Re: [PATCH 1/2] ima_evm_utils: erroneous "verification failed: 0 (invalid padding)" message

[Mimi Zohar <zohar@linux.ibm.com>]

On Wed, 2019-07-17 at 00:37 +0300, Vitaly Chikunov wrote:
> Mimi,
> 
> On Tue, Jul 16, 2019 at 10:30:16AM -0400, Mimi Zohar wrote:
> > When keys are not provided, the default key is used to verify the file
> > signature, resulting in this erroneous message.  Before using the default
> > key to verify the file signature, verify the keyid is correct.
> 
> 1. What is default key when keys are not provided?

The defaults was either "/etc/keys/x509_evm.der" for DIGSIG_VERSION_1
or "/etc/keys/pubkey_evm.pem" for DIGSIG_VERSION_2.

> 2. I don't see keyid verification in the patch.

Since no keys were provided, this patch loads the default key onto the
list of public_keys.  The normal find_keyid() is then called.

> 3. Now we have so complicated keyfile handling.

It would definitely be simpler to remove support for these default
keys, but I'm hesitant to remove it.

> 
> > 
> > Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
> > ---
> >  src/libimaevm.c | 19 ++++++++++---------
> >  1 file changed, 10 insertions(+), 9 deletions(-)
> > 
> > diff --git a/src/libimaevm.c b/src/libimaevm.c
> > index ae487f9fe36c..472ab53c7b42 100644
> > --- a/src/libimaevm.c
> > +++ b/src/libimaevm.c
> > @@ -302,6 +302,9 @@ EVP_PKEY *read_pub_pkey(const char *keyfile, int x509)
> >  	X509 *crt = NULL;
> >  	EVP_PKEY *pkey = NULL;
> >  
> > +	if (!keyfile)
> > +		return NULL;
> > +
> >  	fp = fopen(keyfile, "r");
> >  	if (!fp) {
> >  		log_err("Failed to open keyfile: %s\n", keyfile);
> > @@ -569,27 +572,25 @@ static int get_hash_algo_from_sig(unsigned char *sig)
> >  int verify_hash(const char *file, const unsigned char *hash, int size, unsigned char *sig,
> >  		int siglen)
> >  {
> > -	const char *key;
> > -	int x509;
> > +	const char *key = NULL;
> >  	verify_hash_fn_t verify_hash;
> >  
> >  	/* Get signature type from sig header */
> >  	if (sig[0] == DIGSIG_VERSION_1) {
> >  		verify_hash = verify_hash_v1;
> > +
> >  		/* Read pubkey from RSA key */
> > -		x509 = 0;
> > +		if (!params.keyfile)
> > +			key = "/etc/keys/pubkey_evm.pem";
> >  	} else if (sig[0] == DIGSIG_VERSION_2) {
> >  		verify_hash = verify_hash_v2;
> > +
> >  		/* Read pubkey from x509 cert */
> > -		x509 = 1;
> > +		if (!params.keyfile)
> > +			init_public_keys("/etc/keys/x509_evm.der");
> 
> Also, in "ima-evm-utils: Preload public keys for ima_verify" I call
> init_public_keys in cmd_verify_ima() which calls this verify_hash().
> 
> So there will be double calling of init_public_keys().

init_public_keys() will only be called once, either there or here.  It
depends on whether params.keyfile is defined or not.

> 
> ps. Btw, I think we should remove this verify_hash_fn_t indirect call
> trick and replace it with two normal calls in each if branch.
> 
> verify_hash() calling verify_hash() obscures cscope, and with direct
> calls it will be easier to parse. I may send patch for this.
> 
> Thanks,

Agreed, every time I come back to this code it takes me a few minutes
to remember.

Mimi

> 
> 
> >  	} else
> >  		return -1;
> >  
> > -	/* Determine what key to use for verification*/
> > -	key = params.keyfile ? : x509 ?
> > -			"/etc/keys/x509_evm.der" :
> > -			"/etc/keys/pubkey_evm.pem";
> > -
> >  	return verify_hash(file, hash, size, sig, siglen, key);
> >  }
> >  
> > -- 
> > 2.7.5