Re: [PATCH] crypto: chacha20poly1305 - Fix missing state zeroization in xchacha decrypt

[Eric Biggers <ebiggers@kernel.org>]

On Sat, Oct 03, 2026 at 11:28:20AM +0200, Ard Biesheuvel wrote:
> 
> 
> On Sat, 3 Oct 2026, at 10:57, Eric Biggers wrote:
> > On Sat, Oct 03, 2026 at 09:34:21AM +0200, Ard Biesheuvel wrote:
> >> 
> >> 
> >> On Sat, 3 Oct 2026, at 08:08, Mohamad Raizudeen wrote:
> >> > The `__chacha20poly1305_decrypt` function does not zeroize the chacha
> >> > state, unlike its encrypt counterpart. The regular
> >> > `chacha20poly1305_decrypt` function handles this by manually calling
> >> > chacha_zeroize_state(). However, `xchacha20poly1305_decrypt` returns
> >> > the result directly without clearing the state.
> >> >
> >> > This leaves the derived chacha20 subkey on the stack after the function
> >> > returns. Fix this by storing the return value, calling
> >> > chacha_zeroize_state() and then returning the result, matching the
> >> > logic in `chacha20poly1305_decrypt`.
> >> >
> >> > Fixes: ed20078b7e333 ("crypto: chacha20poly1305 - import construction 
> >> > and selftest from Zinc")
> >> > Cc: stable@vger.kernel.org
> >> > Signed-off-by: Mohamad Raizudeen <raizudeen.kerneldev@gmail.com>
> >> > ---
> >> >  lib/crypto/chacha20poly1305.c | 5 ++++-
> >> >  1 file changed, 4 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/lib/crypto/chacha20poly1305.c 
> >> > b/lib/crypto/chacha20poly1305.c
> >> > index ea42a28f4ff7..03b14d272520 100644
> >> > --- a/lib/crypto/chacha20poly1305.c
> >> > +++ b/lib/crypto/chacha20poly1305.c
> >> > @@ -199,10 +199,13 @@ bool xchacha20poly1305_decrypt(u8 *dst, const u8 
> >> > *src, const size_t src_len,
> >> >  			       const u8 key[at_least CHACHA20POLY1305_KEY_SIZE])
> >> >  {
> >> >  	struct chacha_state chacha_state;
> >> > +	bool ret;
> >> > 
> >> >  	xchacha_init(&chacha_state, key, nonce);
> >> > -	return __chacha20poly1305_decrypt(dst, src, src_len, ad, ad_len,
> >> > +	ret = __chacha20poly1305_decrypt(dst, src, src_len, ad, ad_len,
> >> >  					  &chacha_state);
> >> > +	chacha_zeroize_state(&chacha_state);
> >> > +	return ret;
> >> >  }
> >> >  EXPORT_SYMBOL(xchacha20poly1305_decrypt);
> >> > 
> >> 
> >> Wouldn't it be better to move the existing call from chacha20poly1305_decrypt()
> >> to __chacha20poly1305_decrypt()?
> >
> > Since __chacha20poly1305_decrypt() has two return statements that would
> > need to be considered, I think this patch (which makes each
> > *chacha_init() clearly paired with chacha_zeroize_state()) is slightly
> > cleaner.
> >
> 
> Fair enough.

Actually, looking at the whole file, in the encryption case
__chacha20poly1305_encrypt() already "takes ownership" of the state and
handles zeroizing it.  So given that, I think it does make sense for
__chacha20poly1305_decrypt() to do the same thing, for consistency
between the encryption and decryption cases.

Mohamad, could you send out a v2 that does that?

Also, for the subject prefix, use: "lib/crypto: chacha20poly1305:"

Thanks!

- Eric