Implement Sign() method of Channel_id for NSS

Implement Sign() method of Channel_id for NSS.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=206913

[ramant (doing other things), 2013-06-14 01:16:05]

Hi Wan-Teh,
  Tested SignAndVerify with the Verify functionality from CL
https://codereview.chromium.org/16638016/.

  Disabled QuicCryptoServerStreamTest's DISABLED_ChannelID unittest (ala
https://codereview.chromium.org/16638016/diff/49001/net/quic/quic_crypto_serv...).
After landing these changes, will investigate and fix this test.

  Could you please take a look (implemented changes as per what we had
discussed).

thanks
raman

[wtc, 2013-06-14 22:51:35]

Patch set 1 LGTM.

https://codereview.chromium.org/16126013/diff/4001/net/quic/crypto/channel_id.h
File net/quic/crypto/channel_id.h (right):

https://codereview.chromium.org/16126013/diff/4001/net/quic/crypto/channel_id...
net/quic/crypto/channel_id.h:31: virtual std::string
ChannelIDKeyForHostname(const std::string& hostname) = 0;

This method should include a verb such as "Get". I also suggest omitting
"ChannelID" -- the "Key" should be understood to be a ChannelID key in this
context. So how about "GetKeyForHostname"?

This function was originally in CryptoTestUtils, which is why it needed to
have "ChannelID" in the name.

https://codereview.chromium.org/16126013/diff/4001/net/quic/quic_crypto_serve...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://codereview.chromium.org/16126013/diff/4001/net/quic/quic_crypto_serve...
net/quic/quic_crypto_server_stream_test.cc:254:
TEST_F(QuicCryptoServerStreamTest, DISABLED_ChannelID) {

Please move your TODO comment
    // TODO(rtenneti): Enable testing of ChannelID.
before this line.  Every DISABLED test requires a TODO comment,
which ideally should include a bug number.

https://codereview.chromium.org/16126013/diff/4001/net/quic/quic_crypto_serve...
net/quic/quic_crypto_server_stream_test.cc:268:
EXPECT_EQ(signer->ChannelIDKeyForHostname("test.example.com"),

Hmm... The way this is done is incorrect because
CryptoTestUtils::ChannelIDSignerForTesting() returns a new instance
of the signer.  We should get the ChannelIDSigner of the (fake) client.

Can you try to do this?

https://codereview.chromium.org/16126013/diff/4001/net/quic/test_tools/crypto...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://codereview.chromium.org/16126013/diff/4001/net/quic/test_tools/crypto...
net/quic/test_tools/crypto_test_utils_nss.cc:39: *out_key =
SerializeKey(ecdsa_keypair);

Change SerializeKey to take a "const SECKEYPublicKey* public_key" argument.
Pass ecdsa_keypair->public_key() to SerializeKey().

https://codereview.chromium.org/16126013/diff/4001/net/quic/test_tools/crypto...
net/quic/test_tools/crypto_test_utils_nss.cc:68: std::vector<uint8>
signature_data(kSignatureLength);

Nit: You can declare signature_data as an std::string and uses Chrome's
WriteInto template. This way we don't need to go through std::vector<uint8>.

https://codereview.chromium.org/16126013/diff/4001/net/quic/test_tools/crypto...
net/quic/test_tools/crypto_test_utils_nss.cc:69: SECItem sig = {siBuffer,
&signature_data[0], kSignatureLength};

Nit: add a space after '{' and before '}'.

https://codereview.chromium.org/16126013/diff/4001/net/quic/test_tools/crypto...
net/quic/test_tools/crypto_test_utils_nss.cc:118: unsigned char* data =
public_key->u.ec.publicValue.data;

Add 'const': const unsigned char* data = ...

[ramant (doing other things), 2013-06-15 02:05:30]

Hi Wan-Teh,
  Made all the changes you have suggested. Please take another look. Verified
SignAndVerify unittest works ok (it requires your Verify CL).

thanks
raman

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/crypto/cha...
File net/quic/crypto/channel_id.h (right):

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/crypto/cha...
net/quic/crypto/channel_id.h:31: virtual std::string
ChannelIDKeyForHostname(const std::string& hostname) = 0;
On 2013/06/14 22:51:35, wtc wrote:
> 
> This method should include a verb such as "Get". I also suggest omitting
> "ChannelID" -- the "Key" should be understood to be a ChannelID key in this
> context. So how about "GetKeyForHostname"?
> 
> This function was originally in CryptoTestUtils, which is why it needed to
> have "ChannelID" in the name.

Done.

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/quic_crypt...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/quic_crypt...
net/quic/quic_crypto_server_stream_test.cc:254:
TEST_F(QuicCryptoServerStreamTest, DISABLED_ChannelID) {
On 2013/06/14 22:51:35, wtc wrote:
> 
> Please move your TODO comment
>     // TODO(rtenneti): Enable testing of ChannelID.
> before this line.  Every DISABLED test requires a TODO comment,
> which ideally should include a bug number.

Verified ChannelID is working. Enabled it.

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/quic_crypt...
net/quic/quic_crypto_server_stream_test.cc:268:
EXPECT_EQ(signer->ChannelIDKeyForHostname("test.example.com"),
On 2013/06/14 22:51:35, wtc wrote:
> 
> Hmm... The way this is done is incorrect because
> CryptoTestUtils::ChannelIDSignerForTesting() returns a new instance
> of the signer.  We should get the ChannelIDSigner of the (fake) client.
> 
> Can you try to do this?

Done.

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/test_tools...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/test_tools...
net/quic/test_tools/crypto_test_utils_nss.cc:39: *out_key =
SerializeKey(ecdsa_keypair);
On 2013/06/14 22:51:35, wtc wrote:
> 
> Change SerializeKey to take a "const SECKEYPublicKey* public_key" argument.
> Pass ecdsa_keypair->public_key() to SerializeKey().

Done. 

With this change, there is a small duplication of code to check if keypair and
public_key are not NULL. What do you think?

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/test_tools...
net/quic/test_tools/crypto_test_utils_nss.cc:68: std::vector<uint8>
signature_data(kSignatureLength);
On 2013/06/14 22:51:35, wtc wrote:
> 
> Nit: You can declare signature_data as an std::string and uses Chrome's
> WriteInto template. This way we don't need to go through std::vector<uint8>.

Done.

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/test_tools...
net/quic/test_tools/crypto_test_utils_nss.cc:69: SECItem sig = {siBuffer,
&signature_data[0], kSignatureLength};
On 2013/06/14 22:51:35, wtc wrote:
> 
> Nit: add a space after '{' and before '}'.

Done.

https://chromiumcodereview.appspot.com/16126013/diff/4001/net/quic/test_tools...
net/quic/test_tools/crypto_test_utils_nss.cc:118: unsigned char* data =
public_key->u.ec.publicValue.data;
On 2013/06/14 22:51:35, wtc wrote:
> 
> Add 'const': const unsigned char* data = ...

Done.

[wtc, 2013-06-17 18:13:22]

Patch set 2 LGTM. I suggest some changes. Thanks.

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/ch...
File net/quic/crypto/channel_id.h (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/ch...
net/quic/crypto/channel_id.h:29: // GetKeyForHostname returns the key that
|ChannelIDSigner| will use for the

Nit: add back "ChannelID" before "key" in this comment.

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/ch...
File net/quic/crypto/channel_id_openssl.cc (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/ch...
net/quic/crypto/channel_id_openssl.cc:48: BN_bin2bn(signature_bytes + 32, 32,
sig.s) == NULL) {

Nit this variable renaming destroys the nice alignment of function arguments
in the original code. So I suggest that we just remove all of the extra spaces.
Or at least, the extra spaces between |key_bytes| and '+' on lines
45-46 can be removed.

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/cr...
File net/quic/crypto/crypto_handshake.h (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/crypto/cr...
net/quic/crypto/crypto_handshake.h:359: ChannelIDSigner* channel_id_signer()
const;

I wonder if this method can return a const ChannelIDSigner*.
I guess the GetKeyForHostname() method isn't const because
it may need to generate the ChannelID key if necessary?

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
File net/quic/test_tools/crypto_test_utils.cc (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils.cc:192:
CompareClientAndServerKeys(&client, server);

Instead of returning |crypto_config| as an output argument, I suggest
we check the channel_id here:

  EXPECT_EQ(crypto_config.channel_id_signer()->GetKeyForHostname(
                "test.example.com"),
            server->crypto_negotiated_params().channel_id);

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
File net/quic/test_tools/crypto_test_utils.h (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils.h:60: QuicCryptoClientConfig*
crypto_config,

Please name this argument crypto_client_config to avoid confusion with
QuicCryptoServerConfig* crypto_config argument in other functions.

We should document the crypto_config argument. Is it a pure
output argument, or does this function use it as an in/out
argument?

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:47: }

It is not necessary to check if |public_key| is null. You
can delete lines 44-47.

If |ecdsa_keypair| is valud, its ecdsa_keypair->public_key()
member should not be null.

You can get rid of the |public_key| local variable and just call
SerializeKey(ecdsa_keypair->public_key()).

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:79: SECItem sig = {

Nit: I suggest naming this variable sig_item, to match the
hash_item variable.

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:82: WriteInto(&signature,
kSignatureLength + 1)),

Yes, it is correct to pass kSignatureLength + 1 because the
second argument of WriteInto is length_with_null.

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:103: return
SerializeKey(public_key);

This method can be simply:

  crypto::ECPrivateKey* ecdsa_keypair = HostnameToKey(hostname);
  if (!ecdsa_keypair) {
    return "";
  }
  return SerializeKey(ecdsa_keypair->public_key());

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:124: // crypto::ECPrivateKey is an
ANSI X9.62 public key which, for a P-256

Typo: crypto::ECPrivateKey => public_key->u.ec.publicValue

https://chromiumcodereview.appspot.com/16126013/diff/20002/net/quic/test_tool...
net/quic/test_tools/crypto_test_utils_nss.cc:130: const unsigned int len =
public_key->u.ec.publicValue.len;

Nit: you should do either
  const unsigned char* const data = ...;
  const unsigned int len = ...;
or
  const unsigned char* data = ...;
  unsigned int len = ...;

In the former, |data| and |len| are constants.

[ramant (doing other things), 2013-06-17 19:56:03]

PTAL. thanks.

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/channel_id.h
File net/quic/crypto/channel_id.h (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/channel_i...
net/quic/crypto/channel_id.h:29: // GetKeyForHostname returns the key that
|ChannelIDSigner| will use for the
On 2013/06/17 18:13:22, wtc wrote:
> 
> Nit: add back "ChannelID" before "key" in this comment.

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/channel_i...
File net/quic/crypto/channel_id_openssl.cc (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/channel_i...
net/quic/crypto/channel_id_openssl.cc:48: BN_bin2bn(signature_bytes + 32, 32,
sig.s) == NULL) {
On 2013/06/17 18:13:22, wtc wrote:
> 
> Nit this variable renaming destroys the nice alignment of function arguments
> in the original code. So I suggest that we just remove all of the extra
spaces.
> Or at least, the extra spaces between |key_bytes| and '+' on lines
> 45-46 can be removed.

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/crypto_ha...
File net/quic/crypto/crypto_handshake.h (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/crypto/crypto_ha...
net/quic/crypto/crypto_handshake.h:359: ChannelIDSigner* channel_id_signer()
const;
On 2013/06/17 18:13:22, wtc wrote:
> 
> I wonder if this method can return a const ChannelIDSigner*.
> I guess the GetKeyForHostname() method isn't const because
> it may need to generate the ChannelID key if necessary?

Correct.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
File net/quic/test_tools/crypto_test_utils.cc (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils.cc:192:
CompareClientAndServerKeys(&client, server);
On 2013/06/17 18:13:22, wtc wrote:
> 
> Instead of returning |crypto_config| as an output argument, I suggest
> we check the channel_id here:
> 
>   EXPECT_EQ(crypto_config.channel_id_signer()->GetKeyForHostname(
>                 "test.example.com"),
>             server->crypto_negotiated_params().channel_id);

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
File net/quic/test_tools/crypto_test_utils.h (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils.h:60: QuicCryptoClientConfig*
crypto_config,
On 2013/06/17 18:13:22, wtc wrote:
> 
> Please name this argument crypto_client_config to avoid confusion with
> QuicCryptoServerConfig* crypto_config argument in other functions.
> 
> We should document the crypto_config argument. Is it a pure
> output argument, or does this function use it as an in/out
> argument?

As we had discussed, undid the change to pass QuicCryptoClientConfig* as
argument.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:47: }
On 2013/06/17 18:13:22, wtc wrote:
> 
> It is not necessary to check if |public_key| is null. You
> can delete lines 44-47.
> 
> If |ecdsa_keypair| is valud, its ecdsa_keypair->public_key()
> member should not be null.
> 
> You can get rid of the |public_key| local variable and just call
> SerializeKey(ecdsa_keypair->public_key()).

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:79: SECItem sig = {
On 2013/06/17 18:13:22, wtc wrote:
> 
> Nit: I suggest naming this variable sig_item, to match the
> hash_item variable.

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:103: return
SerializeKey(public_key);
On 2013/06/17 18:13:22, wtc wrote:
> 
> This method can be simply:
> 
>   crypto::ECPrivateKey* ecdsa_keypair = HostnameToKey(hostname);
>   if (!ecdsa_keypair) {
>     return "";
>   }
>   return SerializeKey(ecdsa_keypair->public_key());

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:124: // crypto::ECPrivateKey is an
ANSI X9.62 public key which, for a P-256
On 2013/06/17 18:13:22, wtc wrote:
> 
> Typo: crypto::ECPrivateKey => public_key->u.ec.publicValue

Done.

https://codereview.chromium.org/16126013/diff/20002/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:130: const unsigned int len =
public_key->u.ec.publicValue.len;
On 2013/06/17 18:13:22, wtc wrote:
> 
> Nit: you should do either
>   const unsigned char* const data = ...;
>   const unsigned int len = ...;
> or
>   const unsigned char* data = ...;
>   unsigned int len = ...;
> 
> In the former, |data| and |len| are constants.

Done.

[wtc, 2013-06-17 21:46:52]

Patch set 4 LGTM. Thanks!

https://codereview.chromium.org/16126013/diff/61001/net/quic/quic_crypto_serv...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://codereview.chromium.org/16126013/diff/61001/net/quic/quic_crypto_serv...
net/quic/quic_crypto_server_stream_test.cc:260: // CompleteCryptoHandshake
verifies ChannelID's GetKeyForHostname.

ChannelID's => ChannelIDSigner's ?

But actually, it is stream_.crypto_negotiated_params().channel_id that
we are verifying.

https://codereview.chromium.org/16126013/diff/61001/net/quic/test_tools/crypt...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://codereview.chromium.org/16126013/diff/61001/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:71: static const size_t
kSignatureLength = 32 * 2;

Declare this as unsigned int instead of size_t. This will allow us to get rid
of the static_cast<unsigned int> on line 77.

[ramant (doing other things), 2013-06-17 22:02:13]

https://codereview.chromium.org/16126013/diff/61001/net/quic/quic_crypto_serv...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://codereview.chromium.org/16126013/diff/61001/net/quic/quic_crypto_serv...
net/quic/quic_crypto_server_stream_test.cc:260: // CompleteCryptoHandshake
verifies ChannelID's GetKeyForHostname.
On 2013/06/17 21:46:52, wtc wrote:
> 
> ChannelID's => ChannelIDSigner's ?
> 
> But actually, it is stream_.crypto_negotiated_params().channel_id that
> we are verifying.

Done.

https://codereview.chromium.org/16126013/diff/61001/net/quic/test_tools/crypt...
File net/quic/test_tools/crypto_test_utils_nss.cc (right):

https://codereview.chromium.org/16126013/diff/61001/net/quic/test_tools/crypt...
net/quic/test_tools/crypto_test_utils_nss.cc:71: static const size_t
kSignatureLength = 32 * 2;
On 2013/06/17 21:46:52, wtc wrote:
> 
> Declare this as unsigned int instead of size_t. This will allow us to get rid
> of the static_cast<unsigned int> on line 77.

Done.

[wtc, 2013-06-17 23:09:37]

Patch set 5 LGTM.

https://codereview.chromium.org/16126013/diff/86001/net/quic/quic_crypto_serv...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://codereview.chromium.org/16126013/diff/86001/net/quic/quic_crypto_serv...
net/quic/quic_crypto_server_stream_test.cc:261: //
(stream_.crypto_negotiated_params().channel_id) GetKeyForHostname.

This should say:
  CompleteCryptoHandshake verifies stream_.crypto_negotiated_params().channel_id
is correct.

[ramant (doing other things), 2013-06-17 23:16:46]

https://codereview.chromium.org/16126013/diff/86001/net/quic/quic_crypto_serv...
File net/quic/quic_crypto_server_stream_test.cc (right):

https://codereview.chromium.org/16126013/diff/86001/net/quic/quic_crypto_serv...
net/quic/quic_crypto_server_stream_test.cc:261: //
(stream_.crypto_negotiated_params().channel_id) GetKeyForHostname.
On 2013/06/17 23:09:37, wtc wrote:
> 
> This should say:
>   CompleteCryptoHandshake verifies
stream_.crypto_negotiated_params().channel_id
> is correct.

Done.

[wtc, 2013-06-17 23:19:05]

Patch set 6 LGTM.

[commit-bot: I haz the power, 2013-06-17 23:24:15]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/16126013/99001

[commit-bot: I haz the power, 2013-06-17 23:45:20]

Sorry for I got bad news for ya.
Compile failed with a clobber build on linux_chromeos_clang.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_chro...
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[commit-bot: I haz the power, 2013-06-18 02:05:54]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/16126013/113001

[wtc, 2013-06-18 02:34:07]

Patch set 7 LGTM.

[commit-bot: I haz the power, 2013-06-18 08:19:02]

Change committed as 206913