quic: extract transport logic from Application to Session#64127
Open
pimterry wants to merge 1 commit into
Open
quic: extract transport logic from Application to Session#64127pimterry wants to merge 1 commit into
pimterry wants to merge 1 commit into
Conversation
Collaborator
|
Review requested:
|
nodejs-github-bot
added
c++
Two small fixes notably included: - Fixed a SendPendingData but where datagrams would stall on an idle connection in the nwrite == 0 branch. - Check is_destroyed() after StreamCommit, since it calls JS callbacks which could destroy the session. Signed-off-by: Tim Perry <pimterry@gmail.com>
pimterry
force-pushed
the
extract-h3-refactor
branch
from
19197f6 to
5fdb7e8
Compare
jasnell
reviewed
Jul 5, 2026
| // A block of pending outbound stream data, passed between the application | ||
| // layer (which fills it via GetStreamData) and the send pump (which hands | ||
| // it to ngtcp2_conn_writev_stream and commits the accepted length). | ||
| struct StreamData final { |
Member
There was a problem hiding this comment.
If at all possible, I think this should likely be moved into class Session (e.g. Session::StreamData). It makes more sense as a child class than it's own top-level.
jasnell
reviewed
Jul 5, 2026
| // The stream identifier. If this is a negative value then no stream is | ||
| // identified. | ||
| stream_id id = -1; | ||
| int fin = 0; |
Member
There was a problem hiding this comment.
If we're changing things anyway, making this a bool would probably make sense.
jasnell
reviewed
Jul 5, 2026
| stream_id id = -1; | ||
| int fin = 0; | ||
| ngtcp2_vec data[kMaxVectorCount]{}; | ||
| BaseObjectPtr<Stream> stream; |
Member
There was a problem hiding this comment.
I wonder if we can drop id and just rely on stream.
jasnell
reviewed
Jul 5, 2026
Comment on lines
+449
to
+450
| // If < 0 is returned, either NGTCP2_ERR_WRITE_MORE or a fatal error is | ||
| // returned; the caller must check. If > 0 is returned, the packet is done |
Member
There was a problem hiding this comment.
Suggested change
| // If < 0 is returned, either NGTCP2_ERR_WRITE_MORE or a fatal error is | |
| // returned; the caller must check. If > 0 is returned, the packet is done | |
| // If a negative value is returned, it is either NGTCP2_ERR_WRITE_MORE or | |
| // fatal error; the caller must check. If > 0 is returned, the packet is done |
jasnell
reviewed
Jul 5, 2026
| // the session, encapsulating ngtcp2 transport details here. | ||
|
|
||
| // Open a unidirectional stream, setting *id on success, or returning false | ||
| bool OpenUni(stream_id* id); |
Member
There was a problem hiding this comment.
Prefer to avoid abbreviating. s/OpenUni/OpenUnidirectionalStream
jasnell
reviewed
Jul 5, 2026
|
|
||
| void ExtendMaxStreams(Direction direction, uint64_t max); | ||
|
|
||
| // Signal that we've consumed $len bytes on stream $id to update flow control |
Member
There was a problem hiding this comment.
Suggested change
| // Signal that we've consumed $len bytes on stream $id to update flow control | |
| // Signal that we've consumed `len` bytes on stream $id to update flow control |
jasnell
reviewed
Jul 5, 2026
| PopPendingDatagram(); | ||
| } else { | ||
| Debug(this, "Datagram %" PRIu64 " not accepted into packet", dg.id); | ||
| } |
Member
There was a problem hiding this comment.
Hmm... I think this has a bug. If dg_nwrite is positive, it appears to fall-through to the default: case below?
Member
There was a problem hiding this comment.
(To be fair, I don't think it's a new bug ;-) ...)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Extracting out some standalone refactoring from #63995 for easier review while we keep debating the API. I think this is a helpful refactoring generally, but especially in a world where we're restructuring
Application.This shrinks the application layer by moving generic QUIC logic that will be used by all implementations into the Session itself, and makes the Session act as a clean interface over ngtcp2 - so that the Application exclusively uses Session methods for all connection manipulation, instead of managing ngtcp2 directly. This gives a cleaner split between the low-level QUIC behaviour and the per-application-protocol layer.
There's two small fixes notably included in the migration, contained within the moved code: