Do not pass CRYPT_NOHASHOID to CryptSignHash with CALG_SSL3_SHAMD5....

Do not pass CRYPT_NOHASHOID to CryptSignHash with CALG_SSL3_SHAMD5.
This may fix bug 72041.

R=wtc
BUG=72041
TEST=covered by existing net_unittests

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=74254

[wtc, 2011-02-09 02:18:23]

I did a lot of web search last night.  It seems that we
can just pass 0 as dwFlags to CryptSignHash for
CALG_SSL3_SHAMD5.

[Ryan Sleevi, 2011-02-09 02:20:44]

What about the DSS_SIGN / ECDSA case, which uses CALG_SHA1? See also
doDerEncode, which is what 0 for dwFlags should be doing.

That's why I haven't uploaded a similar CL, because I haven't had a chance to
dig into the CALG_SHA1 case.

[wtc, 2011-02-09 02:32:41]

The CRYPT_NOHASHOID flag applies to PKCS #1 RSA signatures only.
DSA and ECDSA signatures don't contain the DigestInfo structure.

[Ryan Sleevi, 2011-02-09 02:33:11]

On 2011/02/09 02:20:44, Ryan Sleevi wrote:
> What about the DSS_SIGN / ECDSA case, which uses CALG_SHA1? See also
> doDerEncode, which is what 0 for dwFlags should be doing.
> 
> That's why I haven't uploaded a similar CL, because I haven't had a chance to
> dig into the CALG_SHA1 case.

Err, perhaps this is OK. CRYPT_NOHASHOID is documented as only being used with
RSA providers ( http://msdn.microsoft.com/en-us/library/aa380280(VS.85).aspx ) ,
which makes sense, since the OID is a function of the RSA/PKCS#1 signing.

Since I definitely observed SChannel passing 0 for dwFlags for RSA keys, I'd say
LGTM.

[wtc, 2011-02-09 02:37:25]

Please feel free to commit this CL so you get
the credit of finding this solution.  Thanks.

[Ryan Sleevi, 2011-02-09 02:38:46]

On 2011/02/09 02:37:25, wtc wrote:
> Please feel free to commit this CL so you get
> the credit of finding this solution.  Thanks.

Credit doesn't matter to me :-) What matters is 72041 gets fixed. I flagged
commit-bot to land it.

[Ryan Sleevi, 2011-02-09 03:36:42]

On 2011/02/09 02:38:46, Ryan Sleevi wrote:
> On 2011/02/09 02:37:25, wtc wrote:
> > Please feel free to commit this CL so you get
> > the credit of finding this solution.  Thanks.
> 
> Credit doesn't matter to me :-) What matters is 72041 gets fixed. I flagged
> commit-bot to land it.

I saw you submitted try jobs as well, so I wasn't sure if you were planning on
landing it. commit-bot should land it once the try jobs complete (
http://build.chromium.org/p/tryserver.chromium/builders/linux/builds/11759
http://build.chromium.org/p/tryserver.chromium/builders/mac/builds/12173
http://build.chromium.org/p/tryserver.chromium/builders/win/builds/14806 )

[commit-bot: I haz the power, 2011-02-09 03:47:11]

Commit queue patch verification failed

[commit-bot: I haz the power, 2011-02-09 03:54:40]

Commit queue patch verification failed

[commit-bot: I haz the power, 2011-02-09 05:36:21]

Commit queue patch verification failed