The borsh serialization of the HashMap did not follow the borsh specification. It potentially produced non-canonical encodings dependent on insertion order. It also did not perform canonicty checks on decoding.

This can result in consensus splits and cause equivalent objects to be considered distinct.

This was patched in 0.15.1.

In the unique reclaim path of BytesMut::reserve, the condition

if v_capacity >= new_cap + offset

uses an unchecked addition. When new_cap + offset overflows usize in release builds, this condition may incorrectly pass, causing self.cap to be set to a value that exceeds the actual allocated capacity. Subsequent APIs such as spare_capacity_mut() then trust this corrupted cap value and may create out-of-bounds slices, leading to UB.

This behavior is observable in release builds (integer overflow wraps), whereas debug builds panic due to overflow checks.

PoC

use bytes::*;

fn main() {
    let mut a = BytesMut::from(&b"hello world"[..]);
    let mut b = a.split_off(5);

    // Ensure b becomes the unique owner of the backing storage
    drop(a);

    // Trigger overflow in new_cap + offset inside reserve
    b.reserve(usize::MAX - 6);

    // This call relies on the corrupted cap and may cause UB & HBO
    b.put_u8(b'h');
}

Workarounds

Users of BytesMut::reserve are only affected if integer overflow checks are configured to wrap. When integer overflow is configured to panic, this issue does not apply.

Impact

When user-provided input is provided to any type that parses with the RFC 2822 format, a denial of service attack via stack exhaustion is possible. The attack relies on formally deprecated and rarely-used features that are part of the RFC 2822 format used in a malicious manner. Ordinary, non-malicious input will never encounter this scenario.

Patches

A limit to the depth of recursion was added in v0.3.47. From this version, an error will be returned rather than exhausting the stack.

Workarounds

Limiting the length of user input is the simplest way to avoid stack exhaustion, as the amount of the stack consumed would be at most a factor of the length of the input.

A Double Free / Use-After-Free (UAF) vulnerability has been identified in the IntoIter::drop and ThinVec::clear implementations of the thin-vec crate. Both vulnerabilities share the same root cause and can trigger memory corruption using only safe Rust code - no unsafe blocks required. Undefined Behavior has been confirmed via Miri and AddressSanitizer (ASAN).

Details

When a panic occurs during sequential element deallocation, the subsequent length cleanup code (set_len(0)) is never executed. During stack unwinding, the container is dropped again, causing already-freed memory to be re-freed (Double Free / UAF).

Vulnerability 1 - IntoIter::drop

IntoIter::drop transfers ownership of the internal buffer via mem::replace, then sequentially frees elements via ptr::drop_in_place. If a panic occurs during element deallocation, set_len_non_singleton(0) is never reached. During unwinding, vec is dropped again, re-freeing already-freed elements. The standard library's std::vec::IntoIter prevents this with a DropGuard pattern, but thin-vec lacks this defense.

PoC

use thin_vec::ThinVec;

struct PanicBomb(String);

impl Drop for PanicBomb {
    fn drop(&mut self) {
        if self.0 == "panic" {
            panic!("panic!");
        }
        println!("Dropping: {}", self.0);
    }
}

fn main() {
    let mut v = ThinVec::new();
    v.push(PanicBomb(String::from("normal1")));
    v.push(PanicBomb(String::from("panic")));  // trigger element
    v.push(PanicBomb(String::from("normal2")));

    let mut iter = v.into_iter();
    iter.next();
    // When iter is dropped: panic occurs at "panic" element
    // → During unwinding, Double Drop is triggered on "normal1" (already freed)
}

Vulnerability 2 - ThinVec::clear

clear() calls ptr::drop_in_place(&mut self[..]) followed by self.set_len(0) to reset the length. If a panic occurs during element deallocation, set_len(0) is never executed. When the ThinVec itself is subsequently dropped, already-freed elements are freed again.

PoC

use thin_vec::ThinVec;
use std::panic;

struct Poison(Box<usize>, &'static str);

impl Drop for Poison {
    fn drop(&mut self) {
        if self.1 == "panic" {
            panic!("panic!");
        }
        println!("Dropping: {}", self.0);
    }
}

fn main() {
    let mut v = ThinVec::new();
    v.push(Poison(Box::new(1), "normal1")); // index 0
    v.push(Poison(Box::new(2), "panic"));   // index 1 → panic triggered here
    v.push(Poison(Box::new(3), "normal2")); // index 2

    let _ = panic::catch_unwind(panic::AssertUnwindSafe(|| {
        v.clear();
        // panic occurs at "panic" element during clear()
        // → set_len(0) is never called
        // → already-freed elements are re-freed when v goes out of scope
    }));
}

Prerequisites

1. ThinVec stores heap-owning types (String, Vec, Box, etc.)
2. (Vulnerability 1) An iterator is created via into_iter() and dropped before being fully consumed, or (Vulnerability 2) clear() is called while a remaining element's Drop implementation can panic
3. The Drop implementation of a remaining element triggers a panic

When combined with Box<dyn Trait> types, an exploit primitive enabling Arbitrary Code Execution (ACE) via heap spray and vtable hijacking has been confirmed. If the freed fat pointer slot (16 bytes) at the point of Double Drop is reclaimed by an attacker-controlled fake vtable, subsequent Drop calls can be redirected to attacker-controlled code.