Uninitialized buffer exposed to user-provided fn `Reader::read_bytes()`

[JOE1994]

Hello 🦀,
we (Rust group @sslab-gatech) found a memory-safety/soundness issue in this crate while scanning Rust code on crates.io for potential vulnerabilities.

Issue Description

speedy/src/reader.rs

Lines 269 to 273 in 81d4f02

	let mut vec = Vec::with_capacity( length );
	unsafe {
	vec.set_len( length );
	self.read_bytes( T::speedy_slice_as_bytes_mut( &mut vec ) )?;
	}

reader::Reader::read_vec() method creates an uninitialized buffer and passes it to user-provided Reader::read_bytes() implementation. This allows the user-provided implemented read_bytes() to read in uninitialized bytes. This is unsound, because it allows safe Rust code to exhibit an undefined behavior (read from uninitialized memory).

This is very similar to the issue that RFC 2930 tries to resolve. In case a user-provided Read reads from the given buffer, uninitialized buffer can make safe Rust code to cause memory safety errors by miscompilation. Uninitialized values are lowered to LLVM as llvm::UndefValue which may take different random values for each read. Propagation of UndefValue can quickly cause safe Rust code to exhibit undefined behavior.

This part from the Read trait documentation explains the issue:

It is your responsibility to make sure that buf is initialized before calling read. Calling read with an uninitialized buf (of the kind one obtains via MaybeUninit<T>) is not safe, and can lead to undefined behavior.

How to fix the issue?

The Naive & safe way to fix the issue is to always zero-initialize a buffer before lending it to Reader::read_bytes() implementation. Note that this approach will add runtime performance overhead of zero-initializing the buffer.

As of Feb 2021, there is not yet an ideal fix that works with no performance overhead. Below are links to relevant discussions & suggestions for the fix.

• Rust RFC 2930
• Discussion in Rust Internals Forum