nhyatt/putty-source
1
0
Fork 0
mirror of https://git.tartarus.org/simon/putty.git synced 2025-01-09 17:38:00 +00:00
Code Issues Projects Releases Wiki Activity
9fe719f47d
putty-source/ssh2bpp.c

892 lines
32 KiB
C
Raw Normal View History

Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/*
* Binary packet protocol for SSH-2.
*/
#include <assert.h>
#include "putty.h"
#include "ssh.h"
#include "sshbpp.h"
#include "sshcr.h"
struct ssh2_bpp_direction {
unsigned long sequence;
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher *cipher;
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
ssh2_mac *mac;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
int etm_mode;
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
const struct ssh_compression_alg *pending_compression;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
};
struct ssh2_bpp_state {
int crState;
long len, pad, payload, packetlen, maclen, length, maxlen;
unsigned char *buf;
size_t bufsize;
unsigned char *data;
unsigned cipherblk;
PktIn *pktin;
New system for tracking data-limit-based rekeys. I've removed the encrypted_len fields from PktIn and PktOut, which were used to communicate from the BPP to ssh.c how much each packet contributed to the amount of data encrypted with a given set of cipher keys. It seems more sensible to have the BPP itself keep that counter - especially since only one of the three BPPs even needs to count it at all. So now there's a new DataTransferStats structure which the BPP updates, and ssh.c only needs to check it for overflow and reset the limits.
2018-09-19 20:28:21 +00:00
struct DataTransferStats *stats;
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
int cbc_ignore_workaround;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
struct ssh2_bpp_direction in, out;
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
/* comp and decomp logically belong in the per-direction
* substructure, except that they have different types */
ssh_decompressor *in_decomp;
ssh_compressor *out_comp;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
int pending_newkeys;
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
int pending_compression, seen_userauth_success;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
BinaryPacketProtocol bpp;
};
static void ssh2_bpp_free(BinaryPacketProtocol *bpp);
static void ssh2_bpp_handle_input(BinaryPacketProtocol *bpp);
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
static void ssh2_bpp_handle_output(BinaryPacketProtocol *bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
static PktOut *ssh2_bpp_new_pktout(int type);
Add missing 'static' to BPP vtable definitions. Vtable objects only need to be globally visible throughout the code if they're used directly in some interchangeable way, e.g. by passing them to a constructor like cipher_new that's the same for all implementations of the vtable, or by directly looking up public data fields in the vtable itself. But the BPPs are never used like that: each BPP has its own constructor function with a different type signature, so the BPP types are not interchangeable in any way _before_ an instance of one has been constructed. Hence, their vtable objects don't need external linkage.
2018-09-23 08:30:37 +00:00
static const struct BinaryPacketProtocolVtable ssh2_bpp_vtable = {
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
ssh2_bpp_free,
ssh2_bpp_handle_input,
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
ssh2_bpp_handle_output,
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
ssh2_bpp_new_pktout,
Move SSH_MSG_DISCONNECT construction into the BPP. This is a convenient place for it because it abstracts away the difference in disconnect packet formats between SSH-1 and -2, so when I start restructuring, I'll be able to call it even from places that don't know which version of SSH they're running.
2018-09-24 17:14:33 +00:00
ssh2_bpp_queue_disconnect, /* in sshcommon.c */
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
};
Give BPPs a Frontend, so they can do their own logging. The sshverstring quasi-frontend is passed a Frontend pointer at setup time, so that it can generate Event Log entries containing the local and remote version strings and the results of remote bug detection. I'm promoting that field of sshverstring to a field of the public BPP structure, so now all BPPs have the right to talk directly to the frontend if they want to. This means I can move all the log messages of the form 'Initialised so-and-so cipher/MAC/compression' down into the BPPs themselves, where they can live exactly alongside the actual initialisation of those primitives. It also means BPPs will be able to log interesting things they detect at any point in the packet stream, which is about to come in useful for another purpose.
2018-10-07 07:16:44 +00:00
BinaryPacketProtocol *ssh2_bpp_new(
Refactor the LogContext type. LogContext is now the owner of the logevent() function that back ends and so forth are constantly calling. Previously, logevent was owned by the Frontend, which would store the message into its list for the GUI Event Log dialog (or print it to standard error, or whatever) and then pass it _back_ to LogContext to write to the currently open log file. Now it's the other way round: LogContext gets the message from the back end first, writes it to its log file if it feels so inclined, and communicates it back to the front end. This means that lots of parts of the back end system no longer need to have a pointer to a full-on Frontend; the only thing they needed it for was logging, so now they just have a LogContext (which many of them had to have anyway, e.g. for logging SSH packets or session traffic). LogContext itself also doesn't get a full Frontend pointer any more: it now talks back to the front end via a little vtable of its own called LogPolicy, which contains the method that passes Event Log entries through, the old askappend() function that decides whether to truncate a pre-existing log file, and an emergency function for printing an especially prominent message if the log file can't be created. One minor nice effect of this is that console and GUI apps can implement that last function subtly differently, so that Unix console apps can write it with a plain \n instead of the \r\n (harmless but inelegant) that the old centralised implementation generated. One other consequence of this is that the LogContext has to be provided to backend_init() so that it's available to backends from the instant of creation, rather than being provided via a separate API call a couple of function calls later, because backends have typically started doing things that need logging (like making network connections) before the call to backend_provide_logctx. Fortunately, there's no case in the whole code base where we don't already have logctx by the time we make a backend (so I don't actually remember why I ever delayed providing one). So that shortens the backend API by one function, which is always nice. While I'm tidying up, I've also moved the printf-style logeventf() and the handy logevent_and_free() into logging.c, instead of having copies of them scattered around other places. This has also let me remove some stub functions from a couple of outlying applications like Pageant. Finally, I've removed the pointless "_tag" at the end of LogContext's official struct name.
2018-10-10 18:26:18 +00:00
LogContext *logctx, struct DataTransferStats *stats)
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
{
struct ssh2_bpp_state *s = snew(struct ssh2_bpp_state);
memset(s, 0, sizeof(*s));
s->bpp.vt = &ssh2_bpp_vtable;
Refactor the LogContext type. LogContext is now the owner of the logevent() function that back ends and so forth are constantly calling. Previously, logevent was owned by the Frontend, which would store the message into its list for the GUI Event Log dialog (or print it to standard error, or whatever) and then pass it _back_ to LogContext to write to the currently open log file. Now it's the other way round: LogContext gets the message from the back end first, writes it to its log file if it feels so inclined, and communicates it back to the front end. This means that lots of parts of the back end system no longer need to have a pointer to a full-on Frontend; the only thing they needed it for was logging, so now they just have a LogContext (which many of them had to have anyway, e.g. for logging SSH packets or session traffic). LogContext itself also doesn't get a full Frontend pointer any more: it now talks back to the front end via a little vtable of its own called LogPolicy, which contains the method that passes Event Log entries through, the old askappend() function that decides whether to truncate a pre-existing log file, and an emergency function for printing an especially prominent message if the log file can't be created. One minor nice effect of this is that console and GUI apps can implement that last function subtly differently, so that Unix console apps can write it with a plain \n instead of the \r\n (harmless but inelegant) that the old centralised implementation generated. One other consequence of this is that the LogContext has to be provided to backend_init() so that it's available to backends from the instant of creation, rather than being provided via a separate API call a couple of function calls later, because backends have typically started doing things that need logging (like making network connections) before the call to backend_provide_logctx. Fortunately, there's no case in the whole code base where we don't already have logctx by the time we make a backend (so I don't actually remember why I ever delayed providing one). So that shortens the backend API by one function, which is always nice. While I'm tidying up, I've also moved the printf-style logeventf() and the handy logevent_and_free() into logging.c, instead of having copies of them scattered around other places. This has also let me remove some stub functions from a couple of outlying applications like Pageant. Finally, I've removed the pointless "_tag" at the end of LogContext's official struct name.
2018-10-10 18:26:18 +00:00
s->bpp.logctx = logctx;
New system for tracking data-limit-based rekeys. I've removed the encrypted_len fields from PktIn and PktOut, which were used to communicate from the BPP to ssh.c how much each packet contributed to the amount of data encrypted with a given set of cipher keys. It seems more sensible to have the BPP itself keep that counter - especially since only one of the three BPPs even needs to count it at all. So now there's a new DataTransferStats structure which the BPP updates, and ssh.c only needs to check it for overflow and reset the limits.
2018-09-19 20:28:21 +00:00
s->stats = stats;
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
ssh_bpp_common_setup(&s->bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
return &s->bpp;
}
static void ssh2_bpp_free(BinaryPacketProtocol *bpp)
{
Rename FROMFIELD to 'container_of'. Ian Jackson points out that the Linux kernel has a macro of this name with the same purpose, and suggests that it's a good idea to use the same name as they do, so that at least some people reading one code base might recognise it from the other. I never really thought very hard about what order FROMFIELD's parameters should go in, and therefore I'm pleasantly surprised to find that my order agrees with the kernel's, so I don't have to permute every call site as part of making this change :-)
2018-10-05 22:49:08 +00:00
struct ssh2_bpp_state *s = container_of(bpp, struct ssh2_bpp_state, bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
sfree(s->buf);
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->out.cipher)
ssh2_cipher_free(s->out.cipher);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->out.mac)
ssh2_mac_free(s->out.mac);
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->out_comp)
ssh_compressor_free(s->out_comp);
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->in.cipher)
ssh2_cipher_free(s->in.cipher);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->in.mac)
ssh2_mac_free(s->in.mac);
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->in_decomp)
ssh_decompressor_free(s->in_decomp);
Replace PktIn reference count with a 'free queue'. This is a new idea I've had to make memory-management of PktIn even easier. The idea is that a PktIn is essentially _always_ an element of some linked-list queue: if it's not one of the queues by which packets move through ssh.c, then it's a special 'free queue' which holds packets that are unowned and due to be freed. pq_pop() on a PktInQueue automatically relinks the packet to the free queue, and also triggers an idempotent callback which will empty the queue and really free all the packets on it. Hence, you can pop a packet off a real queue, parse it, handle it, and then just assume it'll get tidied up at some point - the only constraint being that you have to finish with it before returning to the application's main loop. The exception is that it's OK to pq_push() the packet back on to some other PktInQueue, because a side effect of that will be to _remove_ it from the free queue again. (And if _all_ the incoming packets get that treatment, then when the free-queue handler eventually runs, it may find it has nothing to do - which is harmless.)
2018-09-23 15:35:29 +00:00
sfree(s->pktin);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
sfree(s);
}
void ssh2_bpp_new_outgoing_crypto(
BinaryPacketProtocol *bpp,
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
const struct ssh2_cipheralg *cipher, const void *ckey, const void *iv,
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
const struct ssh2_macalg *mac, int etm_mode, const void *mac_key,
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
const struct ssh_compression_alg *compression, int delayed_compression)
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
{
struct ssh2_bpp_state *s;
assert(bpp->vt == &ssh2_bpp_vtable);
Rename FROMFIELD to 'container_of'. Ian Jackson points out that the Linux kernel has a macro of this name with the same purpose, and suggests that it's a good idea to use the same name as they do, so that at least some people reading one code base might recognise it from the other. I never really thought very hard about what order FROMFIELD's parameters should go in, and therefore I'm pleasantly surprised to find that my order agrees with the kernel's, so I don't have to permute every call site as part of making this change :-)
2018-10-05 22:49:08 +00:00
s = container_of(bpp, struct ssh2_bpp_state, bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->out.cipher)
ssh2_cipher_free(s->out.cipher);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->out.mac)
ssh2_mac_free(s->out.mac);
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->out_comp)
ssh_compressor_free(s->out_comp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (cipher) {
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
s->out.cipher = ssh2_cipher_new(cipher);
ssh2_cipher_setkey(s->out.cipher, ckey);
ssh2_cipher_setiv(s->out.cipher, iv);
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
s->cbc_ignore_workaround = (
(ssh2_cipher_alg(s->out.cipher)->flags & SSH_CIPHER_IS_CBC) &&
!(s->bpp.remote_bugs & BUG_CHOKES_ON_SSH2_IGNORE));
Give BPPs a Frontend, so they can do their own logging. The sshverstring quasi-frontend is passed a Frontend pointer at setup time, so that it can generate Event Log entries containing the local and remote version strings and the results of remote bug detection. I'm promoting that field of sshverstring to a field of the public BPP structure, so now all BPPs have the right to talk directly to the frontend if they want to. This means I can move all the log messages of the form 'Initialised so-and-so cipher/MAC/compression' down into the BPPs themselves, where they can live exactly alongside the actual initialisation of those primitives. It also means BPPs will be able to log interesting things they detect at any point in the packet stream, which is about to come in useful for another purpose.
2018-10-07 07:16:44 +00:00
bpp_logevent(("Initialised %.200s client->server encryption",
ssh2_cipher_alg(s->out.cipher)->text_name));
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
} else {
s->out.cipher = NULL;
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
s->cbc_ignore_workaround = FALSE;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
s->out.etm_mode = etm_mode;
if (mac) {
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
s->out.mac = ssh2_mac_new(mac, s->out.cipher);
mac->setkey(s->out.mac, mac_key);
Give BPPs a Frontend, so they can do their own logging. The sshverstring quasi-frontend is passed a Frontend pointer at setup time, so that it can generate Event Log entries containing the local and remote version strings and the results of remote bug detection. I'm promoting that field of sshverstring to a field of the public BPP structure, so now all BPPs have the right to talk directly to the frontend if they want to. This means I can move all the log messages of the form 'Initialised so-and-so cipher/MAC/compression' down into the BPPs themselves, where they can live exactly alongside the actual initialisation of those primitives. It also means BPPs will be able to log interesting things they detect at any point in the packet stream, which is about to come in useful for another purpose.
2018-10-07 07:16:44 +00:00
bpp_logevent(("Initialised %.200s client->server"
" MAC algorithm%s%s",
ssh2_mac_alg(s->out.mac)->text_name,
etm_mode ? " (in ETM mode)" : "",
(s->out.cipher &&
ssh2_cipher_alg(s->out.cipher)->required_mac ?
" (required by cipher)" : "")));
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
} else {
s->out.mac = NULL;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
if (delayed_compression && !s->seen_userauth_success) {
s->out.pending_compression = compression;
s->out_comp = NULL;
bpp_logevent(("Will enable %s compression after user authentication",
s->out.pending_compression->text_name));
} else {
s->out.pending_compression = NULL;
/* 'compression' is always non-NULL, because no compression is
* indicated by ssh_comp_none. But this setup call may return a
* null out_comp. */
s->out_comp = ssh_compressor_new(compression);
if (s->out_comp)
bpp_logevent(("Initialised %s compression",
ssh_compressor_alg(s->out_comp)->text_name));
}
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
void ssh2_bpp_new_incoming_crypto(
BinaryPacketProtocol *bpp,
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
const struct ssh2_cipheralg *cipher, const void *ckey, const void *iv,
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
const struct ssh2_macalg *mac, int etm_mode, const void *mac_key,
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
const struct ssh_compression_alg *compression, int delayed_compression)
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
{
struct ssh2_bpp_state *s;
assert(bpp->vt == &ssh2_bpp_vtable);
Rename FROMFIELD to 'container_of'. Ian Jackson points out that the Linux kernel has a macro of this name with the same purpose, and suggests that it's a good idea to use the same name as they do, so that at least some people reading one code base might recognise it from the other. I never really thought very hard about what order FROMFIELD's parameters should go in, and therefore I'm pleasantly surprised to find that my order agrees with the kernel's, so I don't have to permute every call site as part of making this change :-)
2018-10-05 22:49:08 +00:00
s = container_of(bpp, struct ssh2_bpp_state, bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->in.cipher)
ssh2_cipher_free(s->in.cipher);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->in.mac)
ssh2_mac_free(s->in.mac);
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->in_decomp)
ssh_decompressor_free(s->in_decomp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (cipher) {
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
s->in.cipher = ssh2_cipher_new(cipher);
ssh2_cipher_setkey(s->in.cipher, ckey);
ssh2_cipher_setiv(s->in.cipher, iv);
Give BPPs a Frontend, so they can do their own logging. The sshverstring quasi-frontend is passed a Frontend pointer at setup time, so that it can generate Event Log entries containing the local and remote version strings and the results of remote bug detection. I'm promoting that field of sshverstring to a field of the public BPP structure, so now all BPPs have the right to talk directly to the frontend if they want to. This means I can move all the log messages of the form 'Initialised so-and-so cipher/MAC/compression' down into the BPPs themselves, where they can live exactly alongside the actual initialisation of those primitives. It also means BPPs will be able to log interesting things they detect at any point in the packet stream, which is about to come in useful for another purpose.
2018-10-07 07:16:44 +00:00
bpp_logevent(("Initialised %.200s server->client encryption",
ssh2_cipher_alg(s->in.cipher)->text_name));
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
} else {
s->in.cipher = NULL;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
s->in.etm_mode = etm_mode;
if (mac) {
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
s->in.mac = ssh2_mac_new(mac, s->in.cipher);
mac->setkey(s->in.mac, mac_key);
Give BPPs a Frontend, so they can do their own logging. The sshverstring quasi-frontend is passed a Frontend pointer at setup time, so that it can generate Event Log entries containing the local and remote version strings and the results of remote bug detection. I'm promoting that field of sshverstring to a field of the public BPP structure, so now all BPPs have the right to talk directly to the frontend if they want to. This means I can move all the log messages of the form 'Initialised so-and-so cipher/MAC/compression' down into the BPPs themselves, where they can live exactly alongside the actual initialisation of those primitives. It also means BPPs will be able to log interesting things they detect at any point in the packet stream, which is about to come in useful for another purpose.
2018-10-07 07:16:44 +00:00
bpp_logevent(("Initialised %.200s server->client MAC algorithm%s%s",
ssh2_mac_alg(s->in.mac)->text_name,
etm_mode ? " (in ETM mode)" : "",
(s->in.cipher &&
ssh2_cipher_alg(s->in.cipher)->required_mac ?
" (required by cipher)" : "")));
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
} else {
s->in.mac = NULL;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
if (delayed_compression && !s->seen_userauth_success) {
s->in.pending_compression = compression;
s->in_decomp = NULL;
bpp_logevent(("Will enable %s decompression after user authentication",
s->in.pending_compression->text_name));
} else {
s->in.pending_compression = NULL;
/* 'compression' is always non-NULL, because no compression is
* indicated by ssh_comp_none. But this setup call may return a
* null in_decomp. */
s->in_decomp = ssh_decompressor_new(compression);
if (s->in_decomp)
bpp_logevent(("Initialised %s decompression",
ssh_decompressor_alg(s->in_decomp)->text_name));
}
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Clear the pending_newkeys flag, so that handle_input below will
* start consuming the input data again. */
s->pending_newkeys = FALSE;
}
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
int ssh2_bpp_rekey_inadvisable(BinaryPacketProtocol *bpp)
{
struct ssh2_bpp_state *s;
assert(bpp->vt == &ssh2_bpp_vtable);
s = container_of(bpp, struct ssh2_bpp_state, bpp);
return s->pending_compression;
}
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
#define BPP_READ(ptr, len) do \
{ \
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
crMaybeWaitUntilV(s->bpp.input_eof || \
bufchain_try_fetch_consume( \
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
s->bpp.in_raw, ptr, len)); \
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
if (s->bpp.input_eof) \
goto eof; \
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
} while (0)
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
#define userauth_range(pkttype) ((unsigned)((pkttype) - 50) < 20)
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
static void ssh2_bpp_handle_input(BinaryPacketProtocol *bpp)
{
Rename FROMFIELD to 'container_of'. Ian Jackson points out that the Linux kernel has a macro of this name with the same purpose, and suggests that it's a good idea to use the same name as they do, so that at least some people reading one code base might recognise it from the other. I never really thought very hard about what order FROMFIELD's parameters should go in, and therefore I'm pleasantly surprised to find that my order agrees with the kernel's, so I don't have to permute every call site as part of making this change :-)
2018-10-05 22:49:08 +00:00
struct ssh2_bpp_state *s = container_of(bpp, struct ssh2_bpp_state, bpp);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crBegin(s->crState);
while (1) {
s->maxlen = 0;
s->length = 0;
if (s->in.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
s->cipherblk = ssh2_cipher_alg(s->in.cipher)->blksize;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
else
s->cipherblk = 8;
if (s->cipherblk < 8)
s->cipherblk = 8;
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
s->maclen = s->in.mac ? ssh2_mac_alg(s->in.mac)->len : 0;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->in.cipher &&
(ssh2_cipher_alg(s->in.cipher)->flags & SSH_CIPHER_IS_CBC) &&
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->in.mac && !s->in.etm_mode) {
/*
* When dealing with a CBC-mode cipher, we want to avoid the
* possibility of an attacker's tweaking the ciphertext stream
* so as to cause us to feed the same block to the block
* cipher more than once and thus leak information
* (VU#958563). The way we do this is not to take any
* decisions on the basis of anything we've decrypted until
* we've verified it with a MAC. That includes the packet
* length, so we just read data and check the MAC repeatedly,
* and when the MAC passes, see if the length we've got is
* plausible.
*
* This defence is unnecessary in OpenSSH ETM mode, because
* the whole point of ETM mode is that the attacker can't
* tweak the ciphertext stream at all without the MAC
* detecting it before we decrypt anything.
*/
/*
* Make sure we have buffer space for a maximum-size packet.
*/
unsigned buflimit = OUR_V2_PACKETLIMIT + s->maclen;
if (s->bufsize < buflimit) {
s->bufsize = buflimit;
s->buf = sresize(s->buf, s->bufsize, unsigned char);
}
/* Read an amount corresponding to the MAC. */
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->buf, s->maclen);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->packetlen = 0;
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
ssh2_mac_start(s->in.mac);
put_uint32(s->in.mac, s->in.sequence);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
for (;;) { /* Once around this loop per cipher block. */
/* Read another cipher-block's worth, and tack it on to
* the end. */
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->buf + (s->packetlen + s->maclen), s->cipherblk);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Decrypt one more block (a little further back in
* the stream). */
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_decrypt(s->in.cipher,
s->buf + s->packetlen, s->cipherblk);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Feed that block to the MAC. */
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
put_data(s->in.mac,
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->buf + s->packetlen, s->cipherblk);
s->packetlen += s->cipherblk;
/* See if that gives us a valid packet. */
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (ssh2_mac_verresult(s->in.mac, s->buf + s->packetlen) &&
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
((s->len = toint(GET_32BIT(s->buf))) ==
s->packetlen-4))
break;
if (s->packetlen >= (long)OUR_V2_PACKETLIMIT) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh,
"No valid incoming packet found");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
}
s->maxlen = s->packetlen + s->maclen;
/*
* Now transfer the data into an output packet.
*/
s->pktin = snew_plus(PktIn, s->maxlen);
s->pktin->qnode.prev = s->pktin->qnode.next = NULL;
s->pktin->type = 0;
Replace PktIn reference count with a 'free queue'. This is a new idea I've had to make memory-management of PktIn even easier. The idea is that a PktIn is essentially _always_ an element of some linked-list queue: if it's not one of the queues by which packets move through ssh.c, then it's a special 'free queue' which holds packets that are unowned and due to be freed. pq_pop() on a PktInQueue automatically relinks the packet to the free queue, and also triggers an idempotent callback which will empty the queue and really free all the packets on it. Hence, you can pop a packet off a real queue, parse it, handle it, and then just assume it'll get tidied up at some point - the only constraint being that you have to finish with it before returning to the application's main loop. The exception is that it's OK to pq_push() the packet back on to some other PktInQueue, because a side effect of that will be to _remove_ it from the free queue again. (And if _all_ the incoming packets get that treatment, then when the free-queue handler eventually runs, it may find it has nothing to do - which is harmless.)
2018-09-23 15:35:29 +00:00
s->pktin->qnode.on_free_queue = FALSE;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->data = snew_plus_get_aux(s->pktin);
memcpy(s->data, s->buf, s->maxlen);
} else if (s->in.mac && s->in.etm_mode) {
if (s->bufsize < 4) {
s->bufsize = 4;
s->buf = sresize(s->buf, s->bufsize, unsigned char);
}
/*
* OpenSSH encrypt-then-MAC mode: the packet length is
* unencrypted, unless the cipher supports length encryption.
*/
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->buf, 4);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Cipher supports length decryption, so do it */
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
if (s->in.cipher && (ssh2_cipher_alg(s->in.cipher)->flags &
SSH_CIPHER_SEPARATE_LENGTH)) {
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Keep the packet the same though, so the MAC passes */
unsigned char len[4];
memcpy(len, s->buf, 4);
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_decrypt_length(
s->in.cipher, len, 4, s->in.sequence);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->len = toint(GET_32BIT(len));
} else {
s->len = toint(GET_32BIT(s->buf));
}
/*
* _Completely_ silly lengths should be stomped on before they
* do us any more damage.
*/
if (s->len < 0 || s->len > (long)OUR_V2_PACKETLIMIT ||
s->len % s->cipherblk != 0) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh,
"Incoming packet length field was garbled");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
/*
* So now we can work out the total packet length.
*/
s->packetlen = s->len + 4;
/*
* Allocate the packet to return, now we know its length.
*/
s->pktin = snew_plus(PktIn, OUR_V2_PACKETLIMIT + s->maclen);
s->pktin->qnode.prev = s->pktin->qnode.next = NULL;
s->pktin->type = 0;
Replace PktIn reference count with a 'free queue'. This is a new idea I've had to make memory-management of PktIn even easier. The idea is that a PktIn is essentially _always_ an element of some linked-list queue: if it's not one of the queues by which packets move through ssh.c, then it's a special 'free queue' which holds packets that are unowned and due to be freed. pq_pop() on a PktInQueue automatically relinks the packet to the free queue, and also triggers an idempotent callback which will empty the queue and really free all the packets on it. Hence, you can pop a packet off a real queue, parse it, handle it, and then just assume it'll get tidied up at some point - the only constraint being that you have to finish with it before returning to the application's main loop. The exception is that it's OK to pq_push() the packet back on to some other PktInQueue, because a side effect of that will be to _remove_ it from the free queue again. (And if _all_ the incoming packets get that treatment, then when the free-queue handler eventually runs, it may find it has nothing to do - which is harmless.)
2018-09-23 15:35:29 +00:00
s->pktin->qnode.on_free_queue = FALSE;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->data = snew_plus_get_aux(s->pktin);
memcpy(s->data, s->buf, 4);
/*
* Read the remainder of the packet.
*/
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->data + 4, s->packetlen + s->maclen - 4);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/*
* Check the MAC.
*/
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->in.mac && !ssh2_mac_verify(
s->in.mac, s->data, s->len + 4, s->in.sequence)) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh, "Incorrect MAC received on packet");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
/* Decrypt everything between the length field and the MAC. */
if (s->in.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_decrypt(
s->in.cipher, s->data + 4, s->packetlen - 4);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
} else {
if (s->bufsize < s->cipherblk) {
s->bufsize = s->cipherblk;
s->buf = sresize(s->buf, s->bufsize, unsigned char);
}
/*
* Acquire and decrypt the first block of the packet. This will
* contain the length and padding details.
*/
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->buf, s->cipherblk);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (s->in.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_decrypt(
s->in.cipher, s->buf, s->cipherblk);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/*
* Now get the length figure.
*/
s->len = toint(GET_32BIT(s->buf));
/*
* _Completely_ silly lengths should be stomped on before they
* do us any more damage.
*/
if (s->len < 0 || s->len > (long)OUR_V2_PACKETLIMIT ||
(s->len + 4) % s->cipherblk != 0) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh,
"Incoming packet was garbled on decryption");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
/*
* So now we can work out the total packet length.
*/
s->packetlen = s->len + 4;
/*
* Allocate the packet to return, now we know its length.
*/
s->maxlen = s->packetlen + s->maclen;
s->pktin = snew_plus(PktIn, s->maxlen);
s->pktin->qnode.prev = s->pktin->qnode.next = NULL;
s->pktin->type = 0;
Replace PktIn reference count with a 'free queue'. This is a new idea I've had to make memory-management of PktIn even easier. The idea is that a PktIn is essentially _always_ an element of some linked-list queue: if it's not one of the queues by which packets move through ssh.c, then it's a special 'free queue' which holds packets that are unowned and due to be freed. pq_pop() on a PktInQueue automatically relinks the packet to the free queue, and also triggers an idempotent callback which will empty the queue and really free all the packets on it. Hence, you can pop a packet off a real queue, parse it, handle it, and then just assume it'll get tidied up at some point - the only constraint being that you have to finish with it before returning to the application's main loop. The exception is that it's OK to pq_push() the packet back on to some other PktInQueue, because a side effect of that will be to _remove_ it from the free queue again. (And if _all_ the incoming packets get that treatment, then when the free-queue handler eventually runs, it may find it has nothing to do - which is harmless.)
2018-09-23 15:35:29 +00:00
s->pktin->qnode.on_free_queue = FALSE;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->data = snew_plus_get_aux(s->pktin);
memcpy(s->data, s->buf, s->cipherblk);
/*
* Read and decrypt the remainder of the packet.
*/
Macroise the cumbersome read idioms in the BPPs. Now the three 'proper' BPPs each have a BPP_READ() macro that wraps up the fiddly combination of crMaybeWaitUntilV and bufchainery they use to read a fixed-length amount of input data. The sshverstring 'BPP' doesn't read fixed-length data in quite the same way, but it has a similar BPP_WAITFOR macro. No functional change. Mostly this is just a cleanup to make the code more legible, but also, the new macros will be a good place to centralise anything else that needs doing on every read, such as EOF checking.
2018-09-24 13:23:52 +00:00
BPP_READ(s->data + s->cipherblk,
s->packetlen + s->maclen - s->cipherblk);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
/* Decrypt everything _except_ the MAC. */
if (s->in.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_decrypt(
s->in.cipher,
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->data + s->cipherblk, s->packetlen - s->cipherblk);
/*
* Check the MAC.
*/
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
if (s->in.mac && !ssh2_mac_verify(
s->in.mac, s->data, s->len + 4, s->in.sequence)) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh, "Incorrect MAC received on packet");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
}
/* Get and sanity-check the amount of random padding. */
s->pad = s->data[4];
if (s->pad < 4 || s->len - s->pad < 1) {
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
ssh_sw_abort(s->bpp.ssh,
"Invalid padding length on received packet");
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crStopV;
}
/*
* This enables us to deduce the payload length.
*/
s->payload = s->len - s->pad - 1;
s->length = s->payload + 5;
New system for tracking data-limit-based rekeys. I've removed the encrypted_len fields from PktIn and PktOut, which were used to communicate from the BPP to ssh.c how much each packet contributed to the amount of data encrypted with a given set of cipher keys. It seems more sensible to have the BPP itself keep that counter - especially since only one of the three BPPs even needs to count it at all. So now there's a new DataTransferStats structure which the BPP updates, and ssh.c only needs to check it for overflow and reset the limits.
2018-09-19 20:28:21 +00:00
DTS_CONSUME(s->stats, in, s->packetlen);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->pktin->sequence = s->in.sequence++;
s->length = s->packetlen - s->pad;
assert(s->length >= 0);
/*
* Decompress packet payload.
*/
{
unsigned char *newpayload;
int newlen;
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->in_decomp && ssh_decompressor_decompress(
s->in_decomp, s->data + 5, s->length - 5,
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
&newpayload, &newlen)) {
if (s->maxlen < newlen + 5) {
PktIn *old_pktin = s->pktin;
s->maxlen = newlen + 5;
s->pktin = snew_plus(PktIn, s->maxlen);
*s->pktin = *old_pktin; /* structure copy */
s->data = snew_plus_get_aux(s->pktin);
smemclr(old_pktin, s->packetlen + s->maclen);
sfree(old_pktin);
}
s->length = 5 + newlen;
memcpy(s->data + 5, newpayload, newlen);
sfree(newpayload);
}
}
/*
* Now we can identify the semantic content of the packet,
* and also the initial type byte.
*/
if (s->length <= 5) { /* == 5 we hope, but robustness */
/*
* RFC 4253 doesn't explicitly say that completely empty
* packets with no type byte are forbidden. We handle them
* here by giving them a type code larger than 0xFF, which
* will be picked up at the next layer and trigger
* SSH_MSG_UNIMPLEMENTED.
*/
s->pktin->type = SSH_MSG_NO_TYPE_CODE;
Fix log-censoring of incoming SSH-2 session data. The call to ssh2_censor_packet for incoming packets in ssh2bpp was passing the wrong starting position in the packet data - in particular, not the same starting position as the adjacent call to log_packet - so the censor couldn't parse SSH2_MSG_CHANNEL_DATA to identify the string of session data that it should be bleeping out.
2018-09-26 06:39:04 +00:00
s->data += 5;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->length = 0;
} else {
s->pktin->type = s->data[5];
Fix log-censoring of incoming SSH-2 session data. The call to ssh2_censor_packet for incoming packets in ssh2bpp was passing the wrong starting position in the packet data - in particular, not the same starting position as the adjacent call to log_packet - so the censor couldn't parse SSH2_MSG_CHANNEL_DATA to identify the string of session data that it should be bleeping out.
2018-09-26 06:39:04 +00:00
s->data += 6;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
s->length -= 6;
}
Fix log-censoring of incoming SSH-2 session data. The call to ssh2_censor_packet for incoming packets in ssh2bpp was passing the wrong starting position in the packet data - in particular, not the same starting position as the adjacent call to log_packet - so the censor couldn't parse SSH2_MSG_CHANNEL_DATA to identify the string of session data that it should be bleeping out.
2018-09-26 06:39:04 +00:00
BinarySource_INIT(s->pktin, s->data, s->length);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (s->bpp.logctx) {
logblank_t blanks[MAX_BLANKS];
int nblanks = ssh2_censor_packet(
s->bpp.pls, s->pktin->type, FALSE,
make_ptrlen(s->data, s->length), blanks);
log_packet(s->bpp.logctx, PKT_INCOMING, s->pktin->type,
ssh2_pkt_type(s->bpp.pls->kctx, s->bpp.pls->actx,
s->pktin->type),
Fix log-censoring of incoming SSH-2 session data. The call to ssh2_censor_packet for incoming packets in ssh2bpp was passing the wrong starting position in the packet data - in particular, not the same starting position as the adjacent call to log_packet - so the censor couldn't parse SSH2_MSG_CHANNEL_DATA to identify the string of session data that it should be bleeping out.
2018-09-26 06:39:04 +00:00
s->data, s->length, nblanks, blanks,
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
&s->pktin->sequence, 0, NULL);
}
Move SSH_MSG_UNEXPECTED generation into the BPP. Now I've got a list macro defining all the packet types we recognise, I can use it to write a test for 'is this a recognised code?', and use that in turn to centralise detection of completely unrecognised codes into the binary packet protocol, where any such messages will be handled entirely internally and never even be seen by the next level up. This lets me remove another big pile of boilerplate in ssh.c.
2018-09-24 12:45:10 +00:00
if (ssh2_bpp_check_unimplemented(&s->bpp, s->pktin)) {
sfree(s->pktin);
s->pktin = NULL;
continue;
}
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
pq_push(&s->bpp.in_pq, s->pktin);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
{
int type = s->pktin->type;
s->pktin = NULL;
if (type == SSH2_MSG_NEWKEYS) {
/*
* Mild layer violation: in this situation we must
* suspend processing of the input byte stream until
* the transport layer has initialised the new keys by
* calling ssh2_bpp_new_incoming_crypto above.
*/
s->pending_newkeys = TRUE;
crWaitUntilV(!s->pending_newkeys);
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
continue;
}
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
if (type == SSH2_MSG_USERAUTH_SUCCESS) {
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
/*
* Another one: if we were configured with OpenSSH's
* deferred compression which is triggered on receipt
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
* of USERAUTH_SUCCESS, then this is the moment to
* turn on compression.
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
*/
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
if (s->in.pending_compression) {
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
s->in_decomp =
ssh_decompressor_new(s->in.pending_compression);
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
bpp_logevent(("Initialised delayed %s decompression",
ssh_decompressor_alg(
s->in_decomp)->text_name));
s->in.pending_compression = NULL;
}
if (s->out.pending_compression) {
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
s->out_comp =
ssh_compressor_new(s->out.pending_compression);
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
bpp_logevent(("Initialised delayed %s compression",
ssh_compressor_alg(
s->out_comp)->text_name));
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
s->out.pending_compression = NULL;
}
/*
* Whether or not we were doing delayed compression in
* _this_ set of crypto parameters, we should set a
* flag indicating that we're now authenticated, so
* that a delayed compression method enabled in any
* future rekey will be treated as un-delayed.
*/
s->seen_userauth_success = TRUE;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
if (s->pending_compression && userauth_range(type)) {
/*
* Receiving any userauth message at all indicates
* that we're not about to turn on delayed compression
* - either because we just _have_ done, or because
* this message is a USERAUTH_FAILURE or some kind of
* intermediate 'please send more data' continuation
* message. Either way, we turn off the outgoing
* packet blockage for now, and release any queued
* output packets, so that we can make another attempt
* to authenticate. The next userauth packet we send
* will re-block the output direction.
*/
s->pending_compression = FALSE;
queue_idempotent_callback(&s->bpp.ic_out_pq);
}
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
}
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
eof:
if (!s->bpp.expect_close) {
ssh_remote_error(s->bpp.ssh,
"Server unexpectedly closed network connection");
} else {
ssh_remote_eof(s->bpp.ssh, "Server closed network connection");
}
Fix use-after-free on a network error. When any BPP calls ssh_remote_error or ssh_remote_eof, it triggers an immediate cleanup of the BPP itself - so on return from one of those functions we should avoid going straight to the crFinish macro, because that will write to s->crState, which no longer exists.
2018-09-28 10:26:26 +00:00
return; /* avoid touching s now it's been freed */
Move most of ssh.c out into separate source files. I've tried to separate out as many individually coherent changes from this work as I could into their own commits, but here's where I run out and have to commit the rest of this major refactoring as a big-bang change. Most of ssh.c is now no longer in ssh.c: all five of the main coroutines that handle layers of the SSH-1 and SSH-2 protocols now each have their own source file to live in, and a lot of the supporting functions have moved into the appropriate one of those too. The new abstraction is a vtable called 'PacketProtocolLayer', which has an input and output packet queue. Each layer's main coroutine is invoked from the method ssh_ppl_process_queue(), which is usually (though not exclusively) triggered automatically when things are pushed on the input queue. In SSH-2, the base layer is the transport protocol, and it contains a pair of subsidiary queues by which it passes some of its packets to the higher SSH-2 layers - first userauth and then connection, which are peers at the same level, with the former abdicating in favour of the latter at the appropriate moment. SSH-1 is simpler: the whole login phase of the protocol (crypto setup and authentication) is all in one module, and since SSH-1 has no repeat key exchange, that setup layer abdicates in favour of the connection phase when it's done. ssh.c itself is now about a tenth of its old size (which all by itself is cause for celebration!). Its main job is to set up all the layers, hook them up to each other and to the BPP, and to funnel data back and forth between that collection of modules and external things such as the network and the terminal. Once it's set up a collection of packet protocol layers, it communicates with them partly by calling methods of the base layer (and if that's ssh2transport then it will delegate some functionality to the corresponding methods of its higher layer), and partly by talking directly to the connection layer no matter where it is in the stack by means of the separate ConnectionLayer vtable which I introduced in commit 8001dd4cb, and to which I've now added quite a few extra methods replacing services that used to be internal function calls within ssh.c. (One effect of this is that the SSH-1 and SSH-2 channel storage is now no longer shared - there are distinct struct types ssh1_channel and ssh2_channel. That means a bit more code duplication, but on the plus side, a lot fewer confusing conditionals in the middle of half-shared functions, and less risk of a piece of SSH-1 escaping into SSH-2 or vice versa, which I remember has happened at least once in the past.) The bulk of this commit introduces the five new source files, their common header sshppl.h and some shared supporting routines in sshcommon.c, and rewrites nearly all of ssh.c itself. But it also includes a couple of other changes that I couldn't separate easily enough: Firstly, there's a new handling for socket EOF, in which ssh.c sets an 'input_eof' flag in the BPP, and that responds by checking a flag that tells it whether to report the EOF as an error or not. (This is the main reason for those new BPP_READ / BPP_WAITFOR macros - they can check the EOF flag every time the coroutine is resumed.) Secondly, the error reporting itself is changed around again. I'd expected to put some data fields in the public PacketProtocolLayer structure that it could set to report errors in the same way as the BPPs have been doing, but in the end, I decided propagating all those data fields around was a pain and that even the BPPs shouldn't have been doing it that way. So I've reverted to a system where everything calls back to functions in ssh.c itself to report any connection- ending condition. But there's a new family of those functions, categorising the possible such conditions by semantics, and each one has a different set of detailed effects (e.g. how rudely to close the network connection, what exit status should be passed back to the whole application, whether to send a disconnect message and/or display a GUI error box). I don't expect this to be immediately perfect: of course, the code has been through a big upheaval, new bugs are expected, and I haven't been able to do a full job of testing (e.g. I haven't tested every auth or kex method). But I've checked that it _basically_ works - both SSH protocols, all the different kinds of forwarding channel, more than one auth method, Windows and Linux, connection sharing - and I think it's now at the point where the easiest way to find further bugs is to let it out into the wild and see what users can spot.
2018-09-24 17:28:16 +00:00
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
crFinishV;
}
static PktOut *ssh2_bpp_new_pktout(int pkt_type)
{
PktOut *pkt = ssh_new_packet();
pkt->length = 5; /* space for packet length + padding length */
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
pkt->minlen = 0;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
pkt->type = pkt_type;
put_byte(pkt, pkt_type);
pkt->prefix = pkt->length;
return pkt;
}
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
static void ssh2_bpp_format_packet_inner(struct ssh2_bpp_state *s, PktOut *pkt)
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
{
int origlen, cipherblk, maclen, padding, unencrypted_prefix, i;
if (s->bpp.logctx) {
ptrlen pktdata = make_ptrlen(pkt->data + pkt->prefix,
pkt->length - pkt->prefix);
logblank_t blanks[MAX_BLANKS];
int nblanks = ssh2_censor_packet(
s->bpp.pls, pkt->type, TRUE, pktdata, blanks);
log_packet(s->bpp.logctx, PKT_OUTGOING, pkt->type,
ssh2_pkt_type(s->bpp.pls->kctx, s->bpp.pls->actx,
pkt->type),
pktdata.ptr, pktdata.len, nblanks, blanks, &s->out.sequence,
pkt->downstream_id, pkt->additional_log_text);
}
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
cipherblk = s->out.cipher ? ssh2_cipher_alg(s->out.cipher)->blksize : 8;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
cipherblk = cipherblk < 8 ? 8 : cipherblk; /* or 8 if blksize < 8 */
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (s->out_comp) {
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
unsigned char *newpayload;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
int minlen, newlen;
Make compression functions return void. The return value wasn't used to indicate failure; it only indicated whether any compression was being done at all or whether the compression method was ssh_comp_none, and we can tell the latter just as well by the fact that its init function returns a null context pointer.
2018-07-10 20:27:43 +00:00
/*
* Compress packet payload.
*/
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
minlen = pkt->minlen;
if (minlen) {
/*
* Work out how much compressed data we need (at least) to
* make the overall packet length come to pkt->minlen.
*/
if (s->out.mac)
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
minlen -= ssh2_mac_alg(s->out.mac)->len;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
minlen -= 8; /* length field + min padding */
}
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
ssh_compressor_compress(s->out_comp, pkt->data + 5, pkt->length - 5,
&newpayload, &newlen, minlen);
Make compression functions return void. The return value wasn't used to indicate failure; it only indicated whether any compression was being done at all or whether the compression method was ssh_comp_none, and we can tell the latter just as well by the fact that its init function returns a null context pointer.
2018-07-10 20:27:43 +00:00
pkt->length = 5;
put_data(pkt, newpayload, newlen);
sfree(newpayload);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
/*
* Add padding. At least four bytes, and must also bring total
* length (minus MAC) up to a multiple of the block size.
* If pkt->forcepad is set, make sure the packet is at least that size
* after padding.
*/
padding = 4;
unencrypted_prefix = (s->out.mac && s->out.etm_mode) ? 4 : 0;
padding +=
(cipherblk - (pkt->length - unencrypted_prefix + padding) % cipherblk)
% cipherblk;
assert(padding <= 255);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
maclen = s->out.mac ? ssh2_mac_alg(s->out.mac)->len : 0;
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
origlen = pkt->length;
for (i = 0; i < padding; i++)
put_byte(pkt, random_byte());
pkt->data[4] = padding;
PUT_32BIT(pkt->data, origlen + padding - 4);
/* Encrypt length if the scheme requires it */
if (s->out.cipher &&
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
(ssh2_cipher_alg(s->out.cipher)->flags & SSH_CIPHER_SEPARATE_LENGTH)) {
ssh2_cipher_encrypt_length(s->out.cipher, pkt->data, 4,
s->out.sequence);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
Make put_padding() have a consistent argument order. Thanks to Alex Landau for pointing out that commit 8b98fea4a introduced two uses of it with the arguments one way round and one with them the other way round. (Plus a fourth use where it doesn't matter, because the padding at the end of the encrypted blob of an OpenSSH PEM private key consists of n bytes with value n. :-) On the basis of majority vote, I've switched the order in the function definition to match the two of the three call sites that expressed the same opinion, and fixed the third.
2018-06-13 18:42:19 +00:00
put_padding(pkt, maclen, 0);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (s->out.mac && s->out.etm_mode) {
/*
* OpenSSH-defined encrypt-then-MAC protocol.
*/
if (s->out.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_encrypt(s->out.cipher,
pkt->data + 4, origlen + padding - 4);
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
ssh2_mac_generate(s->out.mac, pkt->data, origlen + padding,
s->out.sequence);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
} else {
/*
* SSH-2 standard protocol.
*/
if (s->out.mac)
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
ssh2_mac_generate(s->out.mac, pkt->data, origlen + padding,
s->out.sequence);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
if (s->out.cipher)
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
ssh2_cipher_encrypt(s->out.cipher, pkt->data, origlen + padding);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
s->out.sequence++; /* whether or not we MACed */
New system for tracking data-limit-based rekeys. I've removed the encrypted_len fields from PktIn and PktOut, which were used to communicate from the BPP to ssh.c how much each packet contributed to the amount of data encrypted with a given set of cipher keys. It seems more sensible to have the BPP itself keep that counter - especially since only one of the three BPPs even needs to count it at all. So now there's a new DataTransferStats structure which the BPP updates, and ssh.c only needs to check it for overflow and reset the limits.
2018-09-19 20:28:21 +00:00
DTS_CONSUME(s->stats, out, origlen + padding);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
}
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
static void ssh2_bpp_format_packet(struct ssh2_bpp_state *s, PktOut *pkt)
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
{
Make ssh_compress into a pair of linked classoids. This was mildly fiddly because there's a single vtable structure that implements two distinct interface types, one for compression and one for decompression - and I have actually confused them before now (commit d4304f1b7), so I think it's important to make them actually be separate types!
2018-09-14 08:16:41 +00:00
if (pkt->minlen > 0 && !s->out_comp) {
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
/*
* If we've been told to pad the packet out to a given minimum
* length, but we're not compressing (and hence can't get the
* compression to do the padding by pointlessly opening and
* closing zlib blocks), then our other strategy is to precede
* this message with an SSH_MSG_IGNORE that makes it up to the
* right length.
*
* A third option in principle, and the most obviously
* sensible, would be to set the explicit padding field in the
* packet to more than its minimum value. Sadly, that turns
* out to break some servers (our institutional memory thinks
* Cisco in particular) and so we abandoned that idea shortly
* after trying it.
*/
/*
* Calculate the length we expect the real packet to have.
*/
int block, length;
PktOut *ignore_pkt;
Turn SSH-2 ciphers into a classoid. This is more or less the same job as the SSH-1 case, only more extensive, because we have a wider range of ciphers. I'm a bit disappointed about the AES case, in particular, because I feel as if it ought to have been possible to arrange to combine this layer of vtable dispatch with the subsidiary one that selects between hardware and software implementations of the underlying cipher. I may come back later and have another try at that, in fact.
2018-09-13 13:43:04 +00:00
block = s->out.cipher ? ssh2_cipher_alg(s->out.cipher)->blksize : 0;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
if (block < 8)
block = 8;
length = pkt->length;
length += 4; /* minimum 4 byte padding */
length += block-1;
length -= (length % block);
if (s->out.mac)
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
length += ssh2_mac_alg(s->out.mac)->len;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
if (length < pkt->minlen) {
/*
* We need an ignore message. Calculate its length.
*/
length = pkt->minlen - length;
/*
* And work backwards from that to the length of the
* contained string.
*/
if (s->out.mac)
Turn SSH-2 MACs into a classoid. This piece of tidying-up has come out particularly well in terms of saving tedious repetition and boilerplate. I've managed to remove three pointless methods from every MAC implementation by means of writing them once centrally in terms of the implementation-specific methods; another method (hmacmd5_sink) vanished because I was able to make the interface type 'ssh2_mac' be directly usable as a BinarySink by way of a new delegation system; and because all the method implementations can now find their own vtable, I was even able to merge a lot of keying and output functions that had previously differed only in length parameters by having them look up the lengths in whatever vtable they were passed.
2018-09-13 15:15:17 +00:00
length -= ssh2_mac_alg(s->out.mac)->len;
Move password-packet padding into the BPP module. Now when we construct a packet containing sensitive data, we just set a field saying '... and make it take up at least this much space, to disguise its true size', and nothing in the rest of the system worries about that flag until ssh2bpp.c acts on it. Also, I've changed the strategy for doing the padding. Previously, we were following the real packet with an SSH_MSG_IGNORE to make up the size. But that was only a partial defence: it works OK against passive traffic analysis, but an attacker proxying the TCP stream and dribbling it out one byte at a time could still have found out the size of the real packet by noting when the dribbled data provoked a response. Now I put the SSH_MSG_IGNORE _first_, which should defeat that attack. But that in turn doesn't work when we're doing compression, because we can't predict the compressed sizes accurately enough to make that strategy sensible. Fortunately, compression provides an alternative strategy anyway: if we've got zlib turned on when we send one of these sensitive packets, then we can pad out the compressed zlib data as much as we like by adding empty RFC1951 blocks (effectively chaining ZLIB_PARTIAL_FLUSHes). So both strategies should now be dribble-proof.
2018-07-09 19:30:11 +00:00
length -= 8; /* length field + min padding */
length -= 5; /* type code + string length prefix */
if (length < 0)
length = 0;
ignore_pkt = ssh2_bpp_new_pktout(SSH2_MSG_IGNORE);
put_uint32(ignore_pkt, length);
while (length-- > 0)
put_byte(ignore_pkt, random_byte());
ssh2_bpp_format_packet_inner(s, ignore_pkt);
bufchain_add(s->bpp.out_raw, ignore_pkt->data, ignore_pkt->length);
ssh_free_pktout(ignore_pkt);
}
}
ssh2_bpp_format_packet_inner(s, pkt);
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
bufchain_add(s->bpp.out_raw, pkt->data, pkt->length);
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
}
Reinstate calls to ssh_free_pktout! I think ever since commit 679fa90df last month, PuTTY has been forgetting to free any of its outgoing packet structures after turning them into their encrypted wire format. And apparently no users of the development snapshots have noticed - including me!
2018-07-09 19:59:36 +00:00
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
static void ssh2_bpp_handle_output(BinaryPacketProtocol *bpp)
{
Rename FROMFIELD to 'container_of'. Ian Jackson points out that the Linux kernel has a macro of this name with the same purpose, and suggests that it's a good idea to use the same name as they do, so that at least some people reading one code base might recognise it from the other. I never really thought very hard about what order FROMFIELD's parameters should go in, and therefore I'm pleasantly surprised to find that my order agrees with the kernel's, so I don't have to permute every call site as part of making this change :-)
2018-10-05 22:49:08 +00:00
struct ssh2_bpp_state *s = container_of(bpp, struct ssh2_bpp_state, bpp);
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
PktOut *pkt;
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
int n_userauth;
/*
* Count the userauth packets in the queue.
*/
n_userauth = 0;
for (pkt = pq_first(&s->bpp.out_pq); pkt != NULL;
pkt = pq_next(&s->bpp.out_pq, pkt))
if (userauth_range(pkt->type))
n_userauth++;
if (s->pending_compression && !n_userauth) {
/*
* We're currently blocked from sending any outgoing packets
* until the other end tells us whether we're going to have to
* enable compression or not.
*
* If our end has pushed a userauth packet on the queue, that
* must mean it knows that a USERAUTH_SUCCESS is not
* immediately forthcoming, so we unblock ourselves and send
* up to and including that packet. But in this if statement,
* there aren't any, so we're still blocked.
*/
return;
}
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
if (s->cbc_ignore_workaround) {
/*
* When using a CBC-mode cipher in SSH-2, it's necessary to
* ensure that an attacker can't provide data to be encrypted
* using an IV that they know. We ensure this by inserting an
* SSH_MSG_IGNORE if the last cipher block of the previous
* packet has already been sent to the network (which we
* approximate conservatively by checking if it's vanished
* from out_raw).
*/
if (bufchain_size(s->bpp.out_raw) <
(ssh2_cipher_alg(s->out.cipher)->blksize +
ssh2_mac_alg(s->out.mac)->len)) {
/*
* There's less data in out_raw than the MAC size plus the
* cipher block size, which means at least one byte of
* that cipher block must already have left. Add an
* IGNORE.
*/
pkt = ssh_bpp_new_pktout(&s->bpp, SSH2_MSG_IGNORE);
put_stringz(pkt, "");
ssh2_bpp_format_packet(s, pkt);
}
}
while ((pkt = pq_pop(&s->bpp.out_pq)) != NULL) {
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
if (userauth_range(pkt->type))
n_userauth--;
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
ssh2_bpp_format_packet(s, pkt);
ssh_free_pktout(pkt);
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
Try to decouple directions of delayed compression. I ran 'git push' too soon on the last commit; after a bit more thought I realise that I didn't get the logic quite right in the case where one direction of the connection negotiates delayed compression and the other negotiates ordinary or no compression. For a start, we only need to worry about temporarily delaying outgoing packets to avoid the race condition if delayed compression applies to _outgoing_ packets - that can be disabled in the case where delayed compression is inbound only (though that is admittedly unlikely). Secondly, that means that detecting USERAUTH_SUCCESS to enable compression has to happen even if the output blockage wasn't in place. Thirdly, if we're independently enabling delayed compression in the two directions, we should only print an Event Log entry for the one we actually did! This revised version is probably more robust, although for the moment all of this is theoretical - I haven't tested against a server implementing unidirectional delayed compression.
2018-10-07 11:56:57 +00:00
if (n_userauth == 0 && s->out.pending_compression) {
Support OpenSSH delayed compression without a rekey. The problem with OpenSSH delayed compression is that the spec has a race condition. Compression is enabled when the server sends USERAUTH_SUCCESS. In the server->client direction, that's fine: the USERAUTH_SUCCESS packet is not itself compressed, and the next packet in the same direction is. But in the client->server direction, this specification relies on there being a moment of half-duplex in the connection: the client can't send any outgoing packet _after_ whatever userauth packet the USERAUTH_SUCCESS was a response to, and _before_ finding out whether the response is USERAUTH_SUCCESS or something else. If it emitted, say, an SSH_MSG_IGNORE or initiated a rekey (perhaps due to a timeout), then that might cross in the network with USERAUTH_SUCCESS and the server wouldn't be able to know whether to treat it as compressed. My previous solution was to note the presence of delayed compression options in the server KEXINIT, but not to negotiate them in the initial key exchange. Instead, we conduct the userauth exchange with compression="none", and then once userauth has concluded, we trigger an immediate rekey in which we do accept delayed compression methods - because of course by that time they're no different from the non- delayed versions. And that means compression is enabled by the bidirectional NEWKEYS exchange, which lacks that race condition. I think OpenSSH itself gets away with this because its layer structure is structure so as to never send any such asynchronous transport-layer message in the middle of userauth. Ours is not. But my cunning plan is that now that my BPP abstraction includes a queue of packets to be sent and a callback that processes that queue on to the output raw data bufchain, it's possible to make that callback terminate early, to leave any dangerous transport-layer messages unsent while we wait for a userauth response. Specifically: if we've negotiated a delayed compression method and not yet seen USERAUTH_SUCCESS, then ssh2_bpp_handle_output will emit all packets from its queue up to and including the last one in the userauth type-code range, and keep back any further ones. The idea is that _if_ that last userauth message was one that might provoke USERAUTH_SUCCESS, we don't want to send any difficult things after it; if it's not (e.g. it's in the middle of some ongoing userauth process like k-i or GSS) then the userauth layer will know that, and will emit some further userauth packet on its own initiative which will clue us in that it's OK to release everything up to and including that one. (So in particular it wasn't even necessary to forbid _all_ transport- layer packets during userauth. I could have done that by reordering the output queue - packets in that queue haven't been assigned their sequence numbers yet, so that would have been safe - but it's more elegant not to have to.) One particular case we do have to be careful about is not trying to initiate a _rekey_ during userauth, if delayed compression is in the offing. That's because when we start rekeying, ssh2transport stops sending any higher-layer packets at all, to discourage servers from trying to ignore the KEXINIT and press on regardless - you don't get your higher-layer replies until you actually respond to the lower-layer interrupt. But in this case, if ssh2transport sent a KEXINIT, which ssh2bpp kept back in the queue to avoid a delayed compression race and would only send if another userauth packet followed it, which ssh2transport would never pass on to ssh2bpp's output queue, there'd be a complete protocol deadlock. So instead I defer any attempt to start a rekey until after userauth finishes (using the existing system for starting a deferred rekey at that moment, which was previously used for the _old_ delayed-compression strategy, and still has to be here anyway for GSSAPI purposes).
2018-10-07 08:10:14 +00:00
/*
* This is the last userauth packet in the queue, so
* unless our side decides to send another one in future,
* we have to assume will potentially provoke
* USERAUTH_SUCCESS. Block (non-userauth) outgoing packets
* until we see the reply.
*/
s->pending_compression = TRUE;
return;
}
Give the BPP an input and output packet queue. Now, instead of writing each packet straight on to the raw output bufchain by calling the BPP's format_packet function, the higher protocol layers will put the packets on to a queue, which will automatically trigger a callback (using the new mechanism for embedding a callback in any packet queue) to make the BPP format its queue on to the raw-output bufchain. That in turn triggers a second callback which moves the data to the socket. This means in particular that the CBC ignore-message workaround can be moved into the new BPP routine to process the output queue, which is a good place for it because then it can easily arrange to only put an ignore message at the start of any sequence of packets that are being formatted as a single output blob.
2018-09-24 17:08:09 +00:00
}
Move binary packet protocols and censoring out of ssh.c. sshbpp.h now defines a classoid that encapsulates both directions of an SSH binary packet protocol - that is, a system for reading a bufchain of incoming data and turning it into a stream of PktIn, and another system for taking a PktOut and turning it into data on an outgoing bufchain. The state structure in each of those files contains everything that used to be in the 'rdpkt2_state' structure and its friends, and also quite a lot of bits and pieces like cipher and MAC states that used to live in the main Ssh structure. One minor effect of this layer separation is that I've had to extend the packet dispatch table by one, because the BPP layer can no longer directly trigger sending of SSH_MSG_UNIMPLEMENTED for a message too short to have a type byte. Instead, I extend the PktIn type field to use an out-of-range value to encode that, and the easiest way to make that trigger an UNIMPLEMENTED message is to have the dispatch table contain an entry for it. (That's a system that may come in useful again - I was also wondering about inventing a fake type code to indicate network EOF, so that that could be propagated through the layers and be handled by whichever one currently knew best how to respond.) I've also moved the packet-censoring code into its own pair of files, partly because I was going to want to do that anyway sooner or later, and mostly because it's called from the BPP code, and the SSH-2 version in particular has to be called from both the main SSH-2 BPP and the bare unencrypted protocol used for connection sharing. While I was at it, I took the opportunity to merge the outgoing and incoming censor functions, so that the parts that were common between them (e.g. CHANNEL_DATA messages look the same in both directions) didn't need to be repeated.
2018-06-09 08:09:10 +00:00
}
Reference in New Issue Copy Permalink