[v3,04/10] KEYS: X.509: Parse Key Usage

Message ID 20221214003401.4086781-5-eric.snowberg@oracle.com
State New
Headers
Series Add CA enforcement keyring restrictions |

Commit Message

Eric Snowberg Dec. 14, 2022, 12:33 a.m. UTC
  Parse the X.509 Key Usage.  The key usage extension defines the purpose of
the key contained in the certificate.

   id-ce-keyUsage OBJECT IDENTIFIER ::=  { id-ce 15 }

      KeyUsage ::= BIT STRING {
           digitalSignature        (0),
           contentCommitment       (1),
           keyEncipherment         (2),
           dataEncipherment        (3),
           keyAgreement            (4),
           keyCertSign             (5),
           cRLSign                 (6),
           encipherOnly            (7),
           decipherOnly            (8) }

If the keyCertSign is set, store it in the x509_certificate structure.
This will be used in a follow on patch that requires knowing the
certificate key usage type.

Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
---
 crypto/asymmetric_keys/x509_cert_parser.c | 22 ++++++++++++++++++++++
 crypto/asymmetric_keys/x509_parser.h      |  1 +
 2 files changed, 23 insertions(+)
  

Comments

Mimi Zohar Dec. 15, 2022, 11:25 a.m. UTC | #1
On Tue, 2022-12-13 at 19:33 -0500, Eric Snowberg wrote:
> Parse the X.509 Key Usage.  The key usage extension defines the purpose of
> the key contained in the certificate.
> 
>    id-ce-keyUsage OBJECT IDENTIFIER ::=  { id-ce 15 }
> 
>       KeyUsage ::= BIT STRING {
>            digitalSignature        (0),
>            contentCommitment       (1),
>            keyEncipherment         (2),
>            dataEncipherment        (3),
>            keyAgreement            (4),
>            keyCertSign             (5),
>            cRLSign                 (6),
>            encipherOnly            (7),
>            decipherOnly            (8) }
> 
> If the keyCertSign is set, store it in the x509_certificate structure.
> This will be used in a follow on patch that requires knowing the
> certificate key usage type.

Either in this patch or separately, the "digitalSignature" key usage
flag needs to be saved.

> 
> Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
> ---
>  crypto/asymmetric_keys/x509_cert_parser.c | 22 ++++++++++++++++++++++
>  crypto/asymmetric_keys/x509_parser.h      |  1 +
>  2 files changed, 23 insertions(+)
> 
> diff --git a/crypto/asymmetric_keys/x509_cert_parser.c b/crypto/asymmetric_keys/x509_cert_parser.c
> index b4443e507153..edb22cf04eed 100644
> --- a/crypto/asymmetric_keys/x509_cert_parser.c
> +++ b/crypto/asymmetric_keys/x509_cert_parser.c
> @@ -579,6 +579,28 @@ int x509_process_extension(void *context, size_t hdrlen,
>  		return 0;
>  	}
>  
> +	if (ctx->last_oid == OID_keyUsage) {
> +		/*
> +		 * Get hold of the keyUsage bit string to validate keyCertSign
> +		 * v[1] is the encoding size
> +		 *       (Expect either 0x02 or 0x03, making it 1 or 2 bytes)
> +		 * v[2] is the number of unused bits in the bit string
> +		 *       (If >= 3 keyCertSign is missing)
> +		 * v[3] and possibly v[4] contain the bit string
> +		 * 0x04 is where KeyCertSign lands in this bit string (from
> +		 *      RFC 5280 4.2.1.3)
> +		 */
> +		if (v[0] != ASN1_BTS)
> +			return -EBADMSG;
> +		if (vlen < 4)
> +			return -EBADMSG;
> +		if (v[1] == 0x02 && v[2] <= 2 && (v[3] & 0x04))
> +			ctx->cert->kcs_set = true;
> +		else if (vlen > 4 && v[1] == 0x03 && (v[3] & 0x04))
> +			ctx->cert->kcs_set = true;
> +		return 0;
> +	}
> +
>  	if (ctx->last_oid == OID_authorityKeyIdentifier) {
>  		/* Get hold of the CA key fingerprint */
>  		ctx->raw_akid = v;
> diff --git a/crypto/asymmetric_keys/x509_parser.h b/crypto/asymmetric_keys/x509_parser.h
> index 7c5c0ad1c22e..74a9f929e400 100644
> --- a/crypto/asymmetric_keys/x509_parser.h
> +++ b/crypto/asymmetric_keys/x509_parser.h
> @@ -39,6 +39,7 @@ struct x509_certificate {
>  	bool		unsupported_sig;	/* T if signature uses unsupported crypto */
>  	bool		blacklisted;
>  	bool		root_ca;		/* T if basic constraints CA is set */
> +	bool		kcs_set;		/* T if keyCertSign is set */

Using acronyms as variable names makes reviewing code more difficult. 
Perhaps rename "kcs_set" to either "key_cert_sign" or "keycertsign".

>  };
>  
>  /*
  
Jarkko Sakkinen Jan. 4, 2023, 11:43 a.m. UTC | #2
On Tue, Dec 13, 2022 at 07:33:55PM -0500, Eric Snowberg wrote:
> Parse the X.509 Key Usage.  The key usage extension defines the purpose of
> the key contained in the certificate.
> 
>    id-ce-keyUsage OBJECT IDENTIFIER ::=  { id-ce 15 }
> 
>       KeyUsage ::= BIT STRING {
>            digitalSignature        (0),
>            contentCommitment       (1),
>            keyEncipherment         (2),
>            dataEncipherment        (3),
>            keyAgreement            (4),
>            keyCertSign             (5),
>            cRLSign                 (6),
>            encipherOnly            (7),
>            decipherOnly            (8) }
> 
> If the keyCertSign is set, store it in the x509_certificate structure.
> This will be used in a follow on patch that requires knowing the
> certificate key usage type.
> 
> Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
> ---
>  crypto/asymmetric_keys/x509_cert_parser.c | 22 ++++++++++++++++++++++
>  crypto/asymmetric_keys/x509_parser.h      |  1 +
>  2 files changed, 23 insertions(+)
> 
> diff --git a/crypto/asymmetric_keys/x509_cert_parser.c b/crypto/asymmetric_keys/x509_cert_parser.c
> index b4443e507153..edb22cf04eed 100644
> --- a/crypto/asymmetric_keys/x509_cert_parser.c
> +++ b/crypto/asymmetric_keys/x509_cert_parser.c
> @@ -579,6 +579,28 @@ int x509_process_extension(void *context, size_t hdrlen,
>  		return 0;
>  	}
>  
> +	if (ctx->last_oid == OID_keyUsage) {
> +		/*
> +		 * Get hold of the keyUsage bit string to validate keyCertSign
> +		 * v[1] is the encoding size
> +		 *       (Expect either 0x02 or 0x03, making it 1 or 2 bytes)
> +		 * v[2] is the number of unused bits in the bit string
> +		 *       (If >= 3 keyCertSign is missing)
> +		 * v[3] and possibly v[4] contain the bit string
> +		 * 0x04 is where KeyCertSign lands in this bit string (from
> +		 *      RFC 5280 4.2.1.3)
> +		 */
> +		if (v[0] != ASN1_BTS)
> +			return -EBADMSG;
> +		if (vlen < 4)
> +			return -EBADMSG;
> +		if (v[1] == 0x02 && v[2] <= 2 && (v[3] & 0x04))
> +			ctx->cert->kcs_set = true;
> +		else if (vlen > 4 && v[1] == 0x03 && (v[3] & 0x04))
> +			ctx->cert->kcs_set = true;
> +		return 0;

This is much more easier to follow thanks to explanation.

> +	}
> +
>  	if (ctx->last_oid == OID_authorityKeyIdentifier) {
>  		/* Get hold of the CA key fingerprint */
>  		ctx->raw_akid = v;
> diff --git a/crypto/asymmetric_keys/x509_parser.h b/crypto/asymmetric_keys/x509_parser.h
> index 7c5c0ad1c22e..74a9f929e400 100644
> --- a/crypto/asymmetric_keys/x509_parser.h
> +++ b/crypto/asymmetric_keys/x509_parser.h
> @@ -39,6 +39,7 @@ struct x509_certificate {
>  	bool		unsupported_sig;	/* T if signature uses unsupported crypto */
>  	bool		blacklisted;
>  	bool		root_ca;		/* T if basic constraints CA is set */
> +	bool		kcs_set;		/* T if keyCertSign is set */
>  };
>  
>  /*
> -- 
> 2.27.0
> 

LGTM but I'll hold with reviewed-by's up until the patch set overally
looks good to me and I have actually tested it.

BR, Jarkko
  
Eric Snowberg Jan. 4, 2023, 9:46 p.m. UTC | #3
> On Jan 4, 2023, at 4:43 AM, Jarkko Sakkinen <jarkko@kernel.org> wrote:
> 
> On Tue, Dec 13, 2022 at 07:33:55PM -0500, Eric Snowberg wrote:
>> Parse the X.509 Key Usage.  The key usage extension defines the purpose of
>> the key contained in the certificate.
>> 
>>   id-ce-keyUsage OBJECT IDENTIFIER ::=  { id-ce 15 }
>> 
>>      KeyUsage ::= BIT STRING {
>>           digitalSignature        (0),
>>           contentCommitment       (1),
>>           keyEncipherment         (2),
>>           dataEncipherment        (3),
>>           keyAgreement            (4),
>>           keyCertSign             (5),
>>           cRLSign                 (6),
>>           encipherOnly            (7),
>>           decipherOnly            (8) }
>> 
>> If the keyCertSign is set, store it in the x509_certificate structure.
>> This will be used in a follow on patch that requires knowing the
>> certificate key usage type.
>> 
>> Signed-off-by: Eric Snowberg <eric.snowberg@oracle.com>
>> ---
>> crypto/asymmetric_keys/x509_cert_parser.c | 22 ++++++++++++++++++++++
>> crypto/asymmetric_keys/x509_parser.h      |  1 +
>> 2 files changed, 23 insertions(+)
>> 
>> diff --git a/crypto/asymmetric_keys/x509_cert_parser.c b/crypto/asymmetric_keys/x509_cert_parser.c
>> index b4443e507153..edb22cf04eed 100644
>> --- a/crypto/asymmetric_keys/x509_cert_parser.c
>> +++ b/crypto/asymmetric_keys/x509_cert_parser.c
>> @@ -579,6 +579,28 @@ int x509_process_extension(void *context, size_t hdrlen,
>> 		return 0;
>> 	}
>> 
>> +	if (ctx->last_oid == OID_keyUsage) {
>> +		/*
>> +		 * Get hold of the keyUsage bit string to validate keyCertSign
>> +		 * v[1] is the encoding size
>> +		 *       (Expect either 0x02 or 0x03, making it 1 or 2 bytes)
>> +		 * v[2] is the number of unused bits in the bit string
>> +		 *       (If >= 3 keyCertSign is missing)
>> +		 * v[3] and possibly v[4] contain the bit string
>> +		 * 0x04 is where KeyCertSign lands in this bit string (from
>> +		 *      RFC 5280 4.2.1.3)
>> +		 */
>> +		if (v[0] != ASN1_BTS)
>> +			return -EBADMSG;
>> +		if (vlen < 4)
>> +			return -EBADMSG;
>> +		if (v[1] == 0x02 && v[2] <= 2 && (v[3] & 0x04))
>> +			ctx->cert->kcs_set = true;
>> +		else if (vlen > 4 && v[1] == 0x03 && (v[3] & 0x04))
>> +			ctx->cert->kcs_set = true;
>> +		return 0;
> 
> This is much more easier to follow thanks to explanation.
> 
>> +	}
>> +
>> 	if (ctx->last_oid == OID_authorityKeyIdentifier) {
>> 		/* Get hold of the CA key fingerprint */
>> 		ctx->raw_akid = v;
>> diff --git a/crypto/asymmetric_keys/x509_parser.h b/crypto/asymmetric_keys/x509_parser.h
>> index 7c5c0ad1c22e..74a9f929e400 100644
>> --- a/crypto/asymmetric_keys/x509_parser.h
>> +++ b/crypto/asymmetric_keys/x509_parser.h
>> @@ -39,6 +39,7 @@ struct x509_certificate {
>> 	bool		unsupported_sig;	/* T if signature uses unsupported crypto */
>> 	bool		blacklisted;
>> 	bool		root_ca;		/* T if basic constraints CA is set */
>> +	bool		kcs_set;		/* T if keyCertSign is set */
>> };
>> 
>> /*
>> -- 
>> 2.27.0
>> 
> 
> LGTM but I'll hold with reviewed-by's up until the patch set overally
> looks good to me and I have actually tested it.

Thanks for your review.  I will make all the other changes you brought up with
the other patches in the next round.
  

Patch

diff --git a/crypto/asymmetric_keys/x509_cert_parser.c b/crypto/asymmetric_keys/x509_cert_parser.c
index b4443e507153..edb22cf04eed 100644
--- a/crypto/asymmetric_keys/x509_cert_parser.c
+++ b/crypto/asymmetric_keys/x509_cert_parser.c
@@ -579,6 +579,28 @@  int x509_process_extension(void *context, size_t hdrlen,
 		return 0;
 	}
 
+	if (ctx->last_oid == OID_keyUsage) {
+		/*
+		 * Get hold of the keyUsage bit string to validate keyCertSign
+		 * v[1] is the encoding size
+		 *       (Expect either 0x02 or 0x03, making it 1 or 2 bytes)
+		 * v[2] is the number of unused bits in the bit string
+		 *       (If >= 3 keyCertSign is missing)
+		 * v[3] and possibly v[4] contain the bit string
+		 * 0x04 is where KeyCertSign lands in this bit string (from
+		 *      RFC 5280 4.2.1.3)
+		 */
+		if (v[0] != ASN1_BTS)
+			return -EBADMSG;
+		if (vlen < 4)
+			return -EBADMSG;
+		if (v[1] == 0x02 && v[2] <= 2 && (v[3] & 0x04))
+			ctx->cert->kcs_set = true;
+		else if (vlen > 4 && v[1] == 0x03 && (v[3] & 0x04))
+			ctx->cert->kcs_set = true;
+		return 0;
+	}
+
 	if (ctx->last_oid == OID_authorityKeyIdentifier) {
 		/* Get hold of the CA key fingerprint */
 		ctx->raw_akid = v;
diff --git a/crypto/asymmetric_keys/x509_parser.h b/crypto/asymmetric_keys/x509_parser.h
index 7c5c0ad1c22e..74a9f929e400 100644
--- a/crypto/asymmetric_keys/x509_parser.h
+++ b/crypto/asymmetric_keys/x509_parser.h
@@ -39,6 +39,7 @@  struct x509_certificate {
 	bool		unsupported_sig;	/* T if signature uses unsupported crypto */
 	bool		blacklisted;
 	bool		root_ca;		/* T if basic constraints CA is set */
+	bool		kcs_set;		/* T if keyCertSign is set */
 };
 
 /*