ldisc->input_queue, data, 9);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
bool ldisc_has_input_buffered(Ldisc *ldisc)
|
|
|
|
|
{
|
|
|
|
|
return bufchain_size(&ldisc->input_queue) > 0;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
LdiscInputToken ldisc_get_input_token(Ldisc *ldisc)
|
|
|
|
|
{
|
|
|
|
|
assert(bufchain_size(&ldisc->input_queue) > 0 &&
|
|
|
|
|
"You're not supposed to call this unless there is buffered input!");
|
|
|
|
|
|
|
|
|
|
LdiscInputToken tok;
|
|
|
|
|
|
|
|
|
|
char c;
|
|
|
|
|
bufchain_fetch_consume(&ldisc->input_queue, &c, 1);
|
|
|
|
|
if (c != '\xFF') {
|
|
|
|
|
/* A literal non-FF byte */
|
|
|
|
|
tok.is_special = false;
|
|
|
|
|
tok.chr = c;
|
|
|
|
|
return tok;
|
|
|
|
|
} else {
|
|
|
|
|
char data[8];
|
|
|
|
|
|
|
|
|
|
/* See if the byte after the FF is also FF, indicating a literal FF */
|
|
|
|
|
bufchain_fetch_consume(&ldisc->input_queue, data, 1);
|
|
|
|
|
if (data[0] == '\xFF') {
|
|
|
|
|
tok.is_special = false;
|
|
|
|
|
tok.chr = '\xFF';
|
|
|
|
|
return tok;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* If not, get the rest of an 8-byte chunk and decode a special */
|
|
|
|
|
bufchain_fetch_consume(&ldisc->input_queue, data+1, 7);
|
|
|
|
|
tok.is_special = true;
|
|
|
|
|
tok.code = GET_32BIT_MSB_FIRST(data);
|
|
|
|
|
tok.arg = toint(GET_32BIT_MSB_FIRST(data+4));
|
|
|
|
|
return tok;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void ldisc_check_sendok_callback(void *ctx)
|
|
|
|
|
{
|
|
|
|
|
Ldisc *ldisc = (Ldisc *)ctx;
|
|
|
|
|
|
|
|
|
|
if (!(ldisc->backend && backend_sendok(ldisc->backend)))
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Flush the ldisc input queue into the backend, which is now
|
|
|
|
|
* willing to receive the data.
|
|
|
|
|
*/
|
|
|
|
|
while (bufchain_size(&ldisc->input_queue) > 0) {
|
|
|
|
|
/*
|
|
|
|
|
* Process either a chunk of non-special data, or an FF
|
|
|
|
|
* escape, depending on whether the first thing we see is an
|
|
|
|
|
* FF byte.
|
|
|
|
|
*/
|
|
|
|
|
ptrlen data = bufchain_prefix(&ldisc->input_queue);
|
|
|
|
|
const char *ff = memchr(data.ptr, '\xFF', data.len);
|
|
|
|
|
if (ff != data.ptr) {
|
|
|
|
|
/* Send a maximal block of data not containing any
|
|
|
|
|
* difficult bytes. */
|
|
|
|
|
if (ff)
|
|
|
|
|
data.len = ff - (const char *)data.ptr;
|
|
|
|
|
backend_send(ldisc->backend, data.ptr, data.len);
|
|
|
|
|
bufchain_consume(&ldisc->input_queue, data.len);
|
|
|
|
|
} else {
|
|
|
|
|
/* Decode either a special or an escaped FF byte. The
|
|
|
|
|
* easiest way to do this is to reuse the decoding code
|
|
|
|
|
* already in ldisc_get_input_token. */
|
|
|
|
|
LdiscInputToken tok = ldisc_get_input_token(ldisc);
|
|
|
|
|
if (tok.is_special)
|
|
|
|
|
backend_special(ldisc->backend, tok.code, tok.arg);
|
|
|
|
|
else
|
|
|
|
|
backend_send(ldisc->backend, &tok.chr, 1);
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
2021-09-14 09:13:28 +00:00
|
|
|
void ldisc_check_sendok(Ldisc *ldisc)
|
|
|
|
|
{
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
queue_toplevel_callback(ldisc_check_sendok_callback, ldisc);
|
2021-09-14 09:13:28 +00:00
|
|
|
}
|
|
|
|
|
|
Convert a lot of 'int' variables to 'bool'.
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'!
2018-11-02 19:23:19 +00:00
|
|
|
void ldisc_send(Ldisc *ldisc, const void *vbuf, int len, bool interactive)
|
2002-10-26 10:16:19 +00:00
|
|
|
{
|
2018-05-26 07:31:34 +00:00
|
|
|
const char *buf = (const char *)vbuf;
|
2001-05-09 15:12:26 +00:00
|
|
|
int keyflag = 0;
|
2014-11-22 10:37:14 +00:00
|
|
|
|
|
|
|
|
assert(ldisc->term);
|
|
|
|
|
|
|
|
|
|
if (interactive) {
|
2013-08-17 16:06:40 +00:00
|
|
|
/*
|
|
|
|
|
* Interrupt a paste from the clipboard, if one was in
|
|
|
|
|
* progress when the user pressed a key. This is easier than
|
|
|
|
|
* buffering the current piece of data and saving it until the
|
|
|
|
|
* terminal has finished pasting, and has the potential side
|
|
|
|
|
* benefit of permitting a user to cancel an accidental huge
|
|
|
|
|
* paste.
|
|
|
|
|
*/
|
|
|
|
|
term_nopaste(ldisc->term);
|
|
|
|
|
}
|
|
|
|
|
|
2001-05-09 15:12:26 +00:00
|
|
|
/*
|
|
|
|
|
* Less than zero means null terminated special string.
|
|
|
|
|
*/
|
|
|
|
|
if (len < 0) {
|
2019-09-08 19:29:00 +00:00
|
|
|
len = strlen(buf);
|
|
|
|
|
keyflag = KCTRL('@');
|
2001-05-09 15:12:26 +00:00
|
|
|
}
|
2001-01-24 14:08:20 +00:00
|
|
|
/*
|
|
|
|
|
* Either perform local editing, or just send characters.
|
|
|
|
|
*/
|
|
|
|
|
if (EDITING) {
|
2019-09-08 19:29:00 +00:00
|
|
|
while (len--) {
|
|
|
|
|
int c;
|
|
|
|
|
c = (unsigned char)(*buf++) + keyflag;
|
|
|
|
|
if (!interactive && c == '\r')
|
|
|
|
|
c += KCTRL('@');
|
|
|
|
|
switch (ldisc->quotenext ? ' ' : c) {
|
|
|
|
|
/*
|
|
|
|
|
* ^h/^?: delete, and output BSBs, to return to
|
|
|
|
|
* last character boundary (in UTF-8 mode this may
|
|
|
|
|
* be more than one byte)
|
|
|
|
|
* ^w: delete, and output BSBs, to return to last
|
|
|
|
|
* space/nonspace boundary
|
|
|
|
|
* ^u: delete, and output BSBs, to return to BOL
|
|
|
|
|
* ^c: Do a ^u then send a telnet IP
|
|
|
|
|
* ^z: Do a ^u then send a telnet SUSP
|
|
|
|
|
* ^\: Do a ^u then send a telnet ABORT
|
|
|
|
|
* ^r: echo "^R\n" and redraw line
|
|
|
|
|
* ^v: quote next char
|
|
|
|
|
* ^d: if at BOL, end of file and close connection,
|
|
|
|
|
* else send line and reset to BOL
|
|
|
|
|
* ^m: send line-plus-\r\n and reset to BOL
|
|
|
|
|
*/
|
|
|
|
|
case KCTRL('H'):
|
|
|
|
|
case KCTRL('?'): /* backspace/delete */
|
|
|
|
|
if (ldisc->buflen > 0) {
|
|
|
|
|
do {
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
bsb(ldisc, plen(ldisc, ldisc->buf[ldisc->buflen - 1]));
|
|
|
|
|
ldisc->buflen--;
|
|
|
|
|
} while (!char_start(ldisc, ldisc->buf[ldisc->buflen]));
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case CTRL('W'): /* delete word */
|
|
|
|
|
while (ldisc->buflen > 0) {
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
bsb(ldisc, plen(ldisc, ldisc->buf[ldisc->buflen - 1]));
|
|
|
|
|
ldisc->buflen--;
|
|
|
|
|
if (ldisc->buflen > 0 &&
|
|
|
|
|
isspace((unsigned char)ldisc->buf[ldisc->buflen-1]) &&
|
|
|
|
|
!isspace((unsigned char)ldisc->buf[ldisc->buflen]))
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case CTRL('U'): /* delete line */
|
|
|
|
|
case CTRL('C'): /* Send IP */
|
|
|
|
|
case CTRL('\\'): /* Quit */
|
|
|
|
|
case CTRL('Z'): /* Suspend */
|
|
|
|
|
while (ldisc->buflen > 0) {
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
bsb(ldisc, plen(ldisc, ldisc->buf[ldisc->buflen - 1]));
|
|
|
|
|
ldisc->buflen--;
|
|
|
|
|
}
|
2022-12-21 15:05:04 +00:00
|
|
|
if (c == CTRL('U'))
|
|
|
|
|
break; /* ^U *just* erases a line */
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_EL, 0);
|
2001-10-24 11:50:07 +00:00
|
|
|
/*
|
|
|
|
|
* We don't send IP, SUSP or ABORT if the user has
|
|
|
|
|
* configured telnet specials off! This breaks
|
|
|
|
|
* talkers otherwise.
|
|
|
|
|
*/
|
Post-release destabilisation! Completely remove the struct type
'Config' in putty.h, which stores all PuTTY's settings and includes an
arbitrary length limit on every single one of those settings which is
stored in string form. In place of it is 'Conf', an opaque data type
everywhere outside the new file conf.c, which stores a list of (key,
value) pairs in which every key contains an integer identifying a
configuration setting, and for some of those integers the key also
contains extra parts (so that, for instance, CONF_environmt is a
string-to-string mapping). Everywhere that a Config was previously
used, a Conf is now; everywhere there was a Config structure copy,
conf_copy() is called; every lookup, adjustment, load and save
operation on a Config has been rewritten; and there's a mechanism for
serialising a Conf into a binary blob and back for use with Duplicate
Session.
User-visible effects of this change _should_ be minimal, though I
don't doubt I've introduced one or two bugs here and there which will
eventually be found. The _intended_ visible effects of this change are
that all arbitrary limits on configuration strings and lists (e.g.
limit on number of port forwardings) should now disappear; that list
boxes in the configuration will now be displayed in a sorted order
rather than the arbitrary order in which they were added to the list
(since the underlying data structure is now a sorted tree234 rather
than an ad-hoc comma-separated string); and one more specific change,
which is that local and dynamic port forwardings on the same port
number are now mutually exclusive in the configuration (putting 'D' in
the key rather than the value was a mistake in the first place).
One other reorganisation as a result of this is that I've moved all
the dialog.c standard handlers (dlg_stdeditbox_handler and friends)
out into config.c, because I can't really justify calling them generic
any more. When they took a pointer to an arbitrary structure type and
the offset of a field within that structure, they were independent of
whether that structure was a Config or something completely different,
but now they really do expect to talk to a Conf, which can _only_ be
used for PuTTY configuration, so I've renamed them all things like
conf_editbox_handler and moved them out of the nominally independent
dialog-box management module into the PuTTY-specific config.c.
[originally from svn r9214]
2011-07-14 18:52:21 +00:00
|
|
|
if (!ldisc->telnet_keyboard)
|
2001-10-24 11:50:07 +00:00
|
|
|
goto default_case;
|
2019-09-08 19:29:00 +00:00
|
|
|
if (c == CTRL('C'))
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_IP, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
if (c == CTRL('Z'))
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_SUSP, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
if (c == CTRL('\\'))
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_ABORT, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
case CTRL('R'): /* redraw line */
|
|
|
|
|
if (ECHOING) {
|
|
|
|
|
int i;
|
|
|
|
|
c_write(ldisc, "^R\r\n", 4);
|
|
|
|
|
for (i = 0; i < ldisc->buflen; i++)
|
|
|
|
|
pwrite(ldisc, ldisc->buf[i]);
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
case CTRL('V'): /* quote next char */
|
|
|
|
|
ldisc->quotenext = true;
|
|
|
|
|
break;
|
|
|
|
|
case CTRL('D'): /* logout or send */
|
|
|
|
|
if (ldisc->buflen == 0) {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_EOF, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
} else {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, ldisc->buf, ldisc->buflen);
|
2019-09-08 19:29:00 +00:00
|
|
|
ldisc->buflen = 0;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
/*
|
|
|
|
|
* This particularly hideous bit of code from RDB
|
|
|
|
|
* allows ordinary ^M^J to do the same thing as
|
|
|
|
|
* magic-^M when in Raw protocol. The line `case
|
|
|
|
|
* KCTRL('M'):' is _inside_ the if block. Thus:
|
|
|
|
|
*
|
|
|
|
|
* - receiving regular ^M goes straight to the
|
|
|
|
|
* default clause and inserts as a literal ^M.
|
|
|
|
|
* - receiving regular ^J _not_ directly after a
|
|
|
|
|
* literal ^M (or not in Raw protocol) fails the
|
|
|
|
|
* if condition, leaps to the bottom of the if,
|
|
|
|
|
* and falls through into the default clause
|
|
|
|
|
* again.
|
|
|
|
|
* - receiving regular ^J just after a literal ^M
|
|
|
|
|
* in Raw protocol passes the if condition,
|
|
|
|
|
* deletes the literal ^M, and falls through
|
|
|
|
|
* into the magic-^M code
|
|
|
|
|
* - receiving a magic-^M empties the line buffer,
|
|
|
|
|
* signals end-of-line in one of the various
|
|
|
|
|
* entertaining ways, and _doesn't_ fall out of
|
|
|
|
|
* the bottom of the if and through to the
|
|
|
|
|
* default clause because of the break.
|
|
|
|
|
*/
|
|
|
|
|
case CTRL('J'):
|
|
|
|
|
if (ldisc->protocol == PROT_RAW &&
|
|
|
|
|
ldisc->buflen > 0 && ldisc->buf[ldisc->buflen - 1] == '\r') {
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
bsb(ldisc, plen(ldisc, ldisc->buf[ldisc->buflen - 1]));
|
|
|
|
|
ldisc->buflen--;
|
|
|
|
|
/* FALLTHROUGH */
|
|
|
|
|
case KCTRL('M'): /* send with newline */
|
|
|
|
|
if (ldisc->buflen > 0)
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, ldisc->buf, ldisc->buflen);
|
2019-09-08 19:29:00 +00:00
|
|
|
if (ldisc->protocol == PROT_RAW)
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, "\r\n", 2);
|
2019-09-08 19:29:00 +00:00
|
|
|
else if (ldisc->protocol == PROT_TELNET && ldisc->telnet_newline)
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_EOL, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
else
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, "\r", 1);
|
2019-09-08 19:29:00 +00:00
|
|
|
if (ECHOING)
|
|
|
|
|
c_write(ldisc, "\r\n", 2);
|
|
|
|
|
ldisc->buflen = 0;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
/* FALLTHROUGH */
|
|
|
|
|
default: /* get to this label from ^V handler */
|
2022-08-03 19:48:46 +00:00
|
|
|
default_case:
|
New array-growing macros: sgrowarray and sgrowarrayn.
The idea of these is that they centralise the common idiom along the
lines of
if (logical_array_len >= physical_array_size) {
physical_array_size = logical_array_len * 5 / 4 + 256;
array = sresize(array, physical_array_size, ElementType);
}
which happens at a zillion call sites throughout this code base, with
different random choices of the geometric factor and additive
constant, sometimes forgetting them completely, and generally doing a
lot of repeated work.
The new macro sgrowarray(array,size,n) has the semantics: here are the
array pointer and its physical size for you to modify, now please
ensure that the nth element exists, so I can write into it. And
sgrowarrayn(array,size,n,m) is the same except that it ensures that
the array has size at least n+m (so sgrowarray is just the special
case where m=1).
Now that this is a single centralised implementation that will be used
everywhere, I've also gone to more effort in the implementation, with
careful overflow checks that would have been painful to put at all the
previous call sites.
This commit also switches over every use of sresize(), apart from a
few where I really didn't think it would gain anything. A consequence
of that is that a lot of array-size variables have to have their types
changed to size_t, because the macros require that (they address-take
the size to pass to the underlying function).
2019-02-28 20:07:30 +00:00
|
|
|
sgrowarray(ldisc->buf, ldisc->bufsiz, ldisc->buflen);
|
2019-09-08 19:29:00 +00:00
|
|
|
ldisc->buf[ldisc->buflen++] = c;
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
pwrite(ldisc, (unsigned char) c);
|
|
|
|
|
ldisc->quotenext = false;
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
}
|
2001-01-24 14:08:20 +00:00
|
|
|
} else {
|
2019-09-08 19:29:00 +00:00
|
|
|
if (ldisc->buflen != 0) {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, ldisc->buf, ldisc->buflen);
|
2019-09-08 19:29:00 +00:00
|
|
|
while (ldisc->buflen > 0) {
|
|
|
|
|
bsb(ldisc, plen(ldisc, ldisc->buf[ldisc->buflen - 1]));
|
|
|
|
|
ldisc->buflen--;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
if (len > 0) {
|
|
|
|
|
if (ECHOING)
|
|
|
|
|
c_write(ldisc, buf, len);
|
|
|
|
|
if (keyflag && ldisc->protocol == PROT_TELNET && len == 1) {
|
|
|
|
|
switch (buf[0]) {
|
|
|
|
|
case CTRL('M'):
|
|
|
|
|
if (ldisc->protocol == PROT_TELNET && ldisc->telnet_newline)
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_EOL, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
else
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, "\r", 1);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
case CTRL('?'):
|
|
|
|
|
case CTRL('H'):
|
|
|
|
|
if (ldisc->telnet_keyboard) {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_EC, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case CTRL('C'):
|
|
|
|
|
if (ldisc->telnet_keyboard) {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_IP, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
case CTRL('Z'):
|
|
|
|
|
if (ldisc->telnet_keyboard) {
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_special(ldisc, SS_SUSP, 0);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
}
|
2001-05-09 15:12:26 +00:00
|
|
|
|
2019-09-08 19:29:00 +00:00
|
|
|
default:
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, buf, len);
|
2019-09-08 19:29:00 +00:00
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
} else
|
Complete rework of terminal userpass input system.
The system for handling seat_get_userpass_input has always been
structured differently between GUI PuTTY and CLI tools like Plink.
In the CLI tools, password input is read directly from the OS
terminal/console device by console_get_userpass_input; this means that
you need to ensure the same terminal input data _hasn't_ already been
consumed by the main event loop and sent on to the backend. This is
achieved by the backend_sendok() method, which tells the event loop
when the backend has finished issuing password prompts, and hence,
when it's safe to start passing standard input to backend_send().
But in the GUI tools, input generated by the terminal window has
always been sent straight to backend_send(), regardless of whether
backend_sendok() says it wants it. So the terminal-based
implementation of username and password prompts has to work by
consuming input data that had _already_ been passed to the backend -
hence, any backend that needs to do that must keep its input on a
bufchain, and pass that bufchain to seat_get_userpass_input.
It's awkward that these two totally different systems coexist in the
first place. And now that SSH proxying needs to present interactive
prompts of its own, it's clear which one should win: the CLI style is
the Right Thing. So this change reworks the GUI side of the mechanism
to be more similar: terminal data now goes into a queue in the Ldisc,
and is not sent on to the backend until the backend says it's ready
for it via backend_sendok(). So terminal-based userpass prompts can
now consume data directly from that queue during the connection setup
stage.
As a result, the 'bufchain *' parameter has vanished from all the
userpass_input functions (both the official implementations of the
Seat trait method, and term_get_userpass_input() to which some of
those implementations delegate). The only function that actually used
that bufchain, namely term_get_userpass_input(), now instead reads
from the ldisc's input queue via a couple of new Ldisc functions.
(Not _trivial_ functions, since input buffered by Ldisc can be a
mixture of raw bytes and session specials like SS_EOL! The input queue
inside Ldisc is a bufchain containing a fiddly binary encoding that
can represent an arbitrary interleaving of those things.)
This greatly simplifies the calls to seat_get_userpass_input in
backends, which now don't have to mess about with passing their own
user_input bufchain around, or toggling their want_user_input flag
back and forth to request data to put on to that bufchain.
But the flip side is that now there has to be some _other_ method for
notifying the terminal when there's more input to be consumed during
an interactive prompt, and for notifying the backend when prompt input
has finished so that it can proceed to the next stage of the protocol.
This is done by a pair of extra callbacks: when more data is put on to
Ldisc's input queue, it triggers a call to term_get_userpass_input,
and when term_get_userpass_input finishes, it calls a callback
function provided in the prompts_t.
Therefore, any use of a prompts_t which *might* be asynchronous must
fill in the latter callback when setting up the prompts_t. In SSH, the
callback is centralised into a common PPL helper function, which
reinvokes the same PPL's process_queue coroutine; in rlogin we have to
set it up ourselves.
I'm sorry for this large and sprawling patch: I tried fairly hard to
break it up into individually comprehensible sub-patches, but I just
couldn't tease out any part of it that would stand sensibly alone.
2021-09-14 10:57:21 +00:00
|
|
|
ldisc_to_backend_raw(ldisc, buf, len);
|
2019-09-08 19:29:00 +00:00
|
|
|
}
|
2000-03-11 14:03:04 +00:00
|
|
|
}
|
1999-11-09 12:05:34 +00:00
|
|
|
}
|