Skip to content

Expose the configured IdentityAgent and IdentitiesOnly options - #15

Open
alexey-milovidov wants to merge 795 commits into
masterfrom
continue-pr-115072-identity-agent-option
Open

Expose the configured IdentityAgent and IdentitiesOnly options#15
alexey-milovidov wants to merge 795 commits into
masterfrom
continue-pr-115072-identity-agent-option

Conversation

@alexey-milovidov

@alexey-milovidov alexey-milovidov commented Aug 19, 2026 •
edited
Loading

Copy link
Copy Markdown
Member

Expose SSH_OPTIONS_IDENTITY_AGENT and SSH_OPTIONS_IDENTITIES_ONLY through ssh_options_get, so that consumers can use the values resolved by libssh's full SSH configuration parser instead of parsing ssh_config themselves. IdentitiesOnly is a boolean, and it is reported the same way it is spelled in the configuration file: yes or no.

Used by ClickHouse/ClickHouse#115072, which looks up the SSH key of clickhouse-client the same way ssh does it.

Jakuje and others added 30 commits August 12, 2025 12:13
@Jakuje
poll: Fix memory leak on failed realloc()
5a6e2fd 
In cases where this is the initial allocation, the shrinking of the polltrs
buffer would result in 0B realloc, which really does not make sense. Also,
when this second realloc fails, the memory is never freed as the outer code
believes there is nothing allocated on the poll_ctx

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
poll: Initialize ssh_poll_handle pointers
5691e0f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
socket: Free poll handle when resetting socket state
118a747 
Since 07cb0be we are not closing the user provided FDs,
but the above change also resulted in memory leak during
ssh_disconnect that left the poll_handle allocated during
reset.

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
tests: Enable all key exchange methods in ssh_ping
771e19a 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2025-8277: packet: Adjust packet filter to work when DH-GEX is gu…
4310a69 
…essed wrongly

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@eferollo @Jakuje
CVE-2025-8277: Fix memory leak of unused ephemeral key pair after cli…
ccff22d 
…ent's wrong KEX guess

Signed-off-by: Francesco Rollo <eferollo@gmail.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2025-8277: ecdh: Free previously allocated pubkeys
c9d95ab 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
tests: Invoke all combinations of wrong guesses during rekey
d357a9f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2025-8277: wrapper: Free cv25519 private key on cleanup
9ada7aa 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2025-8277: mbedtls: Avoid leaking ecdh keys
ffed80f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@cryptomilk @Jakuje
CVE-2025-8114: Fix NULL pointer dereference after allocation failure
53ac23d 
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@catenacyber @Jakuje
socket: do not free poll object if it is locked
c992614 
As it may a cause a use after free if `send` fails when
ssh_poll_ctx_dopoll does its callback
ssh_poll_ctx_dopoll still wants to use the poll object later

Signed-off-by: Philippe Antoine <p.antoine@catenacyber.fr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
poll: Use is_locked helper where possible
df4e907 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@DarkPhoenix42 @Jakuje
refactor(pki): Define RSA_MIN_KEY_SIZE and update related checks
e8bbd19 
Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@Jakuje
Add non-namespaced alias sntrup761x25519-sha512 that is being standar…
6ca5930 
…dized

The specification is now in the last call, data point is allocated so there is
no need to stick to the namespaces alias anymore

https://datatracker.ietf.org/doc/draft-ietf-sshm-ntruprime-ssh/

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Sahana Prasad <sahana@redhat.com>
@Jakuje
tests: Avoid prefix matching when selecting algorithmms
3d3b128 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Sahana Prasad <sahana@redhat.com>
@eferollo @Jakuje
fix(bind): Remove code duplication in ssh_bind_listen
c94e2ef 
Signed-off-by: Francesco Rollo <eferollo@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@catenacyber @Jakuje
fuzz: test allocations failures
59a502e 
Signed-off-by: Philippe Antoine <p.antoine@catenacyber.fr>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@Jakuje
tests: Rewrite all fuzzers to LLVMFuzzerInitialize and nalloc
a3c5d3b 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@DarkPhoenix42 @Jakuje
feat(buffer): add ssh_buffer_dup function to duplicate existing buffers
2f77727 
Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@DarkPhoenix42 @Jakuje
test(buffer): add unit tests for ssh_buffer_dup function
8541b65 
Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@DarkPhoenix42 @Jakuje
pki: define RSA_DEFAULT_KEY_SIZE
8c4e337 
Define a new constant for the default RSA key size for consistency.

Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@DarkPhoenix42 @Jakuje
pki: update RSA key generation to use default size when parameter is 0
08cbbea 
Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@DarkPhoenix42 @Jakuje
refactor(pki): separate the sk signature buffer packing to a separate…
55bb909 
… function

The logic for creating the buffer to be verified from an sk signature from ssh_pki_signature_verify has been separated into a new function named pki_prepare_sk_signature_buffer to allow for convenient reuse of this logic.

Signed-off-by: Praneeth Sarode <praneethsarode@gmail.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@Jakuje
mbedtls: Warn about missing featues of mbedTLS build
607dad0 
The libssh requires mbedTLS to have support for threading. Given the
way how the mbedTLS builds are configured (at least to my limited
understanding), by modifying mbedtls_config.h header file, this
changes the silent failure to a message on stderr pointing the
user in the right direction.

Fixes: #304

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
examples: Add more flexibility to set verbosity
c551402 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
sftpserver: Remove some needless parts
a51384f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
sftpserver: Fix loop termination
6aea779 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
tests: Fix wording in comment to make sense
ecb11f1 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
connector: Reformat
8a0aa17 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
Jakuje and others added 25 commits February 5, 2026 15:43
@Jakuje
options: Allow listing all identity files
39d931f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
tests: Check SSH_OPTIONS_NEXT_IDENTITY functionality
8a134e0 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@galorithm @Jakuje
sftp_aio: Test sftp aio with libssh proxyjump
29dd787 
Signed-off-by: Eshan Kelkar <eshankelkar@galorithm.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@galorithm @Jakuje
connector: Fix sftp aio read/write with ProxyJump
dc39902 
Addresses issue #319

The commit description explains:
1. Fix for sftp aio + read
2. Fix for sftp aio + write

1. Fix for sftp aio + read
-------------------------
The reproducer provided in the issue description had a model
as follows (with one jump host):
fd_1---(socket_pair)---fd_2---(connector)----channel(fd_3)-----server

Via debugging, it was noticed that the channel connected directly to
the server stored a lot of unbuffered data (received from the server)
that wasn't being written to fd_2 via the connector API.

(Here on, channel refers to the channel(fd_3) in the diagram connected
directly to the server)

Consider the situation, where after a bit of progress in the transfer,
the server has sent all the requested data (requested via outstanding
requests) and all of that data is stored in channel->stdout_buffer. Say
this data is 10,000 bytes.

At this point, all the client (fd_1) is doing is waiting for all
outstanding requests. (and processing thei responses)

- POLLOUT event callback gets generated indicating that fd_2 is
  available for writing.

- ssh_connector_fd_out_cb() gets called to handle the POLLOUT.

- Assuming connector->in_available was true, 4096 (CHUNKSIZE) bytes
  get read from the channel. (really channel->stdout_buffer) leaving
  10,000 - 4096 = 5904 bytes unread in the channel.

- The read bytes are sent via fd_2 (so that fd_1 can recv them)

- After this, the callback sets connector->in_available to 0 and
  connector->out_wontblock to 0.

- Since out_wontblock has been set to 0 ssh_connector_reset_pollevents()
  (called after the callback returns) will consider POLLOUT events on the
  connector output.

- (Based on assumption before) Since the client (fd_1) is eagerly
  awaiting responses and processing them, the received data gets
  processed quickly and fd_2 is available for sending/writing.

- POLLOUT event gets generated for fd_2 indicating that its available
  for writing/sending to fd_1

- ssh_connector_fd_out_cb() gets called to handle the POLLOUT

- Since connector->in_available is 0 (and
  ssh_connector_channel_data_cb() has not been trigerred in between
  as we have assumed before that all the data has already been received on the
  channel and is stored in the channel->stdout_buffer), ssh_connector_fd_out_cb()
  does nothing besides setting connector->out_wontblock to 1.

- Since out_wontblock has been set to 1 ssh_connector_reset_pollevents()
  (called after the callback returns) will IGNORE POLLOUT events on the
  connector output.

- So, at this point, the channel->buffer contains 5706 bytes and the
  fd_2 is available for writing/sending (out_wontblock is 1), but
  nothing happens and the transfer gets stalled/hanged.

In my opinion, this hanging occurs because connector->in_available was
incorrectly set to 0 despite the channel buffer having 5706 bytes in it.

This commit changes that code to consider the data available to read
on the channel (includes buffered data as well as polled data on
channel's internal fd) and taking that into consideration to set
in_available appropriately. (Instead of unconditionally setting it to 0 as the
current code does) so that the next time POLLOUT gets received on fd_2
the ssh_connector_fd_out_cb() does read from the channel and write to
fd_2 (as the connector->in_available flag would be set).

2. Fix for sftp aio + write
-------------------------------------
On writing tests for sftp aio + proxyjump, it was encountered
that file uploads were also hanging. Though I was not able to
pin point the exact cause for this hanging, the nature of hanging
was observed to be as follows:

- sftp aio write + proxyjump blocks/hangs occasionally (not always)

- It hangs at different points in the test

- hang point 1: Sometimes it hangs after sending the first write request
  (i.e. the second write request call hangs and never returns, at this point
  we are not even waiting for response, just sending data). A lot of pending
  data to write to socket/fd was noticed at this hang point.

- hang point 2: Sometimes it hangs while waiting for the second write request
  response.

- It hangs at ssh_handle_packets_termination (i.e. this is the
  call that never returns), in context to hang point 1, this occurs due to
  trying to flush the channel during sftp_packet_write, and in context to
  hang point 2, this occurs due to trying to read an sftp response packet.

- Not sure why, but more the verbose logging/printing I do, the lesser
  occasionally test hangs (e.g. 1 test in 6-7 test runs), maybe this could
  be a hint for a race condition / thread interaction related bug, but am
  not sure.

Fix: On modifying the connector code to mark out_wontblock
to 0 in case of output channel only when the channel's
remote window is 0, the hanging no longer occured.

Though, as mentioned before, I don't know the exact problem
(i.e. case causing hanging) the fix addresses, but the fix
is logical (if remote window is +ve data can still be written
to channel and hence out_wontblock should not be reset to 0, it should
be set to 1) and fixes the issue hence is added to this commit.

Signed-off-by: Eshan Kelkar <eshankelkar@galorithm.com>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
sftp: Reformat sftp_extensions_* API
f0fdfd4 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
sftp: Fix out-of-bound read from sftp extensions
855a085 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
Reproducer for out of bounds read of SFTP extensions
b90b7f2 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
ed25519: Avoid timing leak when comparing private keys
5654c59 
This affects libgcrypt and mbedTLS backends. The OpenSSL backend is
using OpenSSL implementation of the Ed25519 which is compared correctly.

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
pki: Discourage comparing private keys
b3e13b7 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
sftp: Initialize pointers
28d6d10 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2026-0968: sftp: Sanitize input handling in sftp_parse_longname()
20856f4 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2026-0968 tests: Reproducer for invalid longname data
90a5d8f 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2026-0967 match: Avoid recursive matching (ReDoS)
a411de5 
The specially crafted patterns (from configuration files) could cause
exhaustive search or timeouts.

Previous attempts to fix this by limiting recursion to depth 16 avoided
stack overflow, but not timeouts. This is due to the backtracking,
which caused the exponential time complexity O(N^16) of existing algorithm.

This is code comes from the same function from OpenSSH, where this code
originates from, which is not having this issue (due to not limiting the number
of recursion), but will also easily exhaust stack due to unbound recursion:

openssh/openssh-portable@05bcd0c

This is an attempt to simplify the algorithm by preventing the backtracking
to previous wildcard, which should keep the same behavior for existing inputs
while reducing the complexity to linear O(N*M).

This fixes the long-term issue we had with fuzzing as well as recently reported
security issue by Kang Yang.

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
CVE-2026-0966 misc: Avoid heap buffer underflow in ssh_get_hexa
417a095 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
CVE-2026-0966 tests: Test coverage for ssh_get_hexa
9be8358 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
CVE-2026-0966 doc: Update guided tour to use SHA256 fingerprints
1b2a4f7 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
CVE-2026-0965 config: Do not attempt to read non-regular and too larg…
a5eb30d 
…e configuration files

Changes also the reading of known_hosts to use the new helper function

Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@Jakuje
CVE-2026-0964 scp: Reject invalid paths received through scp
daa8081 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@cryptomilk @Jakuje
cmake: Reformat DefineOptions.cmake with cmake-format
12ccea8 
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
@cryptomilk @Jakuje
CVE-2025-14821 cmake: Fix global config location on Windows
6a7f19e 
Signed-off-by: Andreas Schneider <asn@cryptomilk.org>
Reviewed-by: Jakub Jelen <jjelen@redhat.com>
@Jakuje
misc: Fix possible NULL dereference
60ad19c 
thanks oss-fuzz

https://issues.oss-fuzz.com/u/1/issues/482613826
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
Reviewed-by: Sahana Prasad <sahana@redhat.com>
@Jakuje
fix typos
3232d72 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
Reviewed-by: Sahana Prasad <sahana@redhat.com>
@Jakuje
Copy changelog + ABI files from last minor releases
7e02580 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
@Jakuje
libssh-0.12.0
5031388 
Signed-off-by: Jakub Jelen <jjelen@redhat.com>
Reviewed-by: Pavol Žáčik <pzacik@redhat.com>
Reviewed-by: Andreas Schneider <asn@cryptomilk.org>
@alexey-milovidov
Add IdentityAgent option retrieval
79b6c45
@alexey-milovidov alexey-milovidov mentioned this pull request Aug 19, 2026
Open
@CLAassistant

CLAassistant commented Aug 19, 2026

Copy link
Copy Markdown

CLA assistant check
Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you all sign our Contributor License Agreement before we can accept your contribution.
1 out of 4 committers have signed the CLA.

✅ alexey-milovidov
❌ cryptomilk
❌ Jakuje
❌ galorithm
You have signed the CLA already but the status is still pending? Let us recheck it.

alexey-milovidov added a commit to ClickHouse/ClickHouse that referenced this pull request Aug 20, 2026
@alexey-milovidov
Honor IdentitiesOnly, expand all ssh_config tokens, and fix the t…
91ea124 
…ests

Addresses the review of #115072:

- `IdentityFile` and `IdentityAgent` now expand the whole token set that `ssh`
  supports (`%C`, `%d`, `%h`, `%i`, `%L`, `%l`, `%n`, `%p`, `%r`, `%u`, `%%`),
  instead of leaving the unknown ones as is and silently probing a file that
  cannot exist. A token that is not supported is now an error.
- `IdentitiesOnly yes` is honored: the ssh-agent is no longer asked for an
  arbitrary key when none of the configured identity files exists. Retrieval of
  the option is added to `libssh` (ClickHouse/libssh#15).
- Reading the configuration, which used to happen up to three times per
  connection, happens once, in `getSSHClientConfiguration`.
- Every failure to talk to the ssh-agent is reported as `SSH_AGENT_ERROR`. A
  socket name that is too long for a Unix socket used to escape as a
  `Poco::Exception` and abort the client instead of letting it use the key file.
- `03780_failed_ssh` no longer expects the passphrase prompt, which is now asked
  only for a key that is really encrypted.
- `04908_ssh_key_file_lookup`: the heading of the last case was printed before
  the previous one, and the new cases for `%r` and `IdentitiesOnly` are added.
@alexey-milovidov alexey-milovidov changed the title Expose the configured IdentityAgent option Expose the configured IdentityAgent and IdentitiesOnly options Aug 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.