Encrypt certificate reports before uploading to HTTP URLs

Encrypt certificate reports before uploading to HTTP URLs

This CL introduces a new protobuf to store encrypted
CertLoggerRequests. Serialized certificate reports are encrypted with an
AES-CTR-128-HMAC-SHA256 AEAD (from BoringSSL, thus encrypted reports are
only supported on BoringSSL platforms) before being uploaded to HTTP
endpoints. |CertificateErrorReporter::IsHttpUploadUrlSupported| allows
users of the class to set an HTTP URL if supported.

BUG=461590
Committed: https://crrev.com/0a3351c2a7c81284f82e6531380a21d079f55056
Cr-Commit-Position: refs/heads/master@{#326876}

Committed: https://crrev.com/03206a1fcfe475604570d49dca0a885591e45ad7
Cr-Commit-Position: refs/heads/master@{#326957}

[estark, 2015-04-13 23:43:57]

estark@chromium.org changed reviewers:
+ agl@chromium.org

[estark, 2015-04-13 23:43:58]

agl, could you please take a look?

I had previously sent you a similar CL as a rough draft (975623002), but
certificate reporting has been significantly refactored since then so I thought
it might be better to open a new one. I think I've addressed your comments from
975623002 that are still applicable here.

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:166: bool Seal(const
std::string& aead_key,
I wasn't sure whether to put the AEAD code here or in //crypto so that it can be
used elsewhere if other needs come up. (There's the slight advantage that I
don't have to implement AD support if I put it here, but that's not a big deal.)
Any thoughts?

[estark, 2015-04-13 23:45:02]

cc'ing felt -- feel free to review too if you have comments, obviously

[felt, 2015-04-14 04:30:15]

felt@chromium.org changed reviewers:
+ felt@chromium.org

[felt, 2015-04-14 04:30:16]

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:29: // TODO(estark): insert the
production public key
Why's it missing here?

[estark, 2015-04-14 04:39:16]

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:29: // TODO(estark): insert the
production public key
On 2015/04/14 04:30:16, felt wrote:
> Why's it missing here?

Um, I don't know. I thought I had a reason for not putting in the production one
yet, but now I can't remember what it was. :( I guess I might as well use the
real thing.

[estark, 2015-04-16 15:04:24]

friendly ping, agl :)

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/40001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:29: // TODO(estark): insert the
production public key
On 2015/04/14 04:30:16, felt wrote:
> Why's it missing here?

Done.

[agl, 2015-04-16 17:26:29]

Sorry for missing this email the first time, and I have to run off to a meeting
now, but partial review...

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:6: 
Overall comment, this AEAD is in BoringSSL, as is HKDF. I was assuming that you
would call that version!

We don't yet have it in Linux, but I think it's reasonable to #ifdef the
functionality out until we do. (And you can build with -Duse_openssl=1 for
development.) Do we really need two implementations?

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:27: namespace {
blank line after this.

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:62: void CopyUint64(uint64
value, std::string* input) {
Name it AppendUint64LE or something similar?

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.h:27: bool Seal(const std::string&
aead_key,
Make these static functions in CertificateErrorReporter? Even in a namespace,
|Seal| is a pretty generic name.

[estark, 2015-04-16 18:45:12]

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/80001/chrome/browser/net/cert...
chrome/browser/net/certificate_error_reporter.cc:6: 
On 2015/04/16 17:26:29, agl wrote:
> Overall comment, this AEAD is in BoringSSL, as is HKDF. I was assuming that
you
> would call that version!
> 
> We don't yet have it in Linux, but I think it's reasonable to #ifdef the
> functionality out until we do. (And you can build with -Duse_openssl=1 for
> development.) Do we really need two implementations?

Ah, it didn't occur to me that we could just drop the feature on non-BoringSSL
platforms. Do you know if BoringSSL is used on Android yet?

[agl, 2015-04-16 18:47:30]

On 2015/04/16 18:45:12, estark wrote:
> Ah, it didn't occur to me that we could just drop the feature on non-BoringSSL
> platforms. Do you know if BoringSSL is used on Android yet?

It's everything but Linux, iOS (?) and ChromeOS at the moment I believe. And
davidben and sleevi have plans for a hybrid setup with BoringSSL and NSS for
certificate verification on those platforms.

So if it's critical that we support those platforms then we'll have to duplicate
the code and hopefully clean up later. But if we can live without them for a
little while, it would be nice not to have to worry about the extra code.

[estark, 2015-04-17 03:51:35]

On 2015/04/16 18:47:30, agl wrote:
> On 2015/04/16 18:45:12, estark wrote:
> > Ah, it didn't occur to me that we could just drop the feature on
non-BoringSSL
> > platforms. Do you know if BoringSSL is used on Android yet?
> 
> It's everything but Linux, iOS (?) and ChromeOS at the moment I believe. And
> davidben and sleevi have plans for a hybrid setup with BoringSSL and NSS for
> certificate verification on those platforms.
> 
> So if it's critical that we support those platforms then we'll have to
duplicate
> the code and hopefully clean up later. But if we can live without them for a
> little while, it would be nice not to have to worry about the extra code.

Ok, great, I talked to davidben and his hope is to get BoringSSL into Linux and
ChromeOS for M45 (though meeting that goal is somewhat dependent on rsleevi's
review time). Since there's a concrete plan in the works, I agree that we should
avoid having two implementations. So I'll update this CL to just use the
BoringSSL implementation tomorrow.

[estark, 2015-04-18 00:39:12]

Patchset #9 (id:160001) has been deleted

[estark, 2015-04-18 05:24:43]

Adam, could you please take another look?

In this iteration I'm using BoringSSL's implementation (exposed as
crypto::Aead). Is it okay to add this to //crypto, or should I keep it all in
CertificateErrorReporter and add a section to chrome.gyp to link with BoringSSL?

Another question I had: I'd like to have it so that non-BoringSSL clients still
try to send unencrypted reports over HTTPS, but the way I've done that in this
iteration is kind of clunky -- see inline comment. (Adrienne, maybe you have
thoughts on this?)

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:14: #if defined(USE_OPENSSL)
Should I be checking OPENSSL_IS_BORINGSSL too? It seems to not work (is never
defined).

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.h:87: static bool
IsHttpUploadUrlSupported();
This seems weird to me but I'm not sure what else to do. I want to let users of
this class send reports over HTTPS when sending them encrypted over HTTP isn't
supported.

The name is weird because it only applies to extended reporting, not to Google
pinning violation reports, so maybe I should rename it to
IsHttpUploadUrlSupportedForExtendedReporting?

Open to other suggestions too... maybe users of this class shouldn't specify the
upload URL, and CertificateErrorReporter should just decide itself which URL to
use for a given report based on the report type and BoringSSL support.

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h
File crypto/aead_openssl.h (right):

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:22: explicit Aead(const std::string* key);
Wasn't sure if this should take an algorithm parameter of some kind if only one
algorithm is supported.

[estark, 2015-04-21 20:01:33]

On 2015/04/18 05:24:43, estark wrote:
> Adam, could you please take another look?
> 
> In this iteration I'm using BoringSSL's implementation (exposed as
> crypto::Aead). Is it okay to add this to //crypto, or should I keep it all in
> CertificateErrorReporter and add a section to chrome.gyp to link with
BoringSSL?
> 
> Another question I had: I'd like to have it so that non-BoringSSL clients
still
> try to send unencrypted reports over HTTPS, but the way I've done that in this
> iteration is kind of clunky -- see inline comment. (Adrienne, maybe you have
> thoughts on this?)
> 
>
https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
> File chrome/browser/net/certificate_error_reporter.cc (right):
> 
>
https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
> chrome/browser/net/certificate_error_reporter.cc:14: #if defined(USE_OPENSSL)
> Should I be checking OPENSSL_IS_BORINGSSL too? It seems to not work (is never
> defined).
> 
>
https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
> File chrome/browser/net/certificate_error_reporter.h (right):
> 
>
https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
> chrome/browser/net/certificate_error_reporter.h:87: static bool
> IsHttpUploadUrlSupported();
> This seems weird to me but I'm not sure what else to do. I want to let users
of
> this class send reports over HTTPS when sending them encrypted over HTTP isn't
> supported.
> 
> The name is weird because it only applies to extended reporting, not to Google
> pinning violation reports, so maybe I should rename it to
> IsHttpUploadUrlSupportedForExtendedReporting?
> 
> Open to other suggestions too... maybe users of this class shouldn't specify
the
> upload URL, and CertificateErrorReporter should just decide itself which URL
to
> use for a given report based on the report type and BoringSSL support.
> 
> https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h
> File crypto/aead_openssl.h (right):
> 
>
https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
> crypto/aead_openssl.h:22: explicit Aead(const std::string* key);
> Wasn't sure if this should take an algorithm parameter of some kind if only
one
> algorithm is supported.

agl: this is a ping as requested at lunch :)

[agl, 2015-04-21 20:28:26]

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:14: #if defined(USE_OPENSSL)
On 2015/04/18 05:24:42, estark wrote:
> Should I be checking OPENSSL_IS_BORINGSSL too? It seems to not work (is never
> defined).

No, OpenSSL in Chrome is always BoringSSL.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:54: std::string(),
std::string(), 0, 0,
I'd suggest adding an info argument (the 3rd) of something like "certificate
report".

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:73:
std::string((char*)public_key, crypto::curve25519::kBytes));
sizeof(public_key)

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:97: std::string(),
std::string(), 0, 0,
the info would also have to be updated here.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:113:
decrypted_report->ParseFromString(plaintext);
this can fail, right? Just return it directly?

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.h:27: bool
DecryptCertificateErrorReport(
Could this be a static method on CertificateErrorReporter? It's a tad odd to
have it at the top-level of a namespace.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter_unittest.cc (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter_unittest.cc:114:
uploaded_request.ParseFromString(upload_data);
check return value?

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h
File crypto/aead_openssl.h (right):

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:15: extern const size_t kAeadKeyLength;
these would have to be properties of |Aead|.

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:22: explicit Aead(const std::string* key);
On 2015/04/18 05:24:42, estark wrote:
> Wasn't sure if this should take an algorithm parameter of some kind if only
one
> algorithm is supported.

It should, and the KeyLength and NonceLength() become non-static. See, for
example, hmac.h.

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:26: bool Seal(const base::StringPiece& plaintext,
addition data argument?

[estark, 2015-04-22 03:55:31]

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.cc (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:14: #if defined(USE_OPENSSL)
On 2015/04/21 20:28:25, agl wrote:
> On 2015/04/18 05:24:42, estark wrote:
> > Should I be checking OPENSSL_IS_BORINGSSL too? It seems to not work (is
never
> > defined).
> 
> No, OpenSSL in Chrome is always BoringSSL.

Acknowledged.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:54: std::string(),
std::string(), 0, 0,
On 2015/04/21 20:28:25, agl wrote:
> I'd suggest adding an info argument (the 3rd) of something like "certificate
> report".

Done.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:73:
std::string((char*)public_key, crypto::curve25519::kBytes));
On 2015/04/21 20:28:25, agl wrote:
> sizeof(public_key)

Done.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:97: std::string(),
std::string(), 0, 0,
On 2015/04/21 20:28:25, agl wrote:
> the info would also have to be updated here.

Done.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.cc:113:
decrypted_report->ParseFromString(plaintext);
On 2015/04/21 20:28:25, agl wrote:
> this can fail, right? Just return it directly?

Done.

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.h:27: bool
DecryptCertificateErrorReport(
On 2015/04/21 20:28:25, agl wrote:
> Could this be a static method on CertificateErrorReporter? It's a tad odd to
> have it at the top-level of a namespace.

Done. (Sorry, missed that in your first review.)

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter_unittest.cc (right):

https://codereview.chromium.org/1083493003/diff/220001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter_unittest.cc:114:
uploaded_request.ParseFromString(upload_data);
On 2015/04/21 20:28:25, agl wrote:
> check return value?

Done (and line 117)

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h
File crypto/aead_openssl.h (right):

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:15: extern const size_t kAeadKeyLength;
On 2015/04/21 20:28:26, agl wrote:
> these would have to be properties of |Aead|.

Oops, these are actually dead. Removed.

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:22: explicit Aead(const std::string* key);
On 2015/04/21 20:28:26, agl wrote:
> On 2015/04/18 05:24:42, estark wrote:
> > Wasn't sure if this should take an algorithm parameter of some kind if only
> one
> > algorithm is supported.
> 
> It should, and the KeyLength and NonceLength() become non-static. See, for
> example, hmac.h.

Done.

https://codereview.chromium.org/1083493003/diff/220001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:26: bool Seal(const base::StringPiece& plaintext,
On 2015/04/21 20:28:26, agl wrote:
> addition data argument?

Done.

https://codereview.chromium.org/1083493003/diff/260001/crypto/aead_openssl.h
File crypto/aead_openssl.h (right):

https://codereview.chromium.org/1083493003/diff/260001/crypto/aead_openssl.h#...
crypto/aead_openssl.h:11: struct evp_aead_st;
This seems odd to me (forward-declaring the internal name), not sure if it's
okay or not.

[estark, 2015-04-22 04:02:55]

Rebase, and reverted an accidentally-committed change I made while debugging
something unrelated

[estark, 2015-04-23 17:45:23]

agl, could you please take a look at this again when you have a chance?

[agl, 2015-04-23 20:09:22]

lgtm

https://codereview.chromium.org/1083493003/diff/320001/crypto/aead_openssl.cc
File crypto/aead_openssl.cc (right):

https://codereview.chromium.org/1083493003/diff/320001/crypto/aead_openssl.cc...
crypto/aead_openssl.cc:34: key_ = key;
This should be fixed up afterwards to do there key schedule here, but it's not
important: the API is correct.

[estark, 2015-04-23 22:26:35]

estark@chromium.org changed reviewers:
+ mattm@chromium.org

[estark, 2015-04-23 22:26:36]

mattm@chromium.org: could you please review chrome/browser/safe_browsing?

[mattm, 2015-04-23 22:59:14]

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
File chrome/browser/safe_browsing/ping_manager.cc (right):

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
chrome/browser/safe_browsing/ping_manager.cc:60: :
GURL(kExtendedReportingUploadUrlSecure),
This condition feels a little large/unwieldy to do inline. I guess it would
probably be more readable split into a separate statement.

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
chrome/browser/safe_browsing/ping_manager.cc:63: certificate_error_reporter_ =
nullptr;
unnecessary

[estark, 2015-04-23 23:53:47]

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
File chrome/browser/safe_browsing/ping_manager.cc (right):

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
chrome/browser/safe_browsing/ping_manager.cc:60: :
GURL(kExtendedReportingUploadUrlSecure),
On 2015/04/23 22:59:14, mattm wrote:
> This condition feels a little large/unwieldy to do inline. I guess it would
> probably be more readable split into a separate statement.

Done.

https://codereview.chromium.org/1083493003/diff/320001/chrome/browser/safe_br...
chrome/browser/safe_browsing/ping_manager.cc:63: certificate_error_reporter_ =
nullptr;
On 2015/04/23 22:59:14, mattm wrote:
> unnecessary

Done.

[mattm, 2015-04-24 00:58:03]

safe_browsing lgtm

[estark, 2015-04-24 03:20:22]

Thanks mattm and agl.

felt: do you want to take a quick look before I commit?

[felt, 2015-04-24 19:33:57]

I don't have anything to add to agl & mattm's reviews

https://codereview.chromium.org/1083493003/diff/380001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/380001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.h:80: // over SSL.
nit: seems like this comment ought to be over SendReport, which will DCHECK if
you try to send a report over http on an unsupported platform

[estark, 2015-04-24 20:10:05]

https://codereview.chromium.org/1083493003/diff/380001/chrome/browser/net/cer...
File chrome/browser/net/certificate_error_reporter.h (right):

https://codereview.chromium.org/1083493003/diff/380001/chrome/browser/net/cer...
chrome/browser/net/certificate_error_reporter.h:80: // over SSL.
On 2015/04/24 19:33:57, felt wrote:
> nit: seems like this comment ought to be over SendReport, which will DCHECK if
> you try to send a report over http on an unsupported platform

Done.

[estark, 2015-04-24 20:10:10]

The CQ bit was checked by estark@chromium.org

[estark, 2015-04-24 20:10:10]

The patchset sent to the CQ was uploaded after l-g-t-m from agl@chromium.org,
mattm@chromium.org
Link to the patchset: https://codereview.chromium.org/1083493003/#ps400001
(title: "tweak comments")

[commit-bot: I haz the power, 2015-04-24 20:10:35]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1083493003/400001

[commit-bot: I haz the power, 2015-04-24 21:07:57]

Committed patchset #20 (id:400001)

[commit-bot: I haz the power, 2015-04-24 21:08:44]

Patchset 20 (id:??) landed as
https://crrev.com/0a3351c2a7c81284f82e6531380a21d079f55056
Cr-Commit-Position: refs/heads/master@{#326876}

[Justin Donnelly, 2015-04-24 21:28:42]

A revert of this CL (patchset #20 id:400001) has been created in
https://codereview.chromium.org/1073913004/ by jdonnelly@chromium.org.

The reason for reverting is: This CL broke the Win8 GN bot:

http://build.chromium.org/p/chromium.win/builders/Win8%20GN/builds/6514.

[estark, 2015-04-25 01:03:26]

The CQ bit was checked by estark@chromium.org

[estark, 2015-04-25 01:03:27]

The patchset sent to the CQ was uploaded after l-g-t-m from agl@chromium.org,
mattm@chromium.org
Link to the patchset: https://codereview.chromium.org/1083493003/#ps420001
(title: "add aead files to BUILD.gn")

[commit-bot: I haz the power, 2015-04-25 01:03:47]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1083493003/420001

[commit-bot: I haz the power, 2015-04-25 04:52:35]

Committed patchset #21 (id:420001)

[commit-bot: I haz the power, 2015-04-25 04:54:31]

Patchset 21 (id:??) landed as
https://crrev.com/03206a1fcfe475604570d49dca0a885591e45ad7
Cr-Commit-Position: refs/heads/master@{#326957}