Change WeakPtr<SpdyStream> to raw pointer in SpdyHttpStream.

Change WeakPtr<SpdyStream> to raw pointer in SpdyHttpStream.

This allows for more explicit expression of lifetime expectations.

SpdyStream instances are created in SpdySession::CreateStream() and in
SpdySession::TryCreatePushStream(), and are owned by SpdySession in its
|created_streams_| and |active_streams_| members.  They are destroyed in
SpdySession::CloseActiveStreamIterator() and in
SpdySession::CloseCreatedStreamIterator(),
and SpdySession::DeleteStream() is always called by these methods.
This always calls SpdyStream::OnClose() before deleting the stream,
which in turn calls SpdyStream::Delegate::OnClose().

Every time SpdyHttpStream grabs a handle of a SpdyStream, it immediately
registers itself as a delegate.  That means that every time a SpdyStream
instance is destroyed, if a SpdyHttpStream instance has a handle of it,
it will get notified.  That is, SpdyHttpStream does not actually need to
keep a WeakPtr<SpdyStream>: a raw pointer suffices.

Review-Url: https://codereview.chromium.org/2642133002
Cr-Commit-Position: refs/heads/master@{#445302}
Committed: https://chromium.googlesource.com/chromium/src/+/4e6e774857ae99aa9dc13287e0c36c1e69bf92b6

[Bence, 2017-01-19 16:02:48]

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

[commit-bot: I haz the power, 2017-01-19 16:03:31]

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, 2017-01-19 16:55:16]

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

[commit-bot: I haz the power, 2017-01-19 16:55:16]

Dry run: This issue passed the CQ dry run.

[Bence, 2017-01-19 17:33:36]

bnc@chromium.org changed reviewers:
+ xunjieli@chromium.org

[Bence, 2017-01-19 17:33:36]

xunjieli: PTAL.  Thank you.

[xunjieli, 2017-01-20 02:57:20]

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_http_stream.h
File net/spdy/spdy_http_stream.h (right):

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_http_stream.h...
net/spdy/spdy_http_stream.h:131: SpdyStream* stream_;
Can you add a comment here on who owns |stream_| and when the pointer is
guaranteed to be valid (e.g. after InitializeStream() and before
Delegate::OnClose()).

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc
File net/spdy/spdy_session.cc (right):

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc#ne...
net/spdy/spdy_session.cc:841: if (availability_state_ == STATE_DRAINING)
I think we need to keep the original logic to reset *stream to null before this
line. If we return early without setting |*stream|, the pointer can be in an
uninitialized state.

*stream = nullptr;

[Bence, 2017-01-20 15:53:13]

Thanks.  PTAL.

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_http_stream.h
File net/spdy/spdy_http_stream.h (right):

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_http_stream.h...
net/spdy/spdy_http_stream.h:131: SpdyStream* stream_;
On 2017/01/20 02:57:20, xunjieli wrote:
> Can you add a comment here on who owns |stream_| and when the pointer is
> guaranteed to be valid (e.g. after InitializeStream() and before
> Delegate::OnClose()).

Done.  Also added a comment for |stream_closed_| once we are here.

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc
File net/spdy/spdy_session.cc (right):

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc#ne...
net/spdy/spdy_session.cc:841: if (availability_state_ == STATE_DRAINING)
On 2017/01/20 02:57:20, xunjieli wrote:
> I think we need to keep the original logic to reset *stream to null before
this
> line. If we return early without setting |*stream|, the pointer can be in an
> uninitialized state.
> 
> *stream = nullptr;

Done.  I moved it to the if branch.  BTW the DCHECK in
SpdyHttpStream::InitializeStream() (the only non-test caller) guarantees that
|*stream| is already nullptr (or a null weakptr before this CL), but surely this
does not hurt.

[xunjieli, 2017-01-20 15:57:23]

lgtm!

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc
File net/spdy/spdy_session.cc (right):

https://codereview.chromium.org/2642133002/diff/1/net/spdy/spdy_session.cc#ne...
net/spdy/spdy_session.cc:841: if (availability_state_ == STATE_DRAINING)
On 2017/01/20 15:53:13, Bence wrote:
> On 2017/01/20 02:57:20, xunjieli wrote:
> > I think we need to keep the original logic to reset *stream to null before
> this
> > line. If we return early without setting |*stream|, the pointer can be in an
> > uninitialized state.
> > 
> > *stream = nullptr;
> 
> Done.  I moved it to the if branch.  BTW the DCHECK in
> SpdyHttpStream::InitializeStream() (the only non-test caller) guarantees that
> |*stream| is already nullptr (or a null weakptr before this CL), but surely
this
> does not hurt.

Acknowledged.

[Bence, 2017-01-20 15:59:40]

Thanks.  Now I plan to wait for branch point before landing this in case
something terrible happens.

[Bence, 2017-01-21 10:55:02]

The CQ bit was checked by bnc@chromium.org

[commit-bot: I haz the power, 2017-01-21 10:55:12]

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, 2017-01-21 12:47:29]

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

[commit-bot: I haz the power, 2017-01-21 12:47:30]

Try jobs failed on following builders:
  linux_chromium_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[Bence, 2017-01-22 11:46:53]

The CQ bit was checked by bnc@chromium.org

[commit-bot: I haz the power, 2017-01-22 11:46:56]

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, 2017-01-22 12:59:46]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1485085613416480,
"parent_rev": "dfd9bf1edeca16610876ba4e33150401143b1924", "commit_rev":
"4e6e774857ae99aa9dc13287e0c36c1e69bf92b6"}

[commit-bot: I haz the power, 2017-01-22 13:00:24]

Description was changed from

==========
Change WeakPtr<SpdyStream> to raw pointer in SpdyHttpStream.

This allows for more explicit expression of lifetime expectations.

SpdyStream instances are created in SpdySession::CreateStream() and in
SpdySession::TryCreatePushStream(), and are owned by SpdySession in its
|created_streams_| and |active_streams_| members.  They are destroyed in
SpdySession::CloseActiveStreamIterator() and in
SpdySession::CloseCreatedStreamIterator(),
and SpdySession::DeleteStream() is always called by these methods.
This always calls SpdyStream::OnClose() before deleting the stream,
which in turn calls SpdyStream::Delegate::OnClose().

Every time SpdyHttpStream grabs a handle of a SpdyStream, it immediately
registers itself as a delegate.  That means that every time a SpdyStream
instance is destroyed, if a SpdyHttpStream instance has a handle of it,
it will get notified.  That is, SpdyHttpStream does not actually need to
keep a WeakPtr<SpdyStream>: a raw pointer suffices.
==========

to

==========
Change WeakPtr<SpdyStream> to raw pointer in SpdyHttpStream.

This allows for more explicit expression of lifetime expectations.

SpdyStream instances are created in SpdySession::CreateStream() and in
SpdySession::TryCreatePushStream(), and are owned by SpdySession in its
|created_streams_| and |active_streams_| members.  They are destroyed in
SpdySession::CloseActiveStreamIterator() and in
SpdySession::CloseCreatedStreamIterator(),
and SpdySession::DeleteStream() is always called by these methods.
This always calls SpdyStream::OnClose() before deleting the stream,
which in turn calls SpdyStream::Delegate::OnClose().

Every time SpdyHttpStream grabs a handle of a SpdyStream, it immediately
registers itself as a delegate.  That means that every time a SpdyStream
instance is destroyed, if a SpdyHttpStream instance has a handle of it,
it will get notified.  That is, SpdyHttpStream does not actually need to
keep a WeakPtr<SpdyStream>: a raw pointer suffices.

Review-Url: https://codereview.chromium.org/2642133002
Cr-Commit-Position: refs/heads/master@{#445302}
Committed:
https://chromium.googlesource.com/chromium/src/+/4e6e774857ae99aa9dc13287e0c3...
==========

[commit-bot: I haz the power, 2017-01-22 13:00:25]

Committed patchset #2 (id:20001) as
https://chromium.googlesource.com/chromium/src/+/4e6e774857ae99aa9dc13287e0c3...