This checks if a timeout has been set using ssh_options_set(). If it has
been set it will use that parameter by default for blocking reads.
This is at least what users are expecting.
Fixes T33
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Reviewed-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Set sftp error to SSH_FX_BAD_MESSAGE if an unexpected message is
received.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When an operation fails in sftp subsystem, set the sftp error, so that
it can be obtained by sftp_get_error().
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When an attempt to read a closed channel happens, set the session error
state properly.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When reading a sftp packet and an EOF is received before all requested
bytes are read, set the session and sftp error codes.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* This makes the array constant in the session structure, avoiding
allocations and frees while parsing the file
* It also drops passing the seen array to all the functions,
because it is already part of the passed session
* The test cases are adjusted to match these changes
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
src/libcrypto.c:773:27: warning: comparison of integer expressions of
different signedness: ‘int’ and ‘size_t’ {aka ‘long unsigned int’}
[-Wsign-compare] <--[cc]
if (rc != 1 || outlen != len - aadlen) {
^~
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
clang was reporting a possible memory leak after mkdtemp() call, which
was a false positive, since mkdtemp() returns the same pointer provided
as the parameter, in case of success. This changes the code so that the
static analyser don't get confused.
Found by csbuild runner.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This fixes an access violation when ssh_init() was called after
ssh_finalize() in Windows when using OpenSSL 1.0.2 and libssh statically
linked.
Fixes T120
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This uses a runner configured with the tag "vs2017 or windows".
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
This introduces torture_get_current_working_dir() and
torture_change_dir() to allow changing directories in tests.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Introduced torture_create_temp_file() and torture_close_fp() to allow
creating temprary files transparently in Unix and Windows environment.
This also adds a unit test for the added functions.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Introduces a function to create temporary dir for testing purposes.
Also adds a minimal test for the temporary directory creation.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This completely removes the tests filter code from torture.c and calls
cmocka_set_test_filter() instead, if available. The checks for required
libraries, headers, and the availability of cmocka_set_test_filter()
were added to the cmake configuration.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Define flags S_IRWXO and S_IRWXG in torture_knownhosts_parsing.c when
building for Windows. These flags don't exist in Windows and are not
used.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In Windows environment, there are available low level I/O operations in
io.h, such as _read(), _write(), _open(), and _close().
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Include the header in tests/torture_pki.c only if it is available. It
is not available in Windows environment.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This should not be a problem for well-behaving clients that do not
append the ext-info-c to the rekey, but if they do, we should not
send it either.
Resolves: T121
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The RFC 8308 clearly says, that the additional ext-info-c should
be added only to the first SSH_MSG_KEXINIT.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This was failing when I tried to run the testsuite without
gcrypt devel libraries installed.
Signed-off-by: Jakub Jelen <jakuje@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Valgrind detects many uninitialized memory false positives from
libcrypto's AES-NI internals. Roll back to software AES when running
tests.
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If nread is < 0 and no exception callback is set,
the following code block would cause a buffer overrun.
Signed-off-by: Tilo Eckert <tilo.eckert@flam.de>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If the hostkey type was not in the list of acceptable hostkey
types, the function failed to set the error state. Due to the
fact that the calling function ssh_packet_process() does not
handle the SSH_ERROR return code, the newkeys packet from the
server was silently ignored, stalling the connection until a
timeout occurred.
Signed-off-by: Tilo Eckert <tilo.eckert@flam.de>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
ssh_send_keepalive() should use global_request() to properly configure
the state machine for packet filtering.
Signed-off-by: Nicolas Viennot <nicolas@viennot.biz>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Added missing bn.h include.
Made engine.h include conditional, otherwise it would fail.
DSA_generate_parameters was deprecated long before 1.1.0.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
This commit fixes a couple issues in the CMake configuration files and uses
native features from CMake:
* libssh-build-tree-settings.cmake is deleted. There was a typo that made
this file unusable, anyway.
* use the macros available in CMakePackageConfigHelpers.cmake to generate
the version file and check that the files exist
* Remove the LIBSSH_THREADS_LIBRARY variable, it used the non-existent
LIBSSH_THREADS_LIBRARY_NAME variable.
* Fix the in tree build. libssh can be used uninstalled again.
Test plan:
The values were tested after installing the new files and also without running
'make install'.
Signed-off-by: Christophe Giboudeaux <christophe@krop.fr>
This tests verifies that the only the first seen option is applied
throughout all the configuration files processed. It also verifies
that the configuration files are parsed automatically and that this
behavior can be overridden by configuration option.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This follows the OpenSSH behavior of parsing subseqent configuration
files, while applying only the first option.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This will allow to use the same configuration in client applications
including the users aliases or system wide cryptographic configuration.
As the configuration parsing is the last thing before doing the
actual connection, it might overwrite previously set options.
If this is not intended, the client application can
ask the configuration files to be parsed before setting some other
options that should not ve overwritten. The code ensures that
the configuration is not parsed again.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If either one is missing at this point, fill it with default vaules in
ssh_options_apply().
Previously, when setting up only knownhosts, global_knownhosts file
was left pointing to NULL and the ssh_known_hosts_read_entries()
was trying to open NULL file which is invalid.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Previously, not all of the host keys algorithms were used for algorithm
negotiation. This verifies the algorithms list is sane and ordered
with the key types from known hosts in the first place.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Previously, the file contained the known_hosts strings separated
by NULL bytes which somehow magically worked.
The test was also expecting all the keys from the file will have
the same key type, which was not indeed true.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Previously, only the algorithms we had a keys for in known_hosts
were used, which could lead to no-matching algorithms errors if the
one key we used to depend on was removed from the server.
This commit adds also the other algorithms, but lists them only after
all the key types we have in known_hosts file.
Resolves: T105
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The code as it was written used only the first algorithm from
preferred_hostkeys array and compared it with the list returned
from the known hosts.
This commit is fixing the code so we actually compare each of the
algorithms from both of the lists and returns the intersection.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The conversion from ssh_keytype_e to string does not work for ECDSA keys,
because different key lengths have different string representation.
The usage of type_c should work also for every other key type in future,
but it does not reflrect different signature types (SHA2 extension for RSA
keys), but this early in the key exchange phase, we can not make any
assumptions about supported extensions by the server.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Created the test torture_packet_filter.c which tests if packets are
being correctly filtered.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The packet filter checks required states for the incoming packets and
reject them if they arrived in the wrong state.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When a SSH2_MSG_OPEN_FAILURE arrives, the channel state is checked
to be in SSH_CHANNEL_STATE_OPENING.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When a SSH2_MSG_OPEN_CONFIRMATION arrives, the channel state is checked
to be in SSH_CHANNEL_STATE_OPENING.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
After sending the client token, the auth state is set as
SSH_AUTH_STATE_GSSAPI_MIC_SENT. Then this can be expected to be the
state when a USERAUTH_FAILURE or USERAUTH_SUCCESS arrives.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The introduced auth state allows to identify when a request without
authentication information was sent.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The introduced auth state allows to identify when authentication using
password was tried.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Introduced the states SSH_AUTH_STATE_PUBKEY_OFFER_SENT and
SSH_AUTH_STATE_PUBKEY_AUTH_SENT to know when SSH2_MSG_USERAUTH_PK_OK and
SSH2_MSG_USERAUTH_SUCCESS should be expected.
Fixes T101
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fixes invalid read introduced by commit 21962d.
Accessing tokens[4] for a known_hosts line of
three tokens led to randomly rejected host keys.
This commit completely removes the check because
the optional comments field may contain whitespace.
Signed-off-by: Tilo Eckert <tilo.eckert@flam.de>
The commit also propares the internals throughout the code base
for the inclusion of a new AEAD cipher, because previously, the
source code counted only with chacha20-poly1305 cipher, which
is very specific in many cases.
The SSH_HMAC_AEAD_GCM mac algorithm is not actually used, but the name
needed to be defined so we can match in the algorithms selection per
OpenSSH specification (MACs are ignored in case GCM is select as a cipher [1]).
If the provided OpenSSL does not provide EVP_aes_128_gcm() function,
the AES-GCM ciphers will not be compiled in.
[1] https://cvsweb.openbsd.org/src/usr.bin/ssh/PROTOCOL.chacha20poly1305?annotate=HEAD
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This removes the allocation for the header buffer for each packet we
send.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
The added frees are unnecessary, but the static analyser does not know.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This changes add_cmocka_test() to receive compiler options, the
libraries to be linked to the test, and the linker options. The way the
tests are declared in tests/unittests and tests/client were updated.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure sftp_server_new allocates the packet and payload as
sftp_packet_read now expects the packet and payload to be
pre-allocated.
Similarly, ensure sftp_get_client_message does not free the packet.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The code for calculating SHA 512 in ssh_make_sessionid() had been
duplicated; the cases were unified.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Summary:
buffer variable 's' gets unpacked as char**, the previous code was passing
a char* causing segfaults on all readlink calls inside the unpacking code
Test Plan:
- without patchy examples/samplesftp segfaults in readlink
- with patchy it doesn't
Reviewers: asn
Differential Revision: https://bugs.libssh.org/D14
Signed-off-by: Harald Sitter <sitter@kde.org>
The previous timeout of 500 ms was not enough on slower machines or
while running the tests under valgrind. On much faster machines the
sleep() was bringing unnecessary overhead.
This method opens simple connection to the server verifying it is ready
to accept the connection from the test for 5 seconds. It the server
does not start until then, it fails the tests during initialization,
rather than leaving the cases to run against missing server.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
These frees are unnecessary because the negative tests should not
allocate the keys, but the static analyser reports memory leak errors.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This is no longer needed since libssh can read the private keys
in new OpenSSH format.
This reverts commit 100c9c98ce.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This runs the same test that are ran on the legacy PEM files
also with the new OpenSSH key files.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This implements reading the OpenSSH key format accross the
cryptographic backends. Most of the code is shared and moved
to pki.c, just the building of the keys is implemented in
pki_privkey_build_*() functions.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This extends the torture API to provide a way to request
keys in different formats. This extends the keys with
private keys in the new OpenSSH format (default since
OpenSSH 7.8).
This also needs modifications to the ed25519 tests, which
do not support PEM format and expected the new format out of the
box.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
src/config.c:562:12: error: assuming signed overflow does not occur when
simplifying conditional to constant [-Werror=strict-overflow]
if (args < 1) {
^
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
OpenSSH's sshd does not work well under valgrind so lets avoid tracing it.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Since the verbosity is now set from the setup phase, we do not
need to reset the verbosity, especially not to any arbirary value
such as WARNING.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This allows adjusting the log level of config and options tests using
environment variable LIBSSH_VERBOSITY as it works in most of the other
tests.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Amends f818e63f8, which introduced the constants and matching of this
configuration option, but did not implement the handling of the values
which was causing the configuration parser failing for certain
configurations.
This commit exposes match_pattern_list() from match.c
Red Hat Bugzilla:
https://bugzilla.redhat.com/show_bug.cgi?id=1624425
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The match_hostname() expects comma separated list, while the Host
config keyword in openssh uses spaces separated list by default.
Therefore any subseqent match or negated match in space separated
list will overwrite the previous matches.
This also adjusts the tests to make sure both of the versions work.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This was already done in the torture_threads_pki.
Without the explicit initialization, we can observe random
failures tests (at least of the torture_threads_crypto) from
various threads.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This involves mostly creation of host keys proofs but needs
to follow the same procedure as the client authentication
signatures.
At the same time, the SHA2 extension is enabled in the pkd
so we are able to atomicaly provide correct signatures and
pass tests.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The SHA2 variants should be preferred. Also the buffer needs to be
extended to fit all possible public key algorithms.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This includes intercepting the ext-info-c string from
the client kex proposal, configuring the server to allow using
this extension and sending the SSH_MSG_EXT_INFO packet back
to the client after the new keys are in use.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This does not affect old signatures, where the public key algorithm
matches the public key type.
This is a problem when using SHA2 extension for the RSA keys, where
the new signature algorithsm are introduced in addition to the
exitsing ssh-rsa which was ignored throughout the code.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
SSH_OPTIONS_PUBLICKEY_ACCEPTED_TYPES configuration option can limit
what keys can or can not be used for public key authentication.
This is useful for disabling obsolete algorithms while not completely
removing the support for them or allows to configure what public key
algorithms will be used with the SHA2 RSA extension.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Verify the PUBLICKEY_ACCEPTED_TYPES option is handled correctly
and affects the signature algorithm selection based on the
extensions and can be used to limit list of offered mechanisms
to the server.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This effectively allows to disable using the SHA2 extension, disable
other old public key mechanisms out of the box (hello DSA) or force
the new SHA2-based key algorithm types if needed.
This exposes the default_methods array from kex.c.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This option allows to specify acceptable public key algorithms
and reflects the PubkeyAcceptedTypes configuration option from
OpenSSH.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The new constants for flags are defined in draft-miller-ssh-agent-02
are active if the SHA2 extension is negotiated with the server.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This introduces a new test case for RSA unit tests, verifying that
libraries are able to provide and verify the RSA signatures with
SHA2 hash algorithms.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* This change introduces a new API to request signature using
one key and different hash algorithms. This is used only with
RSA keys, that used to have SHA1 hardcoded, but the new
algorithsms allow to use the SHA2 hashes, if the extension
is negotiated.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This changes the private API by adding one more argument to function
pki_signature_from_blob()
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
RFC 8308: The extension negotiation in Secure Shell (SSH) Protocol
RFC 8332: Use of RSA Keys with SHA-256 and SHA-512
in the Secure Shell (SSH) Protocol
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This is required to work against OpenSSH 7.8, which is now
writing keys in new openssh format by default
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
src/poll.c:1024:9: error: assuming signed overflow does not occur when
simplifying conditional to constant [-Werror=strict-overflow]
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
The target created to copy the file must be the one make dist is
depending on. Otherwise it will not copy the generated files to the
desired path.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Moved the symbols list formatting to the ExtractSymbols.cmake. The
resulting list of symbols is sorted and printed in a more readable way
(one symbol per line). Fixed the script to copy the generated symbols.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
tests/pkd/pkd_hello.c:743:12: error: assuming signed overflow does not
occur when simplifying conditional to constant [-Werror=strict-overflow]
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Fixes, among others, ProxyCommand with %p when no port is used on
commandline or config file, thus using the default port.
Fixes T94
Signed-off-by: Alex Hermann <alex@hexla.nl>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix the targets and output files handling to make the symbols to be
updated correctly when a symbol is added or removed.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This changes amends f561e6bcb3 which
introduces same check in one place, but miss it in other two places.
We encountered this issue with qemu using SFTP to transfer large
data chunks and in some cases, the file transfer was interrupted
without any reason. From the debug messages, it showed up that
last part of data message/packet was not handled in the time
of the sftp_read() call, therefore the ssh_channel_read() returned
zero (there was no more data to read yet), which made the whole
transfer fail hard instead of retrying later.
The proposed change is reusing the code from previously referenced
commit also in the other places.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Updated threading documentation mentioning changes in the requirements
to use libssh in multithread scenarios.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
src/log.c:71:32: error: '%06ld' directive output may be truncated
writing between 6 and 20 bytes into a region of size between 0 and 63
[-Werror=format-truncation=]
snprintf(buf, len, "%s.%06ld", tbuf, (long)tv.tv_usec);
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
This fixes the assignment of variables to comments and makes
the output symmetric with sftp_dequeue().
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If we compile with Visual Studio, we need a DllMain() for running init
and finialize which is the same as a constructor and destructor.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
If the destructor calls finalize, ignore the init counter and finalize
the library anyway.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This changes the condition to use the static error check mutex
initializer. If it is not available, use the default static mutex
initializer.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This adds a cmake module, FindABIMap, which looks for abimap and
provides functions to generate a symbol version linker script.
The module can be included using find_package(ABIMap).
This also adds the option to compile with symbol versioning. The symbol
list is obtained from the header files by filtering those marked with
the LIBSSH_API modifier.
Such symbols are used as input to generate the version script used by
the linker. The version script is automatically updated as new symbols
marked with LIBSSH_API are added to the header files.
If any symbol is removed, the build will fail due to break in the ABI.
Symbol versioning is enabled by default if abimap has been found. It is
disabled in non-UNIX platforms. It can be disabled by passing
"-DWITH_SYMBOL_VERSIONING=OFF" option to cmake.
Pair-Programmed-With: Andreas Schneider <asn@cryptomilk.org>
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
When errors occurred, the session auth state was not being updated,
leading to failures due to the wrong state in following authentication
methods.
Fixes T56
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This can help when debugging in the future
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Since the call is made automatically when the library is loaded, these
calls are no longer required (if the library is not linked statically).
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When pthreads is available, run the threads tests on Windows
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Run the tests from torture_pki_rsa.c on threads. Tests requiring files
to be removed are not tested, since they would require the access to
the files to be synchronized.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The test run crypto test on multiple threads.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The test run buffer tests on multiple threads.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The test runs ssh_init()/ssh_finalize() on multiple threads.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Added Windows threads implementation based on CriticalSection.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This makes unnecessary to call ssh_init() when the library is
dynamically loaded. Also removes the threads shared library. The used
threads implementation is chosen in configuration time, changing the
ssh_threads_get_default() depending on the available threads library.
Internally, it is expected a threads implementation providing:
- void ssh_mutex_lock(void **mutex);
- void ssh_mutex_unlock(void **mutex);
- struct ssh_threads_callbacks_struct *ssh_threads_get_default(void);
and a crypto implementation providing:
- int crypto_thread_init(struct ssh_threads_callbacks_struct *user_callbacks);
- void crypto_thread_finalize(void);
This adds internal threads implementation for pthreads and noop.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This adds support for the ed25519 keys in the pkd framework and adds
openssh-only tests utilizing these host keys (dropbear does not support
them yet).
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Previously, the support was only partial and if the ed25519 key was
the only one, the internal checks were failing the tests.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add a small helper for ssh_buffer to ensure that the buffer has a
certain amount of space already preallocated. This can be useful in case
it is known how much data is going to be added to a buffer, to avoid
multiple reallocations.
Make use of it in few places in the library.
Signed-off-by: Pino Toscano <ptoscano@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This verifies that all the supported host keys can be used and
verified by the client, including the SHA2 extension in RFC 8332.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
torture_connect_double test case used to test the connect only up
to key exchange phase, but not after the new keys are stated to be
used for communication. The keys from previous connectoin were not
cleaned up from the previous invocation as well as the seqence
number was not reset and therefore any further packet was failing
with length-check errors or MAC errors.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This tests multiple items, (1) cross compilation, (2) compilation
on debian and (3) CI run on MIPS systems, tested under qemu. This target
was originally written for GnuTLS by Michael Weiser.
Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Fix the build for mbedTLS:
* set HAVE_CHACHA for non-mbedTLS builds
* only compile chachapoly.c when HAVE_CHACHA
* use empty CHACHA20 in src/kex.c unless HAVE_CHACHA
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In 06b9901e64, invocations of `set_decrypt_key`
and `set_encrypt_key` were moved into the `ssh_packet_newkeys` callback, away
from the packet decrypt and encrypt functions.
Remove the extra `set_decrypt_key` for the case that an `aead_decrypt_length`
is not NULL. At this time, only the chacha20-poly1305@openssh.com cipher
is affected by this change.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
With this change there is less code specific to the
chacha20-poly1305 cipher found in src/wrapper.c.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The local `rc` variable here is never set. Fix a warning that is
emitted due to `-Wunused-variable`.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Make the `test_data` larger so that tests do not read beyond
its length. Observed in testing with an `-fsanitize=address`
build locally.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Avoid setting keys on every packet decrypt or encrypt operation.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
That code is really ugly, but it wasn't meant to be modular at all in the
first place.
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The `-v` is only recognized by `dbclient` when dropbear is built
in its DEBUG_TRACE mode. Omit that flag by default to avoid a
warning log emitted to stderr.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Support for the `blowfish-cbc` cipher has been removed from OpenSSH
as of version 7.6. Remove this cipher from the pkd tests so that
the tests will pass together with a modern OpenSSH client.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Emit a friendly error message for OpenSSH clients older than
7.0. Some of the recent pkd changes now require a modern
client to support some newer config options.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
As of OpenSSH 6.9, support for `ssh-dss` user keys is disabled by default
at runtime. Specify an explicit `-o PubkeyAcceptedKeyTYpes` in the pkd
tests to explicitly enable each user key type being tested, including
`ssh-dss`.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
As of OpenSSH 6.9, support for `ssh-dss` host keys is disabled by default
at runtime. Specify an explicit `-o HostKeyAlgorithms` in the pkd tests
to explicitly enable each host key type being tested, including `ssh-dss`.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Use the socket_wrapper preload shim when running the `pkd_hello`
test with `make test`. The end goal here is to get this test
running alongside normal tests in regular CI. Changes to do
this:
* Configure PKD_ENVIRONMENT for the `pkd_hello_i1` test in the
CMakeLists.txt file.
* Add a `--socket-wrapper-dir|-w` flag that is used to opt-in to
initializing a SOCKET_WRAPPER_DIR as expected by the socket_wrapper
library.
A runtime flag is used here to make it easy to run `pkd_hello`
with the socket_wrapper library while avoiding a hard dependency.
Testing done: observed socker_wrapper in effect with `strace`;
running `make test` uses the wrapper correctly on my local
machine.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add an entry for a `pkd_hello_i1` test which runs one iteration
through each of the pkd algorithm combinations.
Testing done: now `make test` will run `pkd_hello -i1` which
completes in ~25 seconds on my local machine.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to include config.h so that the `HAVE_DSA` value is properly set
when building the pkd tests.
Introduced with 778652460f,
Testing done: with this change, the `pkd_hello` test is passing on an
OpenSSL 1.1.0 build for me. Previously it would fail pubkey exchange
early on for DSA- and ECDSA-type host keys.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix two memory leaks in `ssh_get_pubkey_hash` for some error paths.
The local `h` buffer and `ctx` MD5 context each must be free'd for
the SSH_ERROR cases.
Introduced with 16217454d5.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix `ssh_get_pubkey_hash` indentation to use softabs
with 4 spaces. No change in behavior.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Wrap some function definitions with `HAVE_LIBCRYPTO` ifdefs to
match their usages in `torture_run_tests`.
Fixes this warning I observe when building locally:
torture_pki_ecdsa.c:341:13: warning:
‘torture_pki_ecdsa_write_privkey’ defined but not used
[-Wunused-function]
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to provide the `ssh_string` pubkey blob to the buffer packing
routine when computing the SSH_MSG_KEXDH_REPLY message, rather than
the new `ssh_key` type.
Introduced with 16217454d5.
Testing done: with this change, the `pkd_hello` test is passing on a
mbedTLS build for me. Previously it would segfault during pubkey
exchange with "ecdh-sha2-nistp256".
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to provide the `ssh_string` pubkey blob to the buffer packing
routine when computing the SSH_MSG_KEXDH_REPLY message, rather than
the new `ssh_key` type.
Introduced with 16217454d5.
Testing done: with this change, the `pkd_hello` test is passing on a
libgcrypt build for me. Previously it would segfault during pubkey
exchange with "ecdh-sha2-nistp256".
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to provide the `ssh_string` pubkey blob to the buffer packing
routine when computing the SSH_MSG_KEXDH_REPLY message, rather than
the new `ssh_key` type.
Introduced with 16217454d5.
Testing done: with this change, `pkd_hello` test is passing on an
OpenSSL 1.1.0 build for me. Previously it would segfault during
pubkey exchange with "ecdh-sha2-nistp256".
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When building in a cross-compiling environment, use wine to run the
tests.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This is necessary when building the tests for Windows.
Signed-off-by: Anderson Toshiyuki Sasaki <ansasaki@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Thanks to Tilo Eckart.
The global variable "ge25519_base" is referenced in the module
"src/external/ed25519.c" and initialized in "src/external/ge25519.c".
The lack of the extern keyword in the header results in different
instances being compiled into both translation units with some
compilers.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
This is a regression introduced by 16217454d5
We import the pubkey into current_crypto which is NULL instead of
next_crypto.
Fixes T90
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
This allows compiling and testing with undefined sanitizer.
Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This allows compiling and testing with undefined sanitizer.
Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This allows operating under environments where the username variables
are not present.
Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add a new option SSH_OPTIONS_NODELAY to enable or disable the
Nagle Algorithm (TCP_NODELAY) on the session socket.
Improved performance can be achieved for some applications like
sftp servers by enabling SSH_OPTIONS_NODELAY as typically, the
next request won't arrive until the server replies, which are
typically small writes.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Useful on Windows platforms where SecureZeroMemory is available.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix OSX build error about embedding a directive within macro arguments.
Apparently, snprintf is implemented as a macro on that platform.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
ssh_connector_free fails to remove the in/out channel callbacks
as ssh_connector_remove_event sets the in/out channel variables
to NULL.
Have ssh_connector_free, remove the channel callbacks first before
invoking ssh_connector_remove_event.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
POLLHUP needs to be checked on in_fd, which may be a pipe.
A pipe in Linux signals EOF through POLLHUP (see:
http://www.greenend.org.uk/rjk/tech/poll.html)
Without checking POLLHUP, a client could spin up indefinetely
doing ssh_event_dopoll.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The stray semi-colon in sftp.h is flagged when using -pedantic
which affects clients that include the header and use -pedantic
and -Werror on their codebase.
Signed-off-by: Alberto Aguirre <albaguirre@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
torture_rand and torture_server_x11 call ssh_init without checking
the return value. If mbedTLS is built without threading support
ssh_init fails but the tests continue and then segfault since threading
wasn't correctly initialised.
Add a section that documents requirements for mbedTLS usage in a
multi threaded environment to README.mbedtls.
Signed-off-by: Juraj Vijtiuk <juraj.vijtiuk@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
We need to specify a cipher when we generate a key with a password.
OpenSSH uses aes_128_cbc, so we should use the same.
Thanks to Julian Lunz for the report.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Summary:
When ed25519 was introduced in commit 93c7b81b4e,
the ed25519 case was added after the ecdsa case in src/pki.c. The ecdsa case seems to
have relied on falling through to report an error, when HAVE_ECC is not defined.
If HAVE_ECC is not defined, but ecdsa keys are used, with for example,
ssh_pki_import_pubkey_file, the code fallthroughs into the ed25519 case.
Signed-off-by: Juraj Vijtiuk <juraj.vijtiuk@sartura.hr>
Test Plan: Unit tests passed. No memory leaks found with valgrind.
Reviewers: asn
Differential Revision: https://bugs.libssh.org/D13
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Summary:
This patch adds support for mbedTLS as a crypto backend for libssh.
mbedTLS is an SSL/TLS library that has been designed to mainly be used
in embedded systems. It is loosely coupled and has a low memory
footprint. mbedTLS also provides a cryptography library (libmbedcrypto)
that can be used without the TLS modules.
The patch is unfortunately quite big, since several new files had to
be added.
DSA is disabled at compile time, since mbedTLS doesn't support DSA
Patch review and feedback would be appreciated, and if any issues or
suggestions appear, I'm willing to work on them.
Signed-off-by: Juraj Vijtiuk <juraj.vijtiuk@sartura.hr>
Test Plan:
* The patch has been tested with a Debug and MinSizeRel build, with
libssh unit tests, client tests and the pkd tests.
* All the tests have been run with valgrind's memcheck, drd and helgrind
tools.
* The examples/samplessh client works when built with the patch.
Reviewers: asn, aris
Subscribers: simonsj
Differential Revision: https://bugs.libssh.org/D1
pcap file is generated by the processes writing to the sockets,
which is not allowed for privilege-separated process in new
OpenSSH servers (confined by seccomp filter).
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Summary:
Fix a typo for old libgcrypt builds `k_length` -> `k_len`.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Test Plan: * Spotted this last-minute typo bug in local testing.
Reviewers: asn
Reviewed By: asn
Differential Revision: https://bugs.libssh.org/D10
Summary:
Fix a typo for old libgcrypt builds `k_length` -> `k_len`.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Test Plan: * Spotted this last-minute typo bug in local testing.
Reviewers: asn
Differential Revision: https://bugs.libssh.org/D10
Expose an API 'ssh_server_init_kex' which allows one to change the set of
key exchange, hostkey, ciphers, MACs, and compression algorithms currently
configured for the ssh_session at hand, after having started the
'ssh_handle_key_exchange' process.
One can use this API from the already-existing 'connect_status_function'
callback to dynamically modify the set of algorithms used after having
received the client banner, but before sending out the initial KEXINIT
message.
For example, one might want to prevent advertising the curve25519 key
exchange algorithm for older OpenSSH clients due to interop bugs.
Fixes T25
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
There's a race window between the accept loop's call to
accept(2) and it checking `ctx.keep_going`. Forcefully
close the server socket such that any raced `accept` ends
up failing.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Sometimes, but not always, the pkd tests will fail because they
close the socket at hand a bit too early for the client. The
client in turn may exit non-zero when that happens.
Split up the final close loop so that pkd waits to receive a
channel close from the client, and then socket close, before
finally returning.
With this change I observe that tests are now passing in
environments that would previously tickle the above race
and fail.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a memory leak in the path where parsing returns early due
to seeing a repeated opcode. A testcase is added which
demonstrates the leak and fix with valgrind.
Resolves CID 1374267.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Summary:
Based on Dirkjan's original patch series here:
* https://www.libssh.org/archive/libssh/2015-08/0000029.html
Here the changes are adapted for the current master
branch, and expanded to include libgcrypt support.
Co-Authored-By: Dirkjan Bussink <d.bussink@gmail.com>
Signed-off-by: Jon Simons <jon@jonsimons.org>
Test Plan:
* Ran pkd tests for libcrypto and libgcrypt builds.
* Ran client torture_algorithms.c tests for libcrypto and libgcrypt builds.
* Tested across multiple libgcrypts ("1.6.3" and "1.7.6-beta").
Reviewers: aris, asn
Reviewed By: asn
Tags: #libssh
Differential Revision: https://bugs.libssh.org/D7
Summary:
Based on Dirkjan's original patch series here:
* https://www.libssh.org/archive/libssh/2015-08/0000029.html
Here the changes are adapted for the current master
branch, and expanded to include libgcrypt support.
Co-Authored-By: Dirkjan Bussink <d.bussink@gmail.com>
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Test Plan:
* Ran pkd tests for libcrypto and libgcrypt builds.
* Ran client torture_algorithms.c tests for libcrypto and libgcrypt builds.
* Tested across multiple libgcrypts ("1.6.3" and "1.7.6-beta").
Reviewers: aris, asn
Tags: #libssh
Differential Revision: https://bugs.libssh.org/D7
That way, we will not fail later on key exchange phase when something
unknown is negotiated.
Fixes T37
Signed-off-by: Nikos Mavrogiannopoulos <nmav@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Summary:
Hello, this is a resend for a quick memory leak fix for one of the unit
tests, originally sent to the mailing list here:
* https://www.libssh.org/archive/libssh/2017-07/0000017.html
Test Plan:
* Before the fix and running the test with valgrind:
```
[simonsj@simonsj-lx5 : unittests] valgrind --leak-check=full ./torture_options >/dev/null
==93134== Memcheck, a memory error detector
==93134== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==93134== Using Valgrind-3.10.0 and LibVEX; rerun with -h for copyright info
==93134== Command: ./torture_options
==93134==
[ PASSED ] 10 test(s).
[ PASSED ] 1 test(s).
==93134==
==93134== HEAP SUMMARY:
==93134== in use at exit: 80 bytes in 1 blocks
==93134== total heap usage: 977 allocs, 976 frees, 75,029 bytes allocated
==93134==
==93134== 80 bytes in 1 blocks are definitely lost in loss record 1 of 1
==93134== at 0x4C28C20: malloc (vg_replace_malloc.c:296)
==93134== by 0x41BAB0: ssh_key_new (pki.c:107)
==93134== by 0x40DF90: torture_bind_options_import_key (torture_options.c:222)
==93134== by 0x4E3AA3A: cmocka_run_one_test_or_fixture (cmocka.c:2304)
==93134== by 0x4E3ACEA: cmocka_run_one_tests (cmocka.c:2412)
==93134== by 0x4E3B036: _cmocka_run_group_tests (cmocka.c:2517)
==93134== by 0x40E9E3: torture_run_tests (torture_options.c:276)
==93134== by 0x40DE68: main (torture.c:1100)
==93134==
==93134== LEAK SUMMARY:
==93134== definitely lost: 80 bytes in 1 blocks
==93134== indirectly lost: 0 bytes in 0 blocks
==93134== possibly lost: 0 bytes in 0 blocks
==93134== still reachable: 0 bytes in 0 blocks
==93134== suppressed: 0 bytes in 0 blocks
==93134==
==93134== For counts of detected and suppressed errors, rerun with: -v
==93134== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
```
* And after:
```
[simonsj@simonsj-lx5 : unittests] valgrind --leak-check=full ./torture_options >/dev/null
==93294== Memcheck, a memory error detector
==93294== Copyright (C) 2002-2013, and GNU GPL'd, by Julian Seward et al.
==93294== Using Valgrind-3.10.0 and LibVEX; rerun with -h for copyright info
==93294== Command: ./torture_options
==93294==
[ PASSED ] 10 test(s).
[ PASSED ] 1 test(s).
==93294==
==93294== HEAP SUMMARY:
==93294== in use at exit: 0 bytes in 0 blocks
==93294== total heap usage: 977 allocs, 977 frees, 75,029 bytes allocated
==93294==
==93294== All heap blocks were freed -- no leaks are possible
==93294==
==93294== For counts of detected and suppressed errors, rerun with: -v
==93294== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
```
Reviewers: asn
Reviewed By: asn
Differential Revision: https://bugs.libssh.org/D3
Summary:
Hello, resending this patch series for the `pkd` tests, originally
sent to the mailing list here:
* https://www.libssh.org/archive/libssh/2017-07/0000011.html
Here are a few improvements and fixups for the `pkd` tests, including
a new flag `-m` that can be used to run only certain subsets of the
test passes.
Jon Simons (5):
pkd: rename AES192 cipher suite -> OPENSSHONLY
pkd_daemon.c: mark `pkd_ready` field as volatile
pkd: fixups for updated CMocka CMUnitTest struct
pkd: refactor -t testname lookup-by-name
pkd: support -m to match multiple tests
tests/pkd/pkd_daemon.c | 2 +-
tests/pkd/pkd_daemon.h | 1 +
tests/pkd/pkd_hello.c | 84 +++++++++++++++++++++++++++++++++-----------------
3 files changed, 58 insertions(+), 29 deletions(-)
--
Test Plan:
* I've been using the new `-m` mode locally for a long time to run
only certain groups of tests.
* The CMocka struct fixes can be seen in the pkd output before and
after: after, there are no more extraneous test output strings.
* The fix for the `pkd_ready` field can be observed when building
the libssh tests with `-Os` on a Debian system (before the fix,
pkd would hang, after the fix, it runs as intended).
Reviewers: asn
Reviewed By: asn
Tags: #libssh
Differential Revision: https://bugs.libssh.org/D2
On OpenSSL versions prior to 1.1.0, `EVP_CIPHER_CTX_cleanup` will
dereference its argument regardless of whether it is NULL. This
is not a problem on OpenSSL at or beyond 1.1.0, where
`EVP_CIPHER_CTX_cleanup` (macro to `EVP_CIPHER_CTX_reset`) returns
early upon NULL input.
Move the call to `EVP_CIPHER_CTX_cleanup` under the existing NULL
check in `evp_cipher_cleanup` to avoid the problem.
Introduced with this build-break fix:
* e66f370682
Found in manual testing in an environment with an older OpenSSL.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Relax the cases where `ssh_analyze_banner` fails to extract a
major and minor version from banners which appear like OpenSSH
banners.
Update the tests to demonstrate that now a banner as might be
sent by `ssh-keyscan(1)` ("SSH-2.0-OpenSSH-keyscan") no longer
returns failure.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
With this change, a HAVE_LIBCRYPTO #ifdef is removed from wrapper.c.
Now, the libcrypto-specific logic for EVP_CIPHER_CTX_free is moved
into the ssh_cipher_struct cleanup callback handler for those
ciphers.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a gcrypt build error introduced with
48e7b098f8.
The ssh_cipher_struct only contains a `ctx` field on
the libcrypto builds, so it can't be referenced unless
within HAVE_LIBCRYPTO.
This build fix preserves the original spirit of the
change in 48e7b098f8:
only call `EVP_CIPHER_CTX_free` when `cipher->ctx`
is non-NULL.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix error-checking for `strtoul` in `ssh_analyze_banner`, and
enable some tests which demonstrate the fix before-and-after.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
No changes to code, only whitespace indentation and
an update to the function docs.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a resource leak in `hmac_final`: say `HMAC_CTX_free` instead
of `HMAC_CTX_reset`. This matches the error handling as done in
`hmac_init`. Introduced with cf1e808e2f.
The problem is reproducible running the `pkd_hello` test with:
valgrind --leak-check=full ./pkd_hello -i1 -t torture_pkd_openssh_dsa_rsa_default
Resolves https://red.libssh.org/issues/252.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
On older OpenSSL versions, the EVP_MD_CTX fields within an HMAC_CTX
structure are contained inlined (change here [1]): be sure to not
try to free those fields on those builds.
Found running the `pkd_hello` test with:
valgrind ./pkd_hello -i1 -t torture_pkd_openssh_dsa_rsa_default
^ valgrind will cite "Invalid free() ..." errors which are present
before this fix and absent after, when building with OpenSSL 1.0.1.
[1] 6e59a892db
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
libssh fails to read the configuration from a config file due to a
wrong check in 'ssh_config_parse_line' procedure in 'config.c'; it's
effectively skipping every opcode (and therefore every option) from
the file. The change fixes that behaviour.
Signed-off-by: Artyom V. Poptsov <poptsov.artyom@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Makes sure we free pending keyboard auth prompts
so prompts that have not be replied to do not leak.
Signed-off-by: Peter Volpe <pvolpe@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Check if the size the other party sent is a valid size in the
transmitted buffer.
Thanks to Alex Gaynor for finding and reporting the issue.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Check if the size the other party sent is a valid size in the
transmitted buffer.
Thanks to Alex Gaynor for finding and reporting the issue.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Tokens are not allowed (according to the manpage).
Expansion was introduced by a wrong fix for #127.
This commit reverts part of 6eea08a9ef
Signed-off-by: Alex Hermann <alex@hexla.nl>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This sets the bind private key directly from an ssh_key struct instead
of reading a file.
Signed-off-by: Alfredo Mazzinghi <am2419@cl.cam.ac.uk>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/pki.c (ssh_signature_free): Fix test for ECC using gcrypt.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This has been made opaque and it needs to be a pointer.
This is for OpenSSL 1.1.0 support.
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* include/libssh/crypto.h (struct ssh_crypto_struct): Provide a
suitable 'ecdh_privkey'.
* include/libssh/ecdh.h: Also define 'HAVE_ECDH' if we do ECC using
libgcrypt.
(ecdh_build_k): New prototype.
* src/CMakeLists.txt (libssh_SRCS): Add backend-specific files.
* src/ecdh.c: Move backend-specific parts to...
* src/ecdh_crypto.c: ... this file.
* src/ecdh_gcrypt.c: New file.
* src/wrapper.c (crypto_free): Free 'ecdh_privkey'.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* include/libssh/libgcrypt.h (EVPCTX): Fix type.
(NID_gcrypt_nistp{256,384,521}): New constants.
* src/libgcrypt.c (nid_to_md_algo): New function mapping curves to
digest algorithms.
(evp{,_init,_update,_final}): New functions.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/curve25519.c (ssh_client_curve25519_init): Make use of the
gcrypt-variant of 'bignum_bin2bn'.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When parsing the result of a successful fstat call, make sure to free
the resulting reply message.
Signed-off-by: Pino Toscano <ptoscano@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Comparison ((flags & O_RDONLY) == O_RDONLY) is always true.
Also, O_RDWR, O_WRONLY and O_RDONLY are mutually exclusive => no need to check all of them
Signed-off-by: Tilo Eckert <tilo.eckert@flam.de>
This saves a round-trip with SSHv2 connecting to the server. See RFC
4253 section 5.2 (New Client, Old Server).
Thanks to Yang Yubo <yang@yangyubo.com> for the suggestion.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
According to RFC 4253 the max banner length is 255.
Thanks to Saju Panikulam <spanikulam@ipswitch.com> for the report.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Comply with RFC 4253 compliance section 4.2.
Allow data other than "SSH-" to be sent across prior to the actual
version striong.
Signed-off-by: Ken Reister <reister.kenneth@CIMCOR.COM>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In Cockpit we've seen this memory leak:
at 0x4C2A9C7: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x5B76B03: ssh_userauth_agent (auth.c:778)
by 0x40DD5A: cockpit_ssh_authenticate (cockpitsshtransport.c:327)
BUG: https://red.libssh.org/issues/208
Signed-off-by: Stef Walter <stefw@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/pki_gcrypt.c (pki_export_pubkey_rsa1): Rework to be more
idiomatic. Fix leaking MPIs.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/pki_gcrypt.c (pki_publickey_to_blob): Rework using the new
helper 'ssh_sexp_extract_mpi'.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/pki_gcrypt.c (pki_key_dup): Rework to be more idiomatic.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/options.c (ssh_options_getopt): Only call 'ssh_options_set' if
a port has been given.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* tests/torture.c (torture_setup_create_sshd_config): Rework how the
location of the sftp server is discovered, and add the Debian-specific
location.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
* src/pki_gcrypt.c (asn1_get): Generalize from 'asn1_get_int'.
(asn1_get_int): New function simply calling 'asn1_get'.
Signed-off-by: Justus Winter <justus@g10code.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
- enabled TrustedUserCAKeys option in torture.c
- adds a new set of (signed) keys for bob in a separate dir
The private key used to generate the certs is included, but not required.
Signed-off-by: Axel Eppe <aeppe@google.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Due to a byte/bit confusion, the DH secret was too short. This file was
completely reworked and will be commited in a future version.
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Documentation now is congruent with the code:
- SSH_OK is returned on success,
- SSH_ERROR is returned on error.
Signed-off-by: Younes Serraj <younes.serraj@gmail.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
ssh-agent needs to be executed as the local user and not a fake user or
we will not be able to add identies.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Note that removing ssh_buffer_get_begin() doesn't break API
compatibility, as this functions has never been exposed (it only
has the LIBSSH_API prefix).
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This commit is a preparatory stage for removing ssh_buffer_get_begin().
Note that removing ssh_buffer_get_begin() doesn't break API
compatibility, as this functions has never been exposed (it only has the
LIBSSH_API prefix).
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
As ssh_buffer_get_len() actually calls ssh_buffer_get_rest_len(), let's
just use the first one. This is a preparatory step for removing
ssh_buffer_get_rest_len().
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This is a preparatory step for having the behavior of
ssh_buffer_get_rest_len() in the ssh_buffer_get_len() and then remove
the ssh_buffer_rest_len()
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
/home/ffidenci/src/upstream/libssh/tests/benchmarks/bench_scp.c: In
function ‘benchmarks_scp_down’:
/home/ffidenci/src/upstream/libssh/tests/benchmarks/bench_scp.c:112:14:
warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has
type ‘size_t {aka long unsigned int}’ [-Wformat=]
printf("Only %d bytes available (on %lu requested).\n",size,bytes);
^
/home/ffidenci/src/upstream/libssh/tests/benchmarks/bench_scp.c:116:14:
warning: format ‘%d’ expects argument of type ‘int’, but argument 2 has
type ‘size_t {aka long unsigned int}’ [-Wformat=]
printf("File is %d bytes (on %lu requested). Will cut the end\n"
,size,bytes);
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
/home/ffidenci/src/upstream/libssh/src/pki_gcrypt.c: In function
‘pki_key_compare’:
/home/ffidenci/src/upstream/libssh/src/pki_gcrypt.c:1082:5: warning:
enumeration value ‘SSH_KEYTYPE_DSS_CERT01’ not handled in switch
[-Wswitch]
switch (k1->type) {
^
/home/ffidenci/src/upstream/libssh/src/pki_gcrypt.c:1082:5: warning:
enumeration value ‘SSH_KEYTYPE_RSA_CERT01’ not handled in switch
[-Wswitch]
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Having "ssh_" prefix in the functions' name will avoid possible clashes
when compiling libssh statically.
Signed-off-by: Fabiano Fidêncio <fidencio@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
- ssh_pki_import_cert_base64()
- ssh_pki_import_cert_file()
- ssh_pki_import_cert_blob()
Those functions are currently simple wrappers around their pubkey counterpart.
- ssh_pki_copy_cert_to_privkey()
This function copies the cert-specific data to a private key.
Signed-off-by: Axel Eppe <aeppe@google.com>
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Implicitly fixed unsafe return code handling that could result in use-after-free.
Signed-off-by: Tilo Eckert <tilo.eckert@flam.de>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Allow callers to specify their own socket
for an ssh agent.
Signed-off-by: Peter Volpe <pvolpe@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The previous list is not freed. Since the new session just got
created, an identity list is already allocated and empty.
Signed-off-by: Sebastien Boving <seb@google.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Also consider the host key type at hand when computing whether a
'first_kex_packet_follows' packet matches the current server settings.
Without this change libssh may incorrectly believe that guessed
settings which match by kex algorithm alone fully match: the host
key types must also match. Observed when testing with dropbear
clients.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Return SSH_OK for the case that an incoming SSH_MSG_KEXDH_INIT should be
ignored. That is, for the case that the initial 'first_kex_packet_follows'
guess is incorrect. Before this change sessions served with libssh can be
observed to error out unexpectedly early when testing with dropbear clients
that send an incompatible guess.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
CID: #1250106
This fixes a 1 byte output overflow for large key length (not reachable
in libssh). Pulled from OpenBSD BCrypt PBKDF implementation.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
This is an addition to CVE-2015-3146 to fix the null pointer
dereference. The patch is not required to fix the CVE but prevents
issues in future.
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The state validation in the packet handlers for SSH_MSG_NEWKEYS and
SSH_MSG_KEXDH_REPLY had a bug which did not raise an error.
The issue has been found and reported by Mariusz Ziule.
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
I've improved FindOpenSSL and FindZLIB in that version to work well with
Windows and Linux. This was 2011 it should be old enough that most
distributions have at least this version available.
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Option can be used to filter out irrelevant tests
usage: ./torture_pki '*ed25519'
Signed-off-by: Aris Adamantiadis <aris@0xbadc0de.be>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
torture_server_x11 tests whether a libssh server can correctly parse an
X11 request.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add the ability to generate a keypair (RSA, but tests that focus on
what occurs over a connection may well not care) and an ssh_bind
easily.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
BUG: https://red.libssh.org/issues/168
On some platforms, like OpenBSD, stdin and friends are macros which break
compilation of this struct. I've fixed these names such that it compiles and
runs cleanly on OpenBSD
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If realloc of sftp->ext->name or sftp->ext->data fails, the memory
previously allocated for the respective member is leaked. Fix this by
storing the return value of realloc() in a temporary variable which only
gets assigned to the respective sftp->ext member on success.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix the typo 'stengths' -> 'strengths' and remove the trailing white
space on the same line.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Before this change, dangling pointers can be unintentionally left in the
respective next_crypto kex methods slots. Ensure to set all slots to
NULL in the error-out path.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix DSA signature extraction for the LIBGCRYPT build. Here, the same fix
that was applied to the LIBCRYPTO build for https://red.libssh.org/issues/144
is now adapted for pki_gcrypt. Additionally, ensure to set the resulting
output sig_blob buffer before returning.
Before this fix, one can observe the failure with the pkd test on a LIBGCRYPT
build as so:
# ./pkd_hello -i 1 -t torture_pkd_openssh_dsa_dsa_default
After, runs of 10000 back-to-back iterations of the same test are passing.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Disable the 'make_string_bn_inplace' helper function for the LIBGCRYPT
build, rather than using '#error' to fail the build completely.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Per ecdsa(3ssl), ECDSA_SIG_new does allocate its 'r' and 's' bignum fields.
Fix a bug where the initial 'r' and 's' bignums were being overwritten with
newly-allocated bignums, resulting in a memory leak.
BUG: https://red.libssh.org/issues/175
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a build breakage when '-DWITH_SERVER=OFF' is set: skip building the
pkd test for that case. Add some missing includes for the OpenIndiana
and FreeBSD builds.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Introduce a sample public-key testing daemon to the 'pkd' test directory,
and add support code for cycling through various combinations of different
key exchange, cipher, and MAC algorithms.
The goal of the 'pkd_hello' test is to make it easy to test interactions
between non-libssh clients and a libssh-server, and to provide a starting
point for testing new implementations for key types, ciphers, MACs, and
so on. The thinking is that testing a new algorithm should be as simple
as adding a new line for it in the PKDTESTS_* lists.
Macros are used to generate the tests and helper functions for a couple of
clients -- here, OpenSSH and dropbear are included for the first cut. If
binaries are found for these clients, their test lists will be enabled;
when binaries are not found for a given client, those tests are skipped.
Tests are run in one large batch by default, but can also be run individually
to help with tracking down things like signature bugs that may take many
iterations to reproduce.
Each test logs its stdout and stderr to its own file, which is cleaned up
when a test succeeds. For failures, those logs can be combined with verbose
libssh output from pkd itself to start debugging things.
Some example usages:
pkd_hello
Run all tests with default number of iterations.
pkd_hello --list
List available individual test names.
pkd_hello -i 1000 -t torture_pkd_openssh_ecdsa_256_ecdh_sha2_nistp256
Run only the torture_pkd_openssh_ecdsa_256_ecdh_sha2_nistp256
testcase 1000 times.
pkd_hello -v -v -v -v -e -o
Run all tests with maximum libssh and pkd logging.
Included in the tests are passes for all existing kex, cipher, and MAC
algorithms.
BUG: https://red.libssh.org/issues/144
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add missing initialization of 'type_c' field of a SSH signature in
'pki_do_sign_sessionid' procedure.
If libssh is compiled with GCrypt, 'dh_handshake_server' fails with
"Could not sign the session id" error. The change fixes that.
Signed-off-by: Artyom V. Poptsov <poptsov.artyom@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a missing 'buffer_pack' formatter in 'ssh_message_auth_interactive_request'.
With this fix the 'examples/samplesshd-kbdint' program is working again for me.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Check the 'ssh_buffer_pack' return in ssh_pki_do_sign for the ED25519 case.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fail fast in 'pki_import_pubkey_buffer' for the ED25519 case if a buffer
sized ED25519_PK_LEN can not be retrieved. Before, the 'memcpy' could
have read beyond the bounds of 'ssh_string_data(pubkey)'.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to free the z_stream structures as allocated from
the gzip.c initcompress, initdecompress functions.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
ssh-keygen makes the tests very slow because new keys are generated
at the start of every test.
ssh-keygen on OSX doesn't support ecdsa keys, even if libcrypto does.
Reviewed-by: Andreas Schneider <asn@samba.org>
That function permits chaining of buffer values to minimize buffer handling
in packet sending code.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Add NULL checks to 'is_first_kex_packet_follows_guess_wrong'
to ensure that a 'strdup(NULL)' path can not be taken.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to honor the 'first_kex_packet_follow' field when processing
KEXINIT messages in the 'ssh_packet_kexinit' callback. Until now
libssh would assume that this field is always unset (zero). But
some clients may set this (dropbear at or beyond version 2013.57),
and it needs to be included when computing the session ID.
Also include logic for handling wrongly-guessed key exchange algorithms.
Save whether a client's guess is wrong in a new field in the session
struct: when set, the next KEX_DHINIT message to be processed will be
ignored per RFC 4253, 7.1.
While here, update both 'ssh_packet_kexinit' and 'make_sessionid' to
use softabs with a 4 space indent level throughout, and also convert
various error-checking to store intermediate values into an explicit
'rc'.
Patch adjusted from original to ensure that client tests remain passing
(ie 'torture_connect'): restrict the changes in 'ssh_packet_kexinit'
only for the 'server_kex' case.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to honor the 'first_kex_packet_follow' field when processing
KEXINIT messages in the 'ssh_packet_kexinit' callback. Until now
libssh would assume that this field is always unset (zero). But
some clients may set this (dropbear at or beyond version 2013.57),
and it needs to be included when computing the session ID.
Also include logic for handling wrongly-guessed key exchange algorithms.
Save whether a client's guess is wrong in a new field in the session
struct: when set, the next KEX_DHINIT message to be processed will be
ignored per RFC 4253, 7.1.
While here, update both 'ssh_packet_kexinit' and 'make_sessionid' to
use softabs with a 4 space indent level throughout, and also convert
various error-checking to store intermediate values into an explicit
'rc'.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Enable retrieving the "ecdsa-sha2-nistpNNN" name of ECDSA keys with a
new 'ssh_pki_key_ecdsa_name' API. This gives more information than the
'ssh_key_type_to_char' API, which yields "ssh-ecdsa" for ECDSA keys.
The motivation is that this info is useful to have in a server context.
The torture_pki unit test is updated to include the new API, and a few
more passes are added to additionally test 384 and 521-bit keys.
Signed-off-by: Jon Simons <jon@jonsimons.org>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Tested by
ssh_server_fork -k <an ecdsa key> -k <an rsa key> ...
and connection succeeded with OpenSSH clients both for ECDSA and RSA
(the latter forced via -o HostKeyAlgorithms=ssh-rsa and some known
hosts clearing was necessary). Also
ssh_server_fork -k <an ecdsa key> -k <another ecdsa key> ...
caused the second key to be used, as expected.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
SSH_BIND_OPTIONS_HOSTKEY will now load host keys of any supported type
rather than set the algorithms that the server permits (which seems
like an unhelpful option anyway; it seems you can always control this
by just loading the right keys).
This option has slightly different semantics than the
SSH_BIND_OPTIONS_<x>KEY options because it requires the key file to
exist immediately rather than on ssh_bind_listen or
ssh_bind_accept_fd. The semantics of this option makes more sense to
me.
We also eliminate ssh_bind_options_set_algo, since it is no longer
used.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In packet_send2, rather than issue two separate buffer_prepend_data calls
(each of which may entail realloc + memmove + memcpy), elide the prepend
work into a single buffer_prepend_data: the header information is computed
locally, and a single 5 byte prepend operation is now done instead of
prepending 1, then 4 bytes.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
ssh_channel_read is a wrapper for ssh_channel_read_timeout with timeout
-1 (infinite) so we call that directly.
Signed-off-by: Petar Koretic <petar.koretic@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Defining a non inlined class function in a header will cause multiple
definitions when header is included in more that one file since for each
file function will get defined.
Signed-off-by: Petar Koretic <petar.koretic@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If the sockets are created manually and passed to libssh the internal session
state is set to SSH_SESSION_STATE_SOCKET_CONNECTED. Result of this fix can be
verified by running torture_connect test (torture_connect_socket) with -vvvv
flags.
Signed-off-by: Luka Perkov <luka.perkov@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
On Mar 16, 09:41, Aris Adamantiadis wrote:
> Hi Petar,
> I agree with the principle, but I don't think this code can work...
> RSA_generate_key takes an RSA* as parameter and in our code we probably
> have key->rsa==NULL. (if we don't then the old code had a memory leak).
>
> Does the test case work ?
>
> Aris
>
Yes, you are right. This works, tested with tests/unittests/torture_pki
Signed-off-by: Petar Koretic <petar.koretic@sartura.hr>
The libssh library by default does not allow in-source build (with cmake
MacroEnsureOutOfSourceBuild macro). The INSTALL file (implicitly) suggests
creating a build directory. So lets add build to list of git ignore files to
avoid complaints from git.
Signed-off-by: Luka Perkov <luka.perkov@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
It seems useful to be able to run ssh_server_fork without being
required to load some key of RSA and DSA keytypes. However, with the
current ssh_server_fork, you need to have some keys as some default
value is set by default and you can't unset the value for a keytype
(e.g. by using NULL as an argument). So the "no default keys"
argument turns off the default key assignments.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The commit fixes this build warning:
====
src/server.c:223:8: warning: ‘privkey’ may be used uninitialized in this function [-Wmaybe-uninitialized]
rc = ssh_pki_export_privkey_to_pubkey(*privkey, &pubkey);
^
src/server.c:243:11: note: ‘privkey’ was declared here
ssh_key privkey;
====
Signed-off-by: Luka Perkov <luka.perkov@sartura.hr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In 'ssh_bind_listen', move setting of 'sshbind->bindfd' to only happen after
the listen call: otherwise 'bindfd' can be set to a bogus descriptor for the
case that listen fails.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When accepting a new connection, a forking server based on libssh forks
and the child process handles the request. The RAND_bytes() function of
openssl doesn't reset its state after the fork, but simply adds the
current process id (getpid) to the PRNG state, which is not guaranteed
to be unique.
This can cause several children to end up with same PRNG state which is
a security issue.
After discussion with Aris and it was not obvious enough to understand
the issue we decided to refactor it.
Reviewd-by: Aris Adamantiadis <aris@0xbadc0de.be>
Right now the behavior of packet_{en,de}crypt on len == 0 depends on
the behavior of malloc. Instead, make these consistently fail based
on what I assume the desired behavior is due to the first error
message in each.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
If we receive a packet of length exactly blocksize, then
packet_decrypt gets called on a buffer of size 0. The check at the
beginning of packet_decrypt indicates that the function should be
called on buffers of at least one blocksize, though the check allows
through zero length. As is packet_decrypt can return -1 when len is 0
because malloc can return NULL in this case: according to the ISO C
standard, malloc is free to return NULL or a pointer that can be freed
when size == 0, and uclibc by default will return NULL here (in
"non-glibc-compatible" mode). The net result is that when using
uclibc connections with libssh can anomalously fail.
Alternatively, packet_decrypt (and probably packet_encrypt for
consistency) could be made to always succeed on len == 0 without
depending on the behavior of malloc.
Thanks to Josh Berlin for bringing conneciton failures with uclibc to
my attention.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Follow-up to 4e04ec8, which caused a regression on OS X.
Checking the value of CMAKE_THREAD_LIBS_INIT to decide whether any threading
library is present on a system turns out to be wrong -- in OS X, for
example, usage of pthreads does not depend on any additional linker or
compiler flags, so CMAKE_THREAD_LIBS_INIT is empty and our check in
src/CMakeLists.txt failed (it used to work before 4e04ec8 because
CMAKE_HAVE_THREADS_LIBRARY is set).
Instead, just look for Threads_FOUND, which FindThreads sets just like any
other Find module when it has found what it was looking for.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix bug #142
The mode does need to be an octal numeric string. Mode 0600 now gets sent on the wire as 0384, triggering a "scp: protocol error: bad mode" response, and an "scp status code 1d not valid" message from libssh.
As mentioned in the previous commit, there are cases where
CMAKE_HAVE_THREADS_LIBRARY is not set and pthreads _is_ being used: one can
pass -DTHREADS_HAVE_PTHREAD_ARG=1 to CMake directly so that it just passes
-pthread to the compiler/linker and does not set CMAKE_HAVE_THREADS_LIBRARY.
Since we are only interested in knowing whether any threading library has
been found, we should use CMAKE_THREAD_LIBS_INIT instead (Threads_FOUND
would also work).
Note that, at the moment, there is only a pthreads backend available in
threads/, so if it is not found configuration will fail because CMake will
try to create a library from an empty set of source files.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
libssh is primarily interested in whether pthreads is present and can be
used. Checking for CMAKE_HAVE_THREADS_LIBRARY is not the same thing, as
there are cases where pthread exists but CMAKE_HAVE_THREADS_LIBRARY is not
set (for example, FreeBSD passes -DTHREADS_HAVE_PTHREAD_ARG=1 to CMake by
default as a way to skip the checks for -lpthread, -lpthreads and others and
tell the build system that -pthread is the one expected to be used).
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to check whether the socket at hand is indeed still connected
throughout POLLIN processing in ssh_socket_pollcallback.
Before this change, the POLLIN block in ssh_socket_pollcallback is
predicated against the condition (s->state == SSH_SOCKET_CONNECTED).
Once entered, data from the socket is consumed through the data
callback in this loop:
do {
r = s->callbacks->data(buffer_get_rest(s->in_buffer),
buffer_get_rest_len(s->in_buffer),
s->callbacks->userdata);
buffer_pass_bytes(s->in_buffer,r);
} while (r > 0);
However, it is possible for the socket data callback to change the
state of the socket (closing it, for example). Fix the loop to only
continue so long as the socket remains connected: this also entails
setting the ssh_socket state to SSH_SOCKET_CLOSED upon close.
The bug can be observed before the change by sending a bogus banner
to the server: 'echo -e "A\r\nB\r\n" | nc localhost 22'. Each of
'A' and 'B' will be processed by 'callback_receive_banner', even
though the client socket is closed after rejection of 'A'.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
So that it would match ssh_channel_pty_request_callback as well as the documentation
Signed-off-by: Audrius Butkevicius <audrius.butkevicius@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix the DSA portion of 'pki_signature_to_blob': before this change, it
is possible to sometimes observe DSA signature validation failure when
testing with OpenSSH clients. The problem ended up being the following
snippet which did not account for the case when 'ssh_string_len(x)' may
be less than 20:
r = make_bignum_string(sig->dsa_sig->r);
...
memcpy(buffer,
((char *) ssh_string_data(r)) + ssh_string_len(r) - 20,
20);
Above consider the case that ssh_string_len(r) is 19; in that case the
memcpy unintentionally starts in the wrong place. The same situation
can happen for value 's' in this code.
To fix, adjust the offsets used for the input and output pointers, taking
into account that the lengths of 'r' and 's' can be less than 20. With
the fix I am no longer able to reproduce the original failure mode.
BUG: https://red.libssh.org/issues/144
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
ssh_get_error can actually work on anything with an ssh_common_struct
as its first member. It is already used in examples in the
distribution with ssh_sessions and ssh_binds.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Pad RSA signature blobs to the expected RSA signature length
when processing via 'pki_signature_to_blob'.
Some clients, notably PuTTY, may send unpadded RSA signatures
during the public key exchange: before this change, one can
sometimes observe failure in signature validation when using
PuTTY's 'plink' client, along these lines:
ssh_packet_process: ssh_packet_process: Dispatching handler for packet type 50
ssh_packet_userauth_request: ssh_packet_userauth_request: Auth request for service ssh-connection, method publickey for user 'foo'
ssh_pki_signature_verify_blob: ssh_pki_signature_verify_blob: Going to verify a ssh-rsa type signature
pki_signature_verify: pki_signature_verify: RSA error: error:04091077:rsa routines:INT_RSA_VERIFY:wrong signature length
ssh_packet_userauth_request: ssh_packet_userauth_request: Received an invalid signature from peer
For cross-reference this issue once also existed between
PuTTY and OpenSSH:
http://www.chiark.greenend.org.uk/~sgtatham/putty/wishlist/rsa-verify-failed.htmlhttp://www.openbsd.org/cgi-bin/cvsweb/src/usr.bin/ssh/ssh-rsa.c?rev=1.19;content-type=text%2Fx-cvsweb-markup
With the fix I am unable to reproduce the above failure mode when
testing with 'plink'.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Make sure to explicitly set key pointers to NULL following the use
of 'ssh_key_free' throughout bind.c.
Before this change, a double free can happen via 'ssh_bind_free'
as in this example callpath:
// create an ssh_bind
ssh_bind b = ssh_bind_new();
// provide a path to a wrong key-type
ssh_bind_options_set(b, SSH_BIND_OPTIONS_DSAKEY, path_to_rsa_key);
// initialize set key-type
ssh_bind_listen(b);
-> error path "The DSA host key has the wrong type: %d",
ssh_key_free(sshbind->dsa)
-> ssh_key_clean(key) // OK
-> SAFE_FREE(key) // OK, but, sshbind->dsa is *not* set to NULL
// ssh_bind_listen failed, so clean up ssh_bind
ssh_bind_free(b);
-> ssh_key_free(sshbind->dsa) // double-free here
To fix, set pointers to NULL that have been free'd with 'ssh_key_free'.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
This works same way as ssh_forward_accept() but can return a destination
port of the channel (useful if SSH connection forwarding several TCP/IP
ports).
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Not 100% satisfied of this patch, but the way libgcrypt handles
threading in 1.6 is not compatible with custom handlers. The
new code basicaly uses pthreads in every case. This will probably
not work on windows.
The ssh_userauth_none() call should already be non-blocking. However
this this function is broken in non-blocking mode. It should reveal the
existing bug.
Fix the setting of 'channel->flags' to use '|='. Before this
change, one bug symptom can be that channels are never fully
free'd via ssh_channel_free, resulting in memory leaks.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Ensure to use 'ssh_channel_do_free' in 'ssh_disconnect', when removing and
free'ing up a session's channels. This matches the behavior in 'ssh_free',
and is necessary to fully free any channel which may not have been closed
completely (see usage of flags SSH_CHANNEL_FLAG_CLOSED_REMOTE,
SSH_CHANNEL_FLAG_FREED_LOCAL).
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
In a previous refactoring patch, the code underpinning the
ssh_gssapi_set_creds() API was inadvertently removed. This patch
fixes the problem.
Also clarify what variable holds which credentials and insure that
credentials created within the library are propelry freed.
Signed-off-by: Simo Sorce <simo@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
- Properly acquire and inquitre credentials to get the list of available
credentials.
- Avoid enforcing a specific username it breaks some use cases (k5login).
- Remove confusing references to delegated credentials as there is no code
that actually uses delegated credentials in the initialization case.
Signed-off-by: Siom Sorce <simo@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
For integration with an external mainloop, we need to know how to
replicate libssh's internal poll() calls. We originally through
ssh_get_status() was that API, but it's not really - those flags only
get updated from the *result* of a poll(), where what we really need
is to know how libssh would *start* a poll().
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Otherwise applications providing their own fd end up tripping an
assertion, since the session is just in _CONNECTING.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
When the public key auth handler is executed and returns SSH_OK,
ssh_execute_server_callbacks() still runs some client callbacks,
which may set rc to SSH_AGAIN, which triggers a default reply on
auth, denying auth.
Signed-off-by: Nicolas Viennot <nicolas@viennot.biz>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
The code is careful to reenable compression when rekeying.
Signed-off-by: Nicolas Viennot <nicolas@viennot.biz>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Per RFC 4252, it is required to send back only one of either
SSH_MSG_USERAUTH_PK_OK or SSH_MSG_USERAUTH_FAILURE for public
key probes.
Update the handling of 'auth_pubkey_function' to send back PK_OK
instead of SSH_MSG_USERAUTH_SUCCESS for the case that the state
of the message at hand is SSH_PUBLICKEY_STATE_NONE.
With this change, it is now possible to process an initial key probe
and then subsequent signature validation using the server callbacks.
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Fix a memory leak in 'ssh_poll_ctx_free': issue 'ssh_poll_free'
to remove the poll handle from its context and free it.
BUG: https://red.libssh.org/issues/128
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
A few callback descriptions refer to a non-existent value SSH_AUTH_OK,
which should be SSH_AUTH_SUCCESS. This commit fixes these.
Signed-off-by: Alan Dunn <amdunn@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
# If the source code was changed, but there were no interface changes:
# If the source code was changed, but there were no interface changes:
# Increment REVISION.
# Increment REVISION.
set(LIBRARY_VERSION"4.3.0")
set(LIBRARY_VERSION"4.7.2")
set(LIBRARY_SOVERSION"4")
set(LIBRARY_SOVERSION"4")
# where to look first for cmake modules, before ${CMAKE_ROOT}/Modules/ is checked
# where to look first for cmake modules, before ${CMAKE_ROOT}/Modules/ is checked
set(CMAKE_MODULE_PATH
${CMAKE_SOURCE_DIR}/cmake/Modules
)
# add definitions
# add definitions
include(DefineCMakeDefaults)
include(DefinePlatformDefaults)
include(DefinePlatformDefaults)
include(DefineCompilerFlags)
include(DefineInstallationPaths)
include(DefineInstallationPaths)
include(DefineOptions.cmake)
include(DefineOptions.cmake)
include(CPackConfig.cmake)
include(CPackConfig.cmake)
include(CompilerChecks.cmake)
# disallow in-source build
# disallow in-source build
include(MacroEnsureOutOfSourceBuild)
include(MacroEnsureOutOfSourceBuild)
macro_ensure_out_of_source_build("${PROJECT_NAME} requires an out of source build. Please create a separate build directory and run 'cmake /path/to/${PROJECT_NAME} [options]' there.")
macro_ensure_out_of_source_build("${PROJECT_NAME} requires an out of source build. Please create a separate build directory and run 'cmake /path/to/${PROJECT_NAME} [options]' there.")
@@ -456,5 +456,14 @@ SUCH HOLDER OR OTHER PARTY HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
DAMAGES.
DAMAGES.
Linking with OpenSSL
Linking with OpenSSL
17. In addition, as a special exception, we give permission to link the code of its release of libssh with the OpenSSL project's "OpenSSL" library (or with modified versions of it that use the same license as the "OpenSSL" library), and distribute the linked executables. You must obey the GNU Lesser General Public License in all respects for all of the code used other than "OpenSSL". If you modify this file, you may extend this exception to your version of the file, but you are not obligated to do so. If you do not wish to do so, delete this exception statement from your version.
17. In addition, as a special exception, we give permission to link the code
of its release of libssh with the OpenSSL project's "OpenSSL" library (or with
modified versions of it that use the same license as the "OpenSSL" library),
and distribute the linked executables. You must obey the GNU Lesser General
Public License in all respects for all of the code used other than "OpenSSL".
If you modify this file, you may extend this exception to your version of the
file, but you are not obligated to do so. If you do not wish to do so, delete
style="color:#000000;clip-rule:nonzero;display:inline;overflow:visible;visibility:visible;opacity:1;isolation:auto;mix-blend-mode:normal;color-interpolation:sRGB;color-interpolation-filters:linearRGB;solid-color:#000000;solid-opacity:1;fill:#4d4d4d;fill-opacity:1;fill-rule:nonzero;stroke:none;stroke-width:0.26458332;stroke-linecap:butt;stroke-linejoin:miter;stroke-miterlimit:4;stroke-dasharray:none;stroke-dashoffset:0;stroke-opacity:1;paint-order:stroke fill markers;color-rendering:auto;image-rendering:auto;shape-rendering:auto;text-rendering:auto;enable-background:accumulate"
d="m 0.52916667,292.2374 -0.26458334,0.52925 v 3.43958 H 4.7625001 v -3.43958 H 2.38125 L 2.1166667,292.2374 Z"
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
