netdev - Re: [PATCH -next,v2] gss_krb5: refactor code to return correct PTR_ERR in krb5

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <ZpE9luTCrUnh8RBH@tissot.1015granger.net>
Date: Fri, 12 Jul 2024 10:28:38 -0400
From: Chuck Lever <chuck.lever@...cle.com>
To: Gaosheng Cui <cuigaosheng1@...wei.com>
Cc: trondmy@...nel.org, anna@...nel.org, jlayton@...nel.org, neilb@...e.de,
        kolga@...app.com, Dai.Ngo@...cle.com, tom@...pey.com,
        davem@...emloft.net, edumazet@...gle.com, kuba@...nel.org,
        pabeni@...hat.com, horms@...nel.org, linux-nfs@...r.kernel.org,
        netdev@...r.kernel.org
Subject: Re: [PATCH -next,v2] gss_krb5: refactor code to return correct
 PTR_ERR in krb5_DK

On Fri, Jul 12, 2024 at 09:39:08AM -0400, Chuck Lever wrote:
> On Fri, Jul 12, 2024 at 03:24:23PM +0800, Gaosheng Cui wrote:
> > Refactor the code in krb5_DK to return PTR_ERR when an error occurs.
> 
> My understanding of the current code is that if either
> crypto_alloc_sync_skcipher() or crypto_sync_skcipher_blocksize()
> fails, then krb5_DK() returns -EINVAL. At the only call site for
> krb5_DK(), that return code is unconditionally discarded. Thus I
> don't see that the proposed change is necessary or improves
> anything.

My understanding is wrong  ;-)

The return code isn't discarded. A non-zero return code from
krb5_DK() is carried back up the call stack. The logic in
krb5_derive_key_v2() does not use the kernel's usual error flow
form, so I missed this.

However, it still isn't clear to me why the error behavior here
needs to change. It's possible, for example, that -EINVAL is
perfectly adequate to indicate when sync_skcipher() can't find the
specified encryption algorithm (gk5e->encrypt_name).

Specifying the wrong encryption type: -EINVAL. That makes sense.


> > Signed-off-by: Gaosheng Cui <cuigaosheng1@...wei.com>
> > ---
> > v2: Update IS_ERR to PTR_ERR, thanks very much!
> >  net/sunrpc/auth_gss/gss_krb5_keys.c | 8 ++++++--
> >  1 file changed, 6 insertions(+), 2 deletions(-)
> > 
> > diff --git a/net/sunrpc/auth_gss/gss_krb5_keys.c b/net/sunrpc/auth_gss/gss_krb5_keys.c
> > index 4eb19c3a54c7..5ac8d06ab2c0 100644
> > --- a/net/sunrpc/auth_gss/gss_krb5_keys.c
> > +++ b/net/sunrpc/auth_gss/gss_krb5_keys.c
> > @@ -164,10 +164,14 @@ static int krb5_DK(const struct gss_krb5_enctype *gk5e,
> >  		goto err_return;
> >  
> >  	cipher = crypto_alloc_sync_skcipher(gk5e->encrypt_name, 0, 0);
> > -	if (IS_ERR(cipher))
> > +	if (IS_ERR(cipher)) {
> > +		ret = PTR_ERR(cipher);
> >  		goto err_return;
> > +	}
> > +
> >  	blocksize = crypto_sync_skcipher_blocksize(cipher);
> > -	if (crypto_sync_skcipher_setkey(cipher, inkey->data, inkey->len))
> > +	ret = crypto_sync_skcipher_setkey(cipher, inkey->data, inkey->len);
> > +	if (ret)
> >  		goto err_free_cipher;
> >  
> >  	ret = -ENOMEM;
> > -- 
> > 2.25.1
> > 
> 
> -- 
> Chuck Lever
> 

-- 
Chuck Lever