Fix QUIC crash when ProofSource::GetProof fails

Fix QUIC crash when ProofSource::GetProof fails

This CL fixes a crash in QuicCryptoServerConfig which occurs when
ProofSource::GetProof fails.  The situation occurs in
QuicCryptoServerConfig::EvaluateClientHello, when async codepaths are
disabled.  If the call to proof_source_->GetProof returns false, then we
set get_proof_failed to true and invoke
EvaluateClientHelloAfterGetProof.  However, that method (and
ValidateExpectedLeafCertificate, which it calls) assume that
signed_config->chain is populated, which may not be the case if GetProof
failed.

Merge internal change: 143447261

R=jri@chromium.org
BUG=

[danzh1, 2017-01-09 18:06:15]

The CQ bit was checked by danzh@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-01-09 18:06:40]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-01-09 19:34:01]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-01-09 19:34:02]

Dry run: This issue passed the CQ dry run.

[danzh1, 2017-01-09 20:10:43]

Description was changed from

==========
Fix QUIC crash when ProofSource::GetProof fails

This CL fixes a crash in QuicCryptoServerConfig which occurs when
ProofSource::GetProof fails.  The situation occurs in
QuicCryptoServerConfig::EvaluateClientHello, when async codepaths are
disabled.  If the call to proof_source_->GetProof returns false, then we
set get_proof_failed to true and invoke
EvaluateClientHelloAfterGetProof.  However, that method (and
ValidateExpectedLeafCertificate, which it calls) assume that
signed_config->chain is populated, which may not be the case if GetProof
failed.

Merge internal change: 143447261

R=rch@chromium.org
BUG=
==========

to

==========
Fix QUIC crash when ProofSource::GetProof fails

This CL fixes a crash in QuicCryptoServerConfig which occurs when
ProofSource::GetProof fails.  The situation occurs in
QuicCryptoServerConfig::EvaluateClientHello, when async codepaths are
disabled.  If the call to proof_source_->GetProof returns false, then we
set get_proof_failed to true and invoke
EvaluateClientHelloAfterGetProof.  However, that method (and
ValidateExpectedLeafCertificate, which it calls) assume that
signed_config->chain is populated, which may not be the case if GetProof
failed.

Merge internal change: 143447261

R=jri@chromium.org
BUG=
==========

[danzh1, 2017-01-09 20:10:43]

danzh@chromium.org changed reviewers:
+ jri@chromium.org
- rch@chromium.org

[Jana, 2017-01-09 22:08:08]

Thanks, Dan -- one question

https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
File net/quic/test_tools/failing_proof_source.cc (right):

https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
net/quic/test_tools/failing_proof_source.cc:19: return false;
Can you not just do this in the .h file as in the internal CL? Does compile fail
on some platform?

[danzh1, 2017-01-09 22:20:42]

https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
File net/quic/test_tools/failing_proof_source.cc (right):

https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
net/quic/test_tools/failing_proof_source.cc:19: return false;
On 2017/01/09 22:08:08, Jana wrote:
> Can you not just do this in the .h file as in the internal CL? Does compile
fail
> on some platform?

Yes, chromium compiler failed in my local repo. And it doesn't allow non-empty
method body in .h file.

[Jana, 2017-01-10 00:04:22]

On 2017/01/09 22:20:42, danzh1 wrote:
>
https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
> File net/quic/test_tools/failing_proof_source.cc (right):
> 
>
https://codereview.chromium.org/2626443002/diff/1/net/quic/test_tools/failing...
> net/quic/test_tools/failing_proof_source.cc:19: return false;
> On 2017/01/09 22:08:08, Jana wrote:
> > Can you not just do this in the .h file as in the internal CL? Does compile
> fail
> > on some platform?
> 
> Yes, chromium compiler failed in my local repo. And it doesn't allow non-empty
> method body in .h file.

Gotcha. Can you make sure to backport this change to the internal code?

[Jana, 2017-01-10 00:04:29]

lgtm