putty/sshrsa.c

1021 строка
28 KiB
C
Исходник Обычный вид История

/*
* RSA implementation for PuTTY.
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include "ssh.h"
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
#include "mpint.h"
#include "misc.h"
void BinarySource_get_rsa_ssh1_pub(
BinarySource *src, RSAKey *rsa, RsaSsh1Order order)
{
unsigned bits;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *e, *m;
bits = get_uint32(src);
if (order == RSA_SSH1_EXPONENT_FIRST) {
e = get_mp_ssh1(src);
m = get_mp_ssh1(src);
} else {
m = get_mp_ssh1(src);
e = get_mp_ssh1(src);
}
if (rsa) {
rsa->bits = bits;
rsa->exponent = e;
rsa->modulus = m;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
rsa->bytes = (mp_get_nbits(m) + 7) / 8;
} else {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(e);
mp_free(m);
}
}
void BinarySource_get_rsa_ssh1_priv(
BinarySource *src, RSAKey *rsa)
{
rsa->private_exponent = get_mp_ssh1(src);
}
bool rsa_ssh1_encrypt(unsigned char *data, int length, RSAKey *key)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *b1, *b2;
int i;
unsigned char *p;
if (key->bytes < length + 4)
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
return false; /* RSA key too short! */
memmove(data + key->bytes - length, data, length);
data[0] = 0;
data[1] = 2;
size_t npad = key->bytes - length - 3;
/*
* Generate a sequence of nonzero padding bytes. We do this in a
* reasonably uniform way and without having to loop round
* retrying the random number generation, by first generating an
* integer in [0,2^n) for an appropriately large n; then we
* repeatedly multiply by 255 to give an integer in [0,255*2^n),
* extract the top 8 bits to give an integer in [0,255), and mask
* those bits off before multiplying up again for the next digit.
* This gives us a sequence of numbers in [0,255), and of course
* adding 1 to each of them gives numbers in [1,256) as we wanted.
*
* (You could imagine this being a sort of fixed-point operation:
* given a uniformly random binary _fraction_, multiplying it by k
* and subtracting off the integer part will yield you a sequence
* of integers each in [0,k). I'm just doing that scaled up by a
* power of 2 to avoid the fractions.)
*/
size_t random_bits = (npad + 16) * 8;
mp_int *randval = mp_new(random_bits + 8);
mp_int *tmp = mp_random_bits(random_bits);
mp_copy_into(randval, tmp);
mp_free(tmp);
for (i = 2; i < key->bytes - length - 1; i++) {
mp_mul_integer_into(randval, randval, 255);
uint8_t byte = mp_get_byte(randval, random_bits / 8);
assert(byte != 255);
data[i] = byte + 1;
mp_reduce_mod_2to(randval, random_bits);
}
mp_free(randval);
data[key->bytes - length - 1] = 0;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
b1 = mp_from_bytes_be(make_ptrlen(data, key->bytes));
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
b2 = mp_modpow(b1, key->exponent, key->modulus);
p = data;
for (i = key->bytes; i--;) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
*p++ = mp_get_byte(b2, i);
}
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(b1);
mp_free(b2);
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
return true;
}
/*
* Compute (base ^ exp) % mod, provided mod == p * q, with p,q
* distinct primes, and iqmp is the multiplicative inverse of q mod p.
* Uses Chinese Remainder Theorem to speed computation up over the
* obvious implementation of a single big modpow.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *crt_modpow(mp_int *base, mp_int *exp, mp_int *mod,
mp_int *p, mp_int *q, mp_int *iqmp)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *pm1, *qm1, *pexp, *qexp, *presult, *qresult;
mp_int *diff, *multiplier, *ret0, *ret;
/*
* Reduce the exponent mod phi(p) and phi(q), to save time when
* exponentiating mod p and mod q respectively. Of course, since p
* and q are prime, phi(p) == p-1 and similarly for q.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
pm1 = mp_copy(p);
mp_sub_integer_into(pm1, pm1, 1);
qm1 = mp_copy(q);
mp_sub_integer_into(qm1, qm1, 1);
pexp = mp_mod(exp, pm1);
qexp = mp_mod(exp, qm1);
/*
* Do the two modpows.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *base_mod_p = mp_mod(base, p);
presult = mp_modpow(base_mod_p, pexp, p);
mp_free(base_mod_p);
mp_int *base_mod_q = mp_mod(base, q);
qresult = mp_modpow(base_mod_q, qexp, q);
mp_free(base_mod_q);
/*
* Recombine the results. We want a value which is congruent to
* qresult mod q, and to presult mod p.
*
* We know that iqmp * q is congruent to 1 * mod p (by definition
* of iqmp) and to 0 mod q (obviously). So we start with qresult
* (which is congruent to qresult mod both primes), and add on
* (presult-qresult) * (iqmp * q) which adjusts it to be congruent
* to presult mod p without affecting its value mod q.
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
*
* (If presult-qresult < 0, we add p to it to keep it positive.)
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
unsigned presult_too_small = mp_cmp_hs(qresult, presult);
mp_cond_add_into(presult, presult, p, presult_too_small);
diff = mp_sub(presult, qresult);
multiplier = mp_mul(iqmp, q);
ret0 = mp_mul(multiplier, diff);
mp_add_into(ret0, ret0, qresult);
/*
* Finally, reduce the result mod n.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
ret = mp_mod(ret0, mod);
/*
* Free all the intermediate results before returning.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(pm1);
mp_free(qm1);
mp_free(pexp);
mp_free(qexp);
mp_free(presult);
mp_free(qresult);
mp_free(diff);
mp_free(multiplier);
mp_free(ret0);
return ret;
}
/*
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
* Wrapper on crt_modpow that looks up all the right values from an
* RSAKey.
*/
static mp_int *rsa_privkey_op(mp_int *input, RSAKey *key)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
return crt_modpow(input, key->private_exponent,
key->modulus, key->p, key->q, key->iqmp);
}
mp_int *rsa_ssh1_decrypt(mp_int *input, RSAKey *key)
{
return rsa_privkey_op(input, key);
}
bool rsa_ssh1_decrypt_pkcs1(mp_int *input, RSAKey *key,
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
strbuf *outbuf)
{
strbuf *data = strbuf_new_nm();
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
bool success = false;
BinarySource src[1];
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *b = rsa_ssh1_decrypt(input, key);
for (size_t i = (mp_get_nbits(key->modulus) + 7) / 8; i-- > 0 ;) {
put_byte(data, mp_get_byte(b, i));
}
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(b);
}
BinarySource_BARE_INIT(src, data->u, data->len);
/* Check PKCS#1 formatting prefix */
if (get_byte(src) != 0) goto out;
if (get_byte(src) != 2) goto out;
while (1) {
unsigned char byte = get_byte(src);
if (get_err(src)) goto out;
if (byte == 0)
break;
}
/* Everything else is the payload */
success = true;
put_data(outbuf, get_ptr(src), get_avail(src));
out:
strbuf_free(data);
return success;
}
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
static void append_hex_to_strbuf(strbuf *sb, mp_int *x)
{
if (sb->len > 0)
put_byte(sb, ',');
put_data(sb, "0x", 2);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
char *hex = mp_get_hex(x);
size_t hexlen = strlen(hex);
put_data(sb, hex, hexlen);
smemclr(hex, hexlen);
sfree(hex);
}
char *rsastr_fmt(RSAKey *key)
{
strbuf *sb = strbuf_new();
append_hex_to_strbuf(sb, key->exponent);
append_hex_to_strbuf(sb, key->modulus);
return strbuf_to_str(sb);
}
/*
* Generate a fingerprint string for the key. Compatible with the
* OpenSSH fingerprint code.
*/
char *rsa_ssh1_fingerprint(RSAKey *key)
{
unsigned char digest[16];
strbuf *out;
int i;
/*
* The hash preimage for SSH-1 key fingerprinting consists of the
* modulus and exponent _without_ any preceding length field -
* just the minimum number of bytes to represent each integer,
* stored big-endian, concatenated with no marker at the division
* between them.
*/
ssh_hash *hash = ssh_hash_new(&ssh_md5);
for (size_t i = (mp_get_nbits(key->modulus) + 7) / 8; i-- > 0 ;)
put_byte(hash, mp_get_byte(key->modulus, i));
for (size_t i = (mp_get_nbits(key->exponent) + 7) / 8; i-- > 0 ;)
put_byte(hash, mp_get_byte(key->exponent, i));
ssh_hash_final(hash, digest);
out = strbuf_new();
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
strbuf_catf(out, "%d ", mp_get_nbits(key->modulus));
for (i = 0; i < 16; i++)
strbuf_catf(out, "%s%02x", i ? ":" : "", digest[i]);
if (key->comment)
strbuf_catf(out, " %s", key->comment);
return strbuf_to_str(out);
}
/*
* Verify that the public data in an RSA key matches the private
* data. We also check the private data itself: we ensure that p >
* q and that iqmp really is the inverse of q mod p.
*/
bool rsa_verify(RSAKey *key)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *n, *ed, *pm1, *qm1;
unsigned ok = 1;
/* Preliminary checks: p,q can't be 0 or 1. (Of course no other
* very small value is any good either, but these are the values
* we _must_ check for to avoid assertion failures further down
* this function.) */
if (!(mp_hs_integer(key->p, 2) & mp_hs_integer(key->q, 2)))
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
return false;
/* n must equal pq. */
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
n = mp_mul(key->p, key->q);
ok &= mp_cmp_eq(n, key->modulus);
mp_free(n);
/* e * d must be congruent to 1, modulo (p-1) and modulo (q-1). */
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
pm1 = mp_copy(key->p);
mp_sub_integer_into(pm1, pm1, 1);
ed = mp_modmul(key->exponent, key->private_exponent, pm1);
mp_free(pm1);
ok &= mp_eq_integer(ed, 1);
mp_free(ed);
qm1 = mp_copy(key->q);
mp_sub_integer_into(qm1, qm1, 1);
ed = mp_modmul(key->exponent, key->private_exponent, qm1);
mp_free(qm1);
ok &= mp_eq_integer(ed, 1);
mp_free(ed);
/*
* Ensure p > q.
*
* I have seen key blobs in the wild which were generated with
* p < q, so instead of rejecting the key in this case we
* should instead flip them round into the canonical order of
* p > q. This also involves regenerating iqmp.
*/
mp_int *p_new = mp_max(key->p, key->q);
mp_int *q_new = mp_min(key->p, key->q);
mp_free(key->p);
mp_free(key->q);
mp_free(key->iqmp);
key->p = p_new;
key->q = q_new;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
key->iqmp = mp_invert(key->q, key->p);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
return ok;
}
void rsa_ssh1_public_blob(BinarySink *bs, RSAKey *key,
RsaSsh1Order order)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
put_uint32(bs, mp_get_nbits(key->modulus));
if (order == RSA_SSH1_EXPONENT_FIRST) {
put_mp_ssh1(bs, key->exponent);
put_mp_ssh1(bs, key->modulus);
} else {
put_mp_ssh1(bs, key->modulus);
put_mp_ssh1(bs, key->exponent);
}
}
/* Given an SSH-1 public key blob, determine its length. */
int rsa_ssh1_public_blob_len(ptrlen data)
{
BinarySource src[1];
BinarySource_BARE_INIT_PL(src, data);
/* Expect a length word, then exponent and modulus. (It doesn't
* even matter which order.) */
get_uint32(src);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(get_mp_ssh1(src));
mp_free(get_mp_ssh1(src));
if (get_err(src))
return -1;
/* Return the number of bytes consumed. */
return src->pos;
}
void freersapriv(RSAKey *key)
{
if (key->private_exponent) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->private_exponent);
key->private_exponent = NULL;
}
if (key->p) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->p);
key->p = NULL;
}
if (key->q) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->q);
key->q = NULL;
}
if (key->iqmp) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->iqmp);
key->iqmp = NULL;
}
}
void freersakey(RSAKey *key)
{
freersapriv(key);
if (key->modulus) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->modulus);
key->modulus = NULL;
}
if (key->exponent) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(key->exponent);
key->exponent = NULL;
}
if (key->comment) {
sfree(key->comment);
key->comment = NULL;
}
}
/* ----------------------------------------------------------------------
* Implementation of the ssh-rsa signing key type.
*/
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
static void rsa2_freekey(ssh_key *key); /* forward reference */
static ssh_key *rsa2_new_pub(const ssh_keyalg *self, ptrlen data)
{
BinarySource src[1];
RSAKey *rsa;
BinarySource_BARE_INIT_PL(src, data);
if (!ptrlen_eq_string(get_string(src), "ssh-rsa"))
return NULL;
rsa = snew(RSAKey);
rsa->sshk.vt = &ssh_rsa;
rsa->exponent = get_mp_ssh2(src);
rsa->modulus = get_mp_ssh2(src);
rsa->private_exponent = NULL;
rsa->p = rsa->q = rsa->iqmp = NULL;
rsa->comment = NULL;
if (get_err(src)) {
rsa2_freekey(&rsa->sshk);
return NULL;
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
return &rsa->sshk;
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
static void rsa2_freekey(ssh_key *key)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
freersakey(rsa);
sfree(rsa);
}
static char *rsa2_cache_str(ssh_key *key)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
return rsastr_fmt(rsa);
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
static void rsa2_public_blob(ssh_key *key, BinarySink *bs)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
put_stringz(bs, "ssh-rsa");
put_mp_ssh2(bs, rsa->exponent);
put_mp_ssh2(bs, rsa->modulus);
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
static void rsa2_private_blob(ssh_key *key, BinarySink *bs)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
put_mp_ssh2(bs, rsa->private_exponent);
put_mp_ssh2(bs, rsa->p);
put_mp_ssh2(bs, rsa->q);
put_mp_ssh2(bs, rsa->iqmp);
}
static ssh_key *rsa2_new_priv(const ssh_keyalg *self,
ptrlen pub, ptrlen priv)
{
BinarySource src[1];
ssh_key *sshk;
RSAKey *rsa;
sshk = rsa2_new_pub(self, pub);
if (!sshk)
return NULL;
rsa = container_of(sshk, RSAKey, sshk);
BinarySource_BARE_INIT_PL(src, priv);
rsa->private_exponent = get_mp_ssh2(src);
rsa->p = get_mp_ssh2(src);
rsa->q = get_mp_ssh2(src);
rsa->iqmp = get_mp_ssh2(src);
if (get_err(src) || !rsa_verify(rsa)) {
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
rsa2_freekey(&rsa->sshk);
return NULL;
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
return &rsa->sshk;
}
static ssh_key *rsa2_new_priv_openssh(const ssh_keyalg *self,
BinarySource *src)
{
RSAKey *rsa;
rsa = snew(RSAKey);
rsa->sshk.vt = &ssh_rsa;
rsa->comment = NULL;
rsa->modulus = get_mp_ssh2(src);
rsa->exponent = get_mp_ssh2(src);
rsa->private_exponent = get_mp_ssh2(src);
rsa->iqmp = get_mp_ssh2(src);
rsa->p = get_mp_ssh2(src);
rsa->q = get_mp_ssh2(src);
if (get_err(src) || !rsa_verify(rsa)) {
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
rsa2_freekey(&rsa->sshk);
return NULL;
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
return &rsa->sshk;
}
static void rsa2_openssh_blob(ssh_key *key, BinarySink *bs)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
put_mp_ssh2(bs, rsa->modulus);
put_mp_ssh2(bs, rsa->exponent);
put_mp_ssh2(bs, rsa->private_exponent);
put_mp_ssh2(bs, rsa->iqmp);
put_mp_ssh2(bs, rsa->p);
put_mp_ssh2(bs, rsa->q);
}
static int rsa2_pubkey_bits(const ssh_keyalg *self, ptrlen pub)
{
ssh_key *sshk;
RSAKey *rsa;
int ret;
sshk = rsa2_new_pub(self, pub);
if (!sshk)
return -1;
rsa = container_of(sshk, RSAKey, sshk);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
ret = mp_get_nbits(rsa->modulus);
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
rsa2_freekey(&rsa->sshk);
return ret;
}
static inline const ssh_hashalg *rsa2_hash_alg_for_flags(
unsigned flags, const char **protocol_id_out)
{
const ssh_hashalg *halg;
const char *protocol_id;
if (flags & SSH_AGENT_RSA_SHA2_256) {
halg = &ssh_sha256;
protocol_id = "rsa-sha2-256";
} else if (flags & SSH_AGENT_RSA_SHA2_512) {
halg = &ssh_sha512;
protocol_id = "rsa-sha2-512";
} else {
halg = &ssh_sha1;
protocol_id = "ssh-rsa";
}
if (protocol_id_out)
*protocol_id_out = protocol_id;
return halg;
}
static inline ptrlen rsa_pkcs1_prefix_for_hash(const ssh_hashalg *halg)
{
if (halg == &ssh_sha1) {
/*
* This is the magic ASN.1/DER prefix that goes in the decoded
* signature, between the string of FFs and the actual SHA-1
* hash value. The meaning of it is:
*
* 00 -- this marks the end of the FFs; not part of the ASN.1
* bit itself
*
* 30 21 -- a constructed SEQUENCE of length 0x21
* 30 09 -- a constructed sub-SEQUENCE of length 9
* 06 05 -- an object identifier, length 5
* 2B 0E 03 02 1A -- object id { 1 3 14 3 2 26 }
* (the 1,3 comes from 0x2B = 43 = 40*1+3)
* 05 00 -- NULL
* 04 14 -- a primitive OCTET STRING of length 0x14
* [0x14 bytes of hash data follows]
*
* The object id in the middle there is listed as `id-sha1' in
* ftp://ftp.rsasecurity.com/pub/pkcs/pkcs-1/pkcs-1v2-1d2.asn
* (the ASN module for PKCS #1) and its expanded form is as
* follows:
*
* id-sha1 OBJECT IDENTIFIER ::= {
* iso(1) identified-organization(3) oiw(14) secsig(3)
* algorithms(2) 26 }
*/
static const unsigned char sha1_asn1_prefix[] = {
0x00, 0x30, 0x21, 0x30, 0x09, 0x06, 0x05, 0x2B,
0x0E, 0x03, 0x02, 0x1A, 0x05, 0x00, 0x04, 0x14,
};
return PTRLEN_FROM_CONST_BYTES(sha1_asn1_prefix);
}
if (halg == &ssh_sha256) {
/*
* A similar piece of ASN.1 used for signatures using SHA-256,
* in the same format but differing only in various length
* fields and OID.
*/
static const unsigned char sha256_asn1_prefix[] = {
0x00, 0x30, 0x31, 0x30, 0x0d, 0x06, 0x09, 0x60,
0x86, 0x48, 0x01, 0x65, 0x03, 0x04, 0x02, 0x01,
0x05, 0x00, 0x04, 0x20,
};
return PTRLEN_FROM_CONST_BYTES(sha256_asn1_prefix);
}
if (halg == &ssh_sha512) {
/*
* And one more for SHA-512.
*/
static const unsigned char sha512_asn1_prefix[] = {
0x00, 0x30, 0x51, 0x30, 0x0d, 0x06, 0x09, 0x60,
0x86, 0x48, 0x01, 0x65, 0x03, 0x04, 0x02, 0x03,
0x05, 0x00, 0x04, 0x40,
};
return PTRLEN_FROM_CONST_BYTES(sha512_asn1_prefix);
}
unreachable("bad hash algorithm for RSA PKCS#1");
}
static inline size_t rsa_pkcs1_length_of_fixed_parts(const ssh_hashalg *halg)
{
ptrlen asn1_prefix = rsa_pkcs1_prefix_for_hash(halg);
return halg->hlen + asn1_prefix.len + 2;
}
static unsigned char *rsa_pkcs1_signature_string(
size_t nbytes, const ssh_hashalg *halg, ptrlen data)
{
size_t fixed_parts = rsa_pkcs1_length_of_fixed_parts(halg);
assert(nbytes >= fixed_parts);
size_t padding = nbytes - fixed_parts;
ptrlen asn1_prefix = rsa_pkcs1_prefix_for_hash(halg);
unsigned char *bytes = snewn(nbytes, unsigned char);
bytes[0] = 0;
bytes[1] = 1;
memset(bytes + 2, 0xFF, padding);
memcpy(bytes + 2 + padding, asn1_prefix.ptr, asn1_prefix.len);
ssh_hash *h = ssh_hash_new(halg);
put_datapl(h, data);
ssh_hash_final(h, bytes + 2 + padding + asn1_prefix.len);
return bytes;
}
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
static bool rsa2_verify(ssh_key *key, ptrlen sig, ptrlen data)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
BinarySource src[1];
ptrlen type, in_pl;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *in, *out;
/* If we need to support variable flags on verify, this is where they go */
const ssh_hashalg *halg = rsa2_hash_alg_for_flags(0, NULL);
/* Start by making sure the key is even long enough to encode a
* signature. If not, everything fails to verify. */
size_t nbytes = (mp_get_nbits(rsa->modulus) + 7) / 8;
if (nbytes < rsa_pkcs1_length_of_fixed_parts(halg))
return false;
BinarySource_BARE_INIT_PL(src, sig);
type = get_string(src);
/*
* RFC 4253 section 6.6: the signature integer in an ssh-rsa
* signature is 'without lengths or padding'. That is, we _don't_
* expect the usual leading zero byte if the topmost bit of the
* first byte is set. (However, because of the possibility of
* BUG_SSH2_RSA_PADDING at the other end, we tolerate it if it's
* there.) So we can't use get_mp_ssh2, which enforces that
* leading-byte scheme; instead we use get_string and
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
* mp_from_bytes_be, which will tolerate anything.
*/
in_pl = get_string(src);
if (get_err(src) || !ptrlen_eq_string(type, "ssh-rsa"))
Convert a lot of 'int' variables to 'bool'. My normal habit these days, in new code, is to treat int and bool as _almost_ completely separate types. I'm still willing to use C's implicit test for zero on an integer (e.g. 'if (!blob.len)' is fine, no need to spell it out as blob.len != 0), but generally, if a variable is going to be conceptually a boolean, I like to declare it bool and assign to it using 'true' or 'false' rather than 0 or 1. PuTTY is an exception, because it predates the C99 bool, and I've stuck to its existing coding style even when adding new code to it. But it's been annoying me more and more, so now that I've decided C99 bool is an acceptable thing to require from our toolchain in the first place, here's a quite thorough trawl through the source doing 'boolification'. Many variables and function parameters are now typed as bool rather than int; many assignments of 0 or 1 to those variables are now spelled 'true' or 'false'. I managed this thorough conversion with the help of a custom clang plugin that I wrote to trawl the AST and apply heuristics to point out where things might want changing. So I've even managed to do a decent job on parts of the code I haven't looked at in years! To make the plugin's work easier, I pushed platform front ends generally in the direction of using standard 'bool' in preference to platform-specific boolean types like Windows BOOL or GTK's gboolean; I've left the platform booleans in places they _have_ to be for the platform APIs to work right, but variables only used by my own code have been converted wherever I found them. In a few places there are int values that look very like booleans in _most_ of the places they're used, but have a rarely-used third value, or a distinction between different nonzero values that most users don't care about. In these cases, I've _removed_ uses of 'true' and 'false' for the return values, to emphasise that there's something more subtle going on than a simple boolean answer: - the 'multisel' field in dialog.h's list box structure, for which the GTK front end in particular recognises a difference between 1 and 2 but nearly everything else treats as boolean - the 'urgent' parameter to plug_receive, where 1 vs 2 tells you something about the specific location of the urgent pointer, but most clients only care about 0 vs 'something nonzero' - the return value of wc_match, where -1 indicates a syntax error in the wildcard. - the return values from SSH-1 RSA-key loading functions, which use -1 for 'wrong passphrase' and 0 for all other failures (so any caller which already knows it's not loading an _encrypted private_ key can treat them as boolean) - term->esc_query, and the 'query' parameter in toggle_mode in terminal.c, which _usually_ hold 0 for ESC[123h or 1 for ESC[?123h, but can also hold -1 for some other intervening character that we don't support. In a few places there's an integer that I haven't turned into a bool even though it really _can_ only take values 0 or 1 (and, as above, tried to make the call sites consistent in not calling those values true and false), on the grounds that I thought it would make it more confusing to imply that the 0 value was in some sense 'negative' or bad and the 1 positive or good: - the return value of plug_accepting uses the POSIXish convention of 0=success and nonzero=error; I think if I made it bool then I'd also want to reverse its sense, and that's a job for a separate piece of work. - the 'screen' parameter to lineptr() in terminal.c, where 0 and 1 represent the default and alternate screens. There's no obvious reason why one of those should be considered 'true' or 'positive' or 'success' - they're just indices - so I've left it as int. ssh_scp_recv had particularly confusing semantics for its previous int return value: its call sites used '<= 0' to check for error, but it never actually returned a negative number, just 0 or 1. Now the function and its call sites agree that it's a bool. In a couple of places I've renamed variables called 'ret', because I don't like that name any more - it's unclear whether it means the return value (in preparation) for the _containing_ function or the return value received from a subroutine call, and occasionally I've accidentally used the same variable for both and introduced a bug. So where one of those got in my way, I've renamed it to 'toret' or 'retd' (the latter short for 'returned') in line with my usual modern practice, but I haven't done a thorough job of finding all of them. Finally, one amusing side effect of doing this is that I've had to separate quite a few chained assignments. It used to be perfectly fine to write 'a = b = c = TRUE' when a,b,c were int and TRUE was just a the 'true' defined by stdbool.h, that idiom provokes a warning from gcc: 'suggest parentheses around assignment used as truth value'!
2018-11-02 22:23:19 +03:00
return false;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
in = mp_from_bytes_be(in_pl);
out = mp_modpow(in, rsa->exponent, rsa->modulus);
mp_free(in);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
unsigned diff = 0;
unsigned char *bytes = rsa_pkcs1_signature_string(nbytes, halg, data);
for (size_t i = 0; i < nbytes; i++)
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
diff |= bytes[nbytes-1 - i] ^ mp_get_byte(out, i);
smemclr(bytes, nbytes);
sfree(bytes);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(out);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
return diff == 0;
}
static void rsa2_sign(ssh_key *key, ptrlen data,
unsigned flags, BinarySink *bs)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
unsigned char *bytes;
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
size_t nbytes;
mp_int *in, *out;
const ssh_hashalg *halg;
const char *sign_alg_name;
halg = rsa2_hash_alg_for_flags(flags, &sign_alg_name);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
nbytes = (mp_get_nbits(rsa->modulus) + 7) / 8;
bytes = rsa_pkcs1_signature_string(nbytes, halg, data);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
in = mp_from_bytes_be(make_ptrlen(bytes, nbytes));
smemclr(bytes, nbytes);
sfree(bytes);
out = rsa_privkey_op(in, rsa);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(in);
put_stringz(bs, sign_alg_name);
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
nbytes = (mp_get_nbits(out) + 7) / 8;
put_uint32(bs, nbytes);
for (size_t i = 0; i < nbytes; i++)
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
put_byte(bs, mp_get_byte(out, nbytes - 1 - i));
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(out);
}
char *rsa2_invalid(ssh_key *key, unsigned flags)
{
RSAKey *rsa = container_of(key, RSAKey, sshk);
size_t bits = mp_get_nbits(rsa->modulus), nbytes = (bits + 7) / 8;
const char *sign_alg_name;
const ssh_hashalg *halg = rsa2_hash_alg_for_flags(flags, &sign_alg_name);
if (nbytes < rsa_pkcs1_length_of_fixed_parts(halg)) {
return dupprintf(
"%zu-bit RSA key is too short to generate %s signatures",
bits, sign_alg_name);
}
return NULL;
}
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
const ssh_keyalg ssh_rsa = {
rsa2_new_pub,
rsa2_new_priv,
rsa2_new_priv_openssh,
rsa2_freekey,
rsa2_invalid,
rsa2_sign,
rsa2_verify,
rsa2_public_blob,
rsa2_private_blob,
rsa2_openssh_blob,
rsa2_cache_str,
rsa2_pubkey_bits,
"ssh-rsa",
"rsa2",
NULL,
SSH_AGENT_RSA_SHA2_256 | SSH_AGENT_RSA_SHA2_512,
};
RSAKey *ssh_rsakex_newkey(ptrlen data)
{
ssh_key *sshk = rsa2_new_pub(&ssh_rsa, data);
if (!sshk)
return NULL;
return container_of(sshk, RSAKey, sshk);
}
void ssh_rsakex_freekey(RSAKey *key)
{
Invent a struct type for polymorphic SSH key data. During last week's work, I made a mistake in which I got the arguments backwards in one of the key-blob-generating functions - mistakenly swapped the 'void *' key instance with the 'BinarySink *' output destination - and I didn't spot the mistake until run time, because in C you can implicitly convert both to and from void * and so there was no compile-time failure of type checking. Now that I've introduced the FROMFIELD macro that downcasts a pointer to one field of a structure to retrieve a pointer to the whole structure, I think I might start using that more widely to indicate this kind of polymorphic subtyping. So now all the public-key functions in the struct ssh_signkey vtable handle their data instance in the form of a pointer to a subfield of a new zero-sized structure type 'ssh_key', which outside the key implementations indicates 'this is some kind of key instance but it could be of any type'; they downcast that pointer internally using FROMFIELD in place of the previous ordinary C cast, and return one by returning &foo->sshk for whatever foo they've just made up. The sshk member is not at the beginning of the structure, which means all those FROMFIELDs and &key->sshk are actually adding and subtracting an offset. Of course I could have put the member at the start anyway, but I had the idea that it's actually a feature _not_ to have the two types start at the same address, because it means you should notice earlier rather than later if you absentmindedly cast from one to the other directly rather than by the approved method (in particular, if you accidentally assign one through a void * and back without even _noticing_ you perpetrated a cast). In particular, this enforces that you can't sfree() the thing even once without realising you should instead of called the right freekey function. (I found several bugs by this method during initial testing, so I think it's already proved its worth!) While I'm here, I've also renamed the vtable structure ssh_signkey to ssh_keyalg, because it was a confusing name anyway - it describes the _algorithm_ for handling all keys of that type, not a specific key. So ssh_keyalg is the collection of code, and ssh_key is one instance of the data it handles.
2018-05-27 10:32:21 +03:00
rsa2_freekey(&key->sshk);
}
int ssh_rsakex_klen(RSAKey *rsa)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
return mp_get_nbits(rsa->modulus);
}
static void oaep_mask(const ssh_hashalg *h, void *seed, int seedlen,
void *vdata, int datalen)
{
unsigned char *data = (unsigned char *)vdata;
unsigned count = 0;
while (datalen > 0) {
int i, max = (datalen > h->hlen ? h->hlen : datalen);
ssh_hash *s;
unsigned char hash[MAX_HASH_LEN];
assert(h->hlen <= MAX_HASH_LEN);
s = ssh_hash_new(h);
put_data(s, seed, seedlen);
put_uint32(s, count);
ssh_hash_final(s, hash);
count++;
for (i = 0; i < max; i++)
data[i] ^= hash[i];
data += max;
datalen -= max;
}
}
strbuf *ssh_rsakex_encrypt(RSAKey *rsa, const ssh_hashalg *h, ptrlen in)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *b1, *b2;
int k, i;
char *p;
const int HLEN = h->hlen;
/*
* Here we encrypt using RSAES-OAEP. Essentially this means:
*
* - we have a SHA-based `mask generation function' which
* creates a pseudo-random stream of mask data
* deterministically from an input chunk of data.
*
* - we have a random chunk of data called a seed.
*
* - we use the seed to generate a mask which we XOR with our
* plaintext.
*
* - then we use _the masked plaintext_ to generate a mask
* which we XOR with the seed.
*
* - then we concatenate the masked seed and the masked
* plaintext, and RSA-encrypt that lot.
*
* The result is that the data input to the encryption function
* is random-looking and (hopefully) contains no exploitable
* structure such as PKCS1-v1_5 does.
*
* For a precise specification, see RFC 3447, section 7.1.1.
* Some of the variable names below are derived from that, so
* it'd probably help to read it anyway.
*/
/* k denotes the length in octets of the RSA modulus. */
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
k = (7 + mp_get_nbits(rsa->modulus)) / 8;
/* The length of the input data must be at most k - 2hLen - 2. */
assert(in.len > 0 && in.len <= k - 2*HLEN - 2);
/* The length of the output data wants to be precisely k. */
strbuf *toret = strbuf_new_nm();
int outlen = k;
unsigned char *out = strbuf_append(toret, outlen);
/*
* Now perform EME-OAEP encoding. First set up all the unmasked
* output data.
*/
/* Leading byte zero. */
out[0] = 0;
/* At position 1, the seed: HLEN bytes of random data. */
random_read(out + 1, HLEN);
/* At position 1+HLEN, the data block DB, consisting of: */
/* The hash of the label (we only support an empty label here) */
{
ssh_hash *s = ssh_hash_new(h);
ssh_hash_final(s, out + HLEN + 1);
}
/* A bunch of zero octets */
memset(out + 2*HLEN + 1, 0, outlen - (2*HLEN + 1));
/* A single 1 octet, followed by the input message data. */
out[outlen - in.len - 1] = 1;
memcpy(out + outlen - in.len, in.ptr, in.len);
/*
* Now use the seed data to mask the block DB.
*/
oaep_mask(h, out+1, HLEN, out+HLEN+1, outlen-HLEN-1);
/*
* And now use the masked DB to mask the seed itself.
*/
oaep_mask(h, out+HLEN+1, outlen-HLEN-1, out+1, HLEN);
/*
* Now `out' contains precisely the data we want to
* RSA-encrypt.
*/
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
b1 = mp_from_bytes_be(make_ptrlen(out, outlen));
b2 = mp_modpow(b1, rsa->exponent, rsa->modulus);
p = (char *)out;
for (i = outlen; i--;) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
*p++ = mp_get_byte(b2, i);
}
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(b1);
mp_free(b2);
/*
* And we're done.
*/
return toret;
}
mp_int *ssh_rsakex_decrypt(
RSAKey *rsa, const ssh_hashalg *h, ptrlen ciphertext)
{
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_int *b1, *b2;
int outlen, i;
unsigned char *out;
unsigned char labelhash[64];
ssh_hash *hash;
BinarySource src[1];
const int HLEN = h->hlen;
/*
* Decryption side of the RSA key exchange operation.
*/
/* The length of the encrypted data should be exactly the length
* in octets of the RSA modulus.. */
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
outlen = (7 + mp_get_nbits(rsa->modulus)) / 8;
if (ciphertext.len != outlen)
return NULL;
/* Do the RSA decryption, and extract the result into a byte array. */
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
b1 = mp_from_bytes_be(ciphertext);
b2 = rsa_privkey_op(b1, rsa);
out = snewn(outlen, unsigned char);
for (i = 0; i < outlen; i++)
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
out[i] = mp_get_byte(b2, outlen-1-i);
mp_free(b1);
mp_free(b2);
/* Do the OAEP masking operations, in the reverse order from encryption */
oaep_mask(h, out+HLEN+1, outlen-HLEN-1, out+1, HLEN);
oaep_mask(h, out+1, HLEN, out+HLEN+1, outlen-HLEN-1);
/* Check the leading byte is zero. */
if (out[0] != 0) {
sfree(out);
return NULL;
}
/* Check the label hash at position 1+HLEN */
assert(HLEN <= lenof(labelhash));
hash = ssh_hash_new(h);
ssh_hash_final(hash, labelhash);
if (memcmp(out + HLEN + 1, labelhash, HLEN)) {
sfree(out);
return NULL;
}
/* Expect zero bytes followed by a 1 byte */
for (i = 1 + 2 * HLEN; i < outlen; i++) {
if (out[i] == 1) {
i++; /* skip over the 1 byte */
break;
} else if (out[i] != 1) {
sfree(out);
return NULL;
}
}
/* And what's left is the input message data, which should be
* encoded as an ordinary SSH-2 mpint. */
BinarySource_BARE_INIT(src, out + i, outlen - i);
b1 = get_mp_ssh2(src);
sfree(out);
if (get_err(src) || get_avail(src) != 0) {
Complete rewrite of PuTTY's bignum library. The old 'Bignum' data type is gone completely, and so is sshbn.c. In its place is a new thing called 'mp_int', handled by an entirely new library module mpint.c, with API differences both large and small. The main aim of this change is that the new library should be free of timing- and cache-related side channels. I've written the code so that it _should_ - assuming I haven't made any mistakes - do all of its work without either control flow or memory addressing depending on the data words of the input numbers. (Though, being an _arbitrary_ precision library, it does have to at least depend on the sizes of the numbers - but there's a 'formal' size that can vary separately from the actual magnitude of the represented integer, so if you want to keep it secret that your number is actually small, it should work fine to have a very long mp_int and just happen to store 23 in it.) So I've done all my conditionalisation by means of computing both answers and doing bit-masking to swap the right one into place, and all loops over the words of an mp_int go up to the formal size rather than the actual size. I haven't actually tested the constant-time property in any rigorous way yet (I'm still considering the best way to do it). But this code is surely at the very least a big improvement on the old version, even if I later find a few more things to fix. I've also completely rewritten the low-level elliptic curve arithmetic from sshecc.c; the new ecc.c is closer to being an adjunct of mpint.c than it is to the SSH end of the code. The new elliptic curve code keeps all coordinates in Montgomery-multiplication transformed form to speed up all the multiplications mod the same prime, and only converts them back when you ask for the affine coordinates. Also, I adopted extended coordinates for the Edwards curve implementation. sshecc.c has also had a near-total rewrite in the course of switching it over to the new system. While I was there, I've separated ECDSA and EdDSA more completely - they now have separate vtables, instead of a single vtable in which nearly every function had a big if statement in it - and also made the externally exposed types for an ECDSA key and an ECDH context different. A minor new feature: since the new arithmetic code includes a modular square root function, we can now support the compressed point representation for the NIST curves. We seem to have been getting along fine without that so far, but it seemed a shame not to put it in, since it was suddenly easy. In sshrsa.c, one major change is that I've removed the RSA blinding step in rsa_privkey_op, in which we randomise the ciphertext before doing the decryption. The purpose of that was to avoid timing leaks giving away the plaintext - but the new arithmetic code should take that in its stride in the course of also being careful enough to avoid leaking the _private key_, which RSA blinding had no way to do anything about in any case. Apart from those specific points, most of the rest of the changes are more or less mechanical, just changing type names and translating code into the new API.
2018-12-31 16:53:41 +03:00
mp_free(b1);
return NULL;
}
/* Success! */
return b1;
}
static const struct ssh_rsa_kex_extra ssh_rsa_kex_extra_sha1 = { 1024 };
static const struct ssh_rsa_kex_extra ssh_rsa_kex_extra_sha256 = { 2048 };
static const ssh_kex ssh_rsa_kex_sha1 = {
"rsa1024-sha1", NULL, KEXTYPE_RSA,
&ssh_sha1, &ssh_rsa_kex_extra_sha1,
};
static const ssh_kex ssh_rsa_kex_sha256 = {
"rsa2048-sha256", NULL, KEXTYPE_RSA,
&ssh_sha256, &ssh_rsa_kex_extra_sha256,
};
static const ssh_kex *const rsa_kex_list[] = {
&ssh_rsa_kex_sha256,
&ssh_rsa_kex_sha1
};
const ssh_kexes ssh_rsa_kex = { lenof(rsa_kex_list), rsa_kex_list };