[net/dns] Reland of 218616 (Simultaneous A/AAAA queries).

Reland of 218616.  It was reverted in 219027 due to crashes in the case of
responses with no addresses.

> [net/dns] Perform A/AAAA queries for AF_UNSPEC resolutions in parallel. 
> 
> The second DnsTransaction is scheduled as a second job on the resolver's
> PrioritizedDispatcher, at the beginning of its priority queue. The two
> DnsTransactions run independently of each other, although if one of them
> finishes with an error, the other one will be scrapped immediately, and
> the Job will fall back to ProcTask. 
> 
> BUG=174992
> 
> Original Review URL: https://chromiumcodereview.appspot.com/19498003
Revert URL: https://codereview.chromium.org/23102009

BUG=174992, 277625
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=221469

[mmenke, 2013-08-22 18:20:39]

Thanks for pointing out where the problem was, made things simpler.

Was able to get a crash in roughly the same spot by forcing all DNS transactions
to be AAAA (Since the unit test doesn't go far enough to crash in the same way,
wanted to double check this fixes things.  One top crasher a week is enough for
me).

[mmenke, 2013-08-28 19:44:31]

Fixed the other crasher.  Looked through the other browser crashes, and all
net:: crashes looked to be due to one of those two bugs, or unrelated.

Patch set 1 is what I originally landed, except I forgot to include
dns_probe_test_util.cc.

PTAL when you get a chance.

[szym, 2013-08-30 19:25:41]

Excellent work.
LGTM

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
File net/base/prioritized_dispatcher_unittest.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
net/base/prioritized_dispatcher_unittest.cc:444: TestJob* job_e =
AddJobAtHead('e', MEDIUM);
nit: To test one thing at a time, I'd suggest using AddJob() so that the order
is "ab..cd.e."

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
net/base/prioritized_dispatcher_unittest.cc:509:
TEST_F(PrioritizedDispatcherTest, ZeroLimitsThenIncreatePriority) {
nit: Increase

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
File net/dns/host_resolver_impl.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl.cc:2320: // DnsTasks, since do so may start new jobs.
nit: doing

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
File net/dns/host_resolver_impl_unittest.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1295: AddDnsRule("slow_ok",
dns_protocol::kTypeA, MockDnsClientRule::FAIL, true);
suggest slow_nx

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1947: // trying different limits is
important for this test:  Different limits
nit-picky nit: Trying

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1964: host += i;
nit: std::string host = "nx" + i; or host = base::StringPrintf("nx%i, i);

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1977: proc_->SignalMultiple(19u);
maximum_dns_failures() + 3

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1984: EXPECT_EQ(OK,
requests_[16]->WaitForResult());
maximum_dns_failures()
and below:
+ 1
+ 2

[mmenke, 2013-08-30 19:52:25]

Thanks for the feedback!

I think I'll land Wednesday or Thursday, since I'm taking two vacation days next
week, and don't want to land this CL when I won't be around to clean up if
things go wrong.

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
File net/base/prioritized_dispatcher_unittest.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
net/base/prioritized_dispatcher_unittest.cc:444: TestJob* job_e =
AddJobAtHead('e', MEDIUM);
On 2013/08/30 19:25:41, szym wrote:
> nit: To test one thing at a time, I'd suggest using AddJob() so that the order
> is "ab..cd.e."

Done.  I did it this way because I noticed we never did an AddJobAtHead() call
on an empty dispatcher, but I think you're right.

https://codereview.chromium.org/22909037/diff/110001/net/base/prioritized_dis...
net/base/prioritized_dispatcher_unittest.cc:509:
TEST_F(PrioritizedDispatcherTest, ZeroLimitsThenIncreatePriority) {
On 2013/08/30 19:25:41, szym wrote:
> nit: Increase

Done.

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
File net/dns/host_resolver_impl.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl.cc:2320: // DnsTasks, since do so may start new jobs.
On 2013/08/30 19:25:41, szym wrote:
> nit: doing

Done.

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
File net/dns/host_resolver_impl_unittest.cc (right):

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1295: AddDnsRule("slow_ok",
dns_protocol::kTypeA, MockDnsClientRule::FAIL, true);
On 2013/08/30 19:25:41, szym wrote:
> suggest slow_nx

Done.  Changed what it did, but forgot to rename it.  :)

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1947: // trying different limits is
important for this test:  Different limits
On 2013/08/30 19:25:41, szym wrote:
> nit-picky nit: Trying

Done.

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1964: host += i;
On 2013/08/30 19:25:41, szym wrote:
> nit: std::string host = "nx" + i; or host = base::StringPrintf("nx%i, i);

Done.  Don't think the first works, since it would be adding i to a string,
which would just increment the pointer, so went with the second.

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1977: proc_->SignalMultiple(19u);
On 2013/08/30 19:25:41, szym wrote:
> maximum_dns_failures() + 3

Done.

https://codereview.chromium.org/22909037/diff/110001/net/dns/host_resolver_im...
net/dns/host_resolver_impl_unittest.cc:1984: EXPECT_EQ(OK,
requests_[16]->WaitForResult());
On 2013/08/30 19:25:41, szym wrote:
> maximum_dns_failures()
> and below:
> + 1
> + 2

Done.

[commit-bot: I haz the power, 2013-09-05 15:35:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/22909037/138001

[commit-bot: I haz the power, 2013-09-05 18:26:54]

Change committed as 221469