[webcrypto] Implement RSA-PSS using BoringSSL.

[webcrypto] Implement RSA-PSS using BoringSSL.

BUG=399090
Committed: https://crrev.com/8793ece2780573ef52e6adf163ccfe2b2d711dec
Cr-Commit-Position: refs/heads/master@{#300327}

[eroman, 2014-10-15 23:36:32]

Patchset #1 (id:1) has been deleted

[eroman, 2014-10-15 23:36:38]

Patchset #1 (id:20001) has been deleted

[eroman, 2014-10-15 23:44:05]

Patchset #1 (id:40001) has been deleted

[eroman, 2014-10-15 23:44:24]

eroman@chromium.org changed reviewers:
+ rsleevi@chromium.org

[Ryan Sleevi, 2014-10-17 21:08:13]

rsleevi@chromium.org changed reviewers:
+ davidben@chromium.org

[Ryan Sleevi, 2014-10-17 21:08:13]

rsleevi@chromium.org changed required reviewers:
+ davidben@chromium.org

[Ryan Sleevi, 2014-10-17 21:08:14]

David should review the BoringSSL bits. He was suggesting there was a bug in
BoringSSL when I was in Boston last week. Otherwise, LGTM

https://codereview.chromium.org/661653002/diff/50015/content/child/webcrypto/...
File content/child/webcrypto/openssl/rsa_pss_openssl.cc (right):

https://codereview.chromium.org/661653002/diff/50015/content/child/webcrypto/...
content/child/webcrypto/openssl/rsa_pss_openssl.cc:26: // TODO(eroman): Is this
right? Not enumerated in WebCrypto spec.
Where do you see that? It is -
https://dvcs.w3.org/hg/webcrypto-api/raw-file/tip/spec/Overview.html#rsa-pss

[eroman, 2014-10-17 22:53:38]

https://codereview.chromium.org/661653002/diff/50015/content/child/webcrypto/...
File content/child/webcrypto/openssl/rsa_pss_openssl.cc (right):

https://codereview.chromium.org/661653002/diff/50015/content/child/webcrypto/...
content/child/webcrypto/openssl/rsa_pss_openssl.cc:26: // TODO(eroman): Is this
right? Not enumerated in WebCrypto spec.
On 2014/10/17 21:08:14, Ryan Sleevi wrote:
> Where do you see that? It is -
> https://dvcs.w3.org/hg/webcrypto-api/raw-file/tip/spec/Overview.html#rsa-pss

Confirmed, thanks! (Removed the comment).

... My confusion was in consulting:
  "A.1. Algorithm mappings"

Which does not enumerate it. And then right below that there was an editor
comment asking whether "RSA-PSS with SHA-1" should be specified.

Reading the spec more carefully I see that A.1. Says it is non-normative, and to
defer to the algorithm-specific sections.
I notice now that the section clearly states it is non-normative.

[davidben, 2014-10-19 03:42:34]

lgtm with a couple of comments.

> David should review the BoringSSL bits. He was suggesting there was a bug in
BoringSSL when I was in Boston last week.

That was
https://boringssl.googlesource.com/boringssl/+/5db5d4e2b075f46ca80698a8baf168....
I believe it's been rolled into Chromium. (You would have noticed if it hadn't;
the tests would have completely failed to work.)

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
File content/child/webcrypto/openssl/rsa_sign_openssl.cc (right):

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
content/child/webcrypto/openssl/rsa_sign_openssl.cc:61:
salt_length_bytes_int.ValueOrDie())) {
These should all be != 1 rather than !. The random functions that back onto
EVP_PKEY_CTX_ctrl can return -1 or -2 to signal random things. (I wonder if we
can fix that because that's obnoxious...)

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
File content/child/webcrypto/test/rsa_pss_unittest.cc (right):

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
content/child/webcrypto/test/rsa_pss_unittest.cc:194: // Sign twice.
Nit: This comment seems off.

[eroman, 2014-10-20 18:44:36]

Patchset #4 (id:110001) has been deleted

[eroman, 2014-10-20 18:44:53]

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
File content/child/webcrypto/openssl/rsa_sign_openssl.cc (right):

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
content/child/webcrypto/openssl/rsa_sign_openssl.cc:61:
salt_length_bytes_int.ValueOrDie())) {
On 2014/10/19 03:42:34, David Benjamin wrote:
> These should all be != 1 rather than !. The random functions that back onto
> EVP_PKEY_CTX_ctrl can return -1 or -2 to signal random things. (I wonder if we
> can fix that because that's obnoxious...)

Done. Thanks for spotting this!

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
File content/child/webcrypto/test/rsa_pss_unittest.cc (right):

https://codereview.chromium.org/661653002/diff/90001/content/child/webcrypto/...
content/child/webcrypto/test/rsa_pss_unittest.cc:194: // Sign twice.
On 2014/10/19 03:42:34, David Benjamin wrote:
> Nit: This comment seems off.

Done. (Copy paste error).

[eroman, 2014-10-20 18:45:09]

The CQ bit was checked by eroman@chromium.org

[commit-bot: I haz the power, 2014-10-20 18:46:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/661653002/130001

[eroman, 2014-10-20 18:48:16]

The CQ bit was unchecked by eroman@chromium.org

[eroman, 2014-10-20 19:03:40]

The CQ bit was checked by eroman@chromium.org

[commit-bot: I haz the power, 2014-10-20 19:04:46]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/661653002/170001

[commit-bot: I haz the power, 2014-10-20 20:47:24]

Committed patchset #6 (id:170001)

[commit-bot: I haz the power, 2014-10-20 20:48:41]

Patchset 6 (id:??) landed as
https://crrev.com/8793ece2780573ef52e6adf163ccfe2b2d711dec
Cr-Commit-Position: refs/heads/master@{#300327}