I believe this fixes https://github.com/golang/go/issues/36126 .
The problem was that it was keeping only the first known key of each
type found.
If you have a server advertising multiple keys of the same type,
you might get a missmatch key error.
Per sshd(8) man page, it should allow reapeatable hosts with
different host keys, although it don't specify anything about
hosts being from different types:
"It is permissible (but not recommended) to have several lines or
different host keys for the same names. This will inevitably happen when
short forms of host names from different domains are put in the file. It
is possible that the files contain conflicting information;
authentication is accepted if valid information can be found from either
file."
So, this changes knownhosts behavior to accept any of the keys for a
given host, regardless of type.
Fixes#36126
Change-Id: I3450ff954259a403f2471082d013a5f79def0e16
GitHub-Last-Rev: 361bd2bcd20348956aaf114ef159a5350397eaf4
GitHub-Pull-Request: golang/crypto#254
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/478535
Reviewed-by: Junyang Shao <shaojunyang@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
In the SSH protocol, clients and servers execute the key exchange to
generate one-time session keys used for encryption and authentication.
The key exchange is performed initially after the connection is
established and then periodically after a configurable amount of data.
While a key exchange is in progress, we add the received packets to an
internal queue until we receive SSH_MSG_KEXINIT from the other side.
This can result in high memory usage if the other party is slow to
respond to the SSH_MSG_KEXINIT packet, or memory exhaustion if a
malicious client never responds to an SSH_MSG_KEXINIT packet during a
large file transfer.
We now limit the internal queue to 64 packets: this means 2MB with the
typical 32KB packet size.
When the internal queue is full we block further writes until the
pending key exchange is completed or there is a read or write error.
Thanks to Yuichi Watanabe for reporting this issue.
Change-Id: I1ce2214cc16e08b838d4bc346c74c72addafaeec
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/652135
Reviewed-by: Neal Patel <nealpatel@google.com>
Auto-Submit: Gopher Robot <gobot@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Users of the the ssh package seem to extremely commonly misuse the
PublicKeyCallback API, assuming that the key passed in the last call
before a connection is established is the key used for authentication.
Some users then make authorization decisions based on this key. This
property is not documented, and may not be correct, due to the caching
behavior of the package, resulting in users making incorrect
authorization decisions about the connection.
This change makes the cache a one entry FIFO cache, making the assumed
property, that the last call to PublicKeyCallback represents the key
actually used for authentication, actually hold.
Thanks to Damien Tournoud, Patrick Dawkins, Vince Parker, and
Jules Duvivier from the Platform.sh / Upsun engineering team
for reporting this issue.
Fixesgolang/go#70779
Fixes CVE-2024-45337
Change-Id: Ife7c7b4045d8b6bcd7e3a417bdfae370c709797f
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/635315
Reviewed-by: Roland Shoemaker <roland@golang.org>
Auto-Submit: Gopher Robot <gobot@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Seems the OpenSSH server running on windows fails keyboard-interactive
auth this way without sending any prompt to client. In such case the
golang ssh client should not retry keyboard-interactive auth when the
auth method is wrapped in a RetryableAuthMethod(). Rather the auth
method should be immediately marked as tried&failed and the client auth
process should move on to next available and acceptable auth method.
Fixesgolang/go#67855
Change-Id: I6c64ae58ff8325774e37af716601b112f8833d8f
GitHub-Last-Rev: 7fafc4d1c81284b31000d7d6ccadd934dda26d24
GitHub-Pull-Request: golang/crypto#297
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/590956
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
When adding a new key, if we already have a Signer with the same public
key, we now replace it with the new one instead of duplicating it.
Before this change we had this:
$ ssh-add -l
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA)
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA-CERT)
$ ssh-add /home/nicola/ssh_certs/id_rsa
Identity added: /home/nicola/ssh_certs/id_rsa (nicola@p1)
Certificate added: /home/nicola/ssh_certs/id_rsa-cert.pub (myid)
$ ssh-add -l
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA)
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA-CERT)
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA)
3072 SHA256:bsBRHC/xgiqBJdSuvSTNpJNLTISP/G356jNMCRYC5Es nicola@p1 (RSA-CERT)
Change-Id: Iad1b1a6dc94f68f53f05d7d1172f0017839976fc
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/602955
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
The original SSH RFC 4253 explicitly disallows padding. This applies to
ssh-rsa signatures.
The updated SSH RFC 8332 which defines the SHA2 RSA signature variants
explicitly calls out the existence of signers who produce short
signatures and specifies that verifiers may allow this behavior.
In practice, PuTTY 0.81 and prior versions, as well as SSH.NET prior to
2024.1.0 always generated short signatures. Furthermore, PuTTY is
embedded in other software like WinSCP and FileZilla, which are updated
on their own schedules as well. This leads to occasional unexplained
login errors, when using RSA keys.
OpenSSH server allows these short signatures for all RSA algorithms.
Fixesgolang/go#68286
Change-Id: Ia60ece21bf9c111c490fac0c066443ed5ff7dd29
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/598534
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
on RHEL 9 OpenSSH does not support DSA keys and RSA keys with size less
than 2048 bits, furthermore signing with ssh-rsa (SHA-1 signatures) and
ssh-dss is not allowed, therefore:
1) replaced the 1024-bit RSA key used in the test with a new 2048-bit
one
2) removed DSA key from itegration tests
3) allowed signature errors using ssh-rsa in agent integration tests, we
also check SHA-2 variants that are not skipped
Fixesgolang/go#65581
Change-Id: I54bf997b61ef4d91d38eb624275737ba7291bb20
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/562755
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: David Chase <drchase@google.com>
Reviewed-by: Filippo Valsorda <filippo@golang.org>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
We don't yet understand why these tests fail, but the Apple sshd seems
to have some non-trivial vendor patches, so it is plausibly a
platform-specific bug in the test. Let's skip that failure mode on the
whole platform until/unless someone has time to reproduce and
investigate the failure.
For golang/go#64959.
Cq-Include-Trybots: luci.golang.try:x_crypto-gotip-darwin-amd64-longtest,x_crypto-gotip-linux-amd64-longtest,x_crypto-gotip-windows-amd64-longtest
Change-Id: I9e43579469de3fe9329c093b5916bbed0edd3751
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/554077
Reviewed-by: Dmitri Shuralyov <dmitshur@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Bryan Mills <bcmills@google.com>
Implement the "strict KEX" protocol changes, as described in section
1.9 of the OpenSSH PROTOCOL file (as of OpenSSH version 9.6/9.6p1).
Namely this makes the following changes:
* Both the server and the client add an additional algorithm to the
initial KEXINIT message, indicating support for the strict KEX mode.
* When one side of the connection sees the strict KEX extension
algorithm, the strict KEX mode is enabled for messages originating
from the other side of the connection. If the sequence number for
the side which requested the extension is not 1 (indicating that it
has already received non-KEXINIT packets), the connection is
terminated.
* When strict kex mode is enabled, unexpected messages during the
handshake are considered fatal. Additionally when a key change
occurs (on the receipt of the NEWKEYS message) the message sequence
numbers are reset.
Thanks to Fabian Bäumer, Marcus Brinkmann, and Jörg Schwenk from Ruhr
University Bochum for reporting this issue.
Fixes CVE-2023-48795
Fixesgolang/go#64784
Change-Id: I96b53afd2bd2fb94d2b6f2a46a5dacf325357604
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/550715
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Tatiana Bradley <tatianabradley@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Run-TryBot: Roland Shoemaker <roland@golang.org>
Reviewed-by: Damien Neil <dneil@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Sending a window adjustment after every read is unnecessarily chatty,
especially with a series of small reads like with TTY interactions.
Copy OpenSSH's logic for deferring these, which seemingly hasn't changed
since 2007. Note that since channelWindowSize and c.maxIncomingPayload
are currently constants here, the two checks could be combined into a
single check for c.myWindow < 2 MiB - 96 KiB (with the current values
of the constants).
Fixesgolang/go#57424.
Change-Id: Ifcef5be76fcc3f0b1a6dc396096bed9c50d64f21
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/459915
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Michael Knyszek <mknyszek@google.com>
Run-TryBot: Nicola Murino <nicola.murino@gmail.com>
Auto-Submit: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Commit-Queue: Nicola Murino <nicola.murino@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
These tests try to ensure better compatibility of our server implementation
with the ssh CLI.
With these tests in place:
1) before merging CL 447757 we would have noticed that our server
implementation was broken with OpenSSH 8.8+
2) after merging CL 447757 we would have noticed that our server
implementation was broken with OpenSSH 7.2-7.7
The ssh CLI from $PATH is used by default, but can be overridden using
the SSH_CLI_PATH environment variable.
Change-Id: I93d64be41c7613132b0364afac8397f57c2dcbca
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/506837
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Benny Siegert <bsiegert@gmail.com>
Reviewed-by: Han-Wen Nienhuys <hanwen@google.com>
Run-TryBot: Nicola Murino <nicola.murino@gmail.com>
This is a follow-up to CL 317169, which dropped go1.12 compatibility,
and made the golang.org/x/crypto/ed25519 package an alias / wrapper for
crypto/ed25519 in stdlib.
This patch updates uses within this repository to use stdlib instead of
depending on the wrapper. With this patch applied, the only remaining
use of the wrapper is in ed25519_test, which appears to be in place to
verify compatibility of the wrapper itself.
Change-Id: I0195396102a75ae20bdd82ca8ab59855c0eb5cea
GitHub-Last-Rev: 24dbec563cbd84bc47bdc7736b0245fc83dd3353
GitHub-Pull-Request: golang/crypto#238
Reviewed-on: https://go-review.googlesource.com/c/crypto/+/448238
Reviewed-by: Bryan Mills <bcmills@google.com>
Run-TryBot: Nicola Murino <nicola.murino@gmail.com>
Reviewed-by: Nicola Murino <nicola.murino@gmail.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Joedian Reid <joedian@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Than McIntosh <thanm@google.com>
Nicola MurinoandDamien Miller
Carlos A Becker
Ilia Mirkin
cuishuang
Sean Liao
Brad Fitzpatrick
Roland Shoemaker
samiponkanen
Filippo Valsorda
Mariano Cano
Andrew Lytvynov
Bryan C. Mills
Roland Shoemaker
Edoardo Spadolini
Will Mortensen
Pavel Repin
Heschi Kreinick
Randy Reddig
Matt Dainty
Dmitri Shuralyov
Sebastiaan van Stijn