Add nghttp2_option_set_server_fallback_rfc7540_priorities. If it is
set to nonzero, and server submits SETTINGS_NO_RFC7540_PRIORITIES = 1,
but it does not receive SETTINGS_NO_RFC7540_PRIORITIES from client,
server falls back to RFC 7540 priorities. Only minimal set of
features are enabled in this fallback case.
This commit adds PRIORITY_UPDATE frame support. Applying incoming
PRIORITY_UPDATE frame to server push stream is not implemented.
Client can send PRIORITY_UPDATE frame by calling
nghttp2_submit_priority_update.
Server opts to receive PRIORITY_UPDATE frame by the call
nghttp2_option_set_builtin_recv_extension_type(option,
NGHTTP2_PRIORITY_UPDATE), and passing the option to
nghttp2_session_server_new2 or nghttp2_session_server_new3.
This commit implements RFC 9218 extensible prioritization scheme. It
is enabled when a local endpoint submits
SETTINGS_NO_RFC7540_PRIORITIES = 1. This commit only handles priority
signal in HTTP request header field. Priority header field in
PUSH_PROMISE is not supported.
HTTP messaging must be enabled to take advantage of this
prioritization scheme because HTTP fields are not parsed if HTTP
messaging is disabled.
Previously, if automatic window update is enabled (which is default),
after window size is set to 0 by
nghttp2_session_set_local_window_size, once the receiving window is
exhausted, even after window size is increased by
nghttp2_session_set_local_window_size, no more data cannot be
received. This is because nghttp2_session_set_local_window_size does
not submit WINDOW_UPDATE. It is only triggered when new data arrives
but since window is filled up, no more data cannot be received, thus
dead lock happens.
This commit fixes this issue. nghttp2_session_set_local_window_size
submits WINDOW_UPDATE if necessary.
https://github.com/curl/curl/issues/4939
The maximum number of outgoing concurrent streams is initially
limited to 100 to avoid issues when the local endpoint submits
lots of requests before receiving initial SETTINGS frame from
the remote endpoint, since sending them at once to the remote
endpoint could lead to rejection of some of the requests.
This initial limit is overwritten with the value advertised in
SETTINGS_MAX_CONCURRENT_STREAMS setting by the remote endpoint,
but previously, it wasn't lifted if the remote endpoint didn't
advertise that setting (implying no limits), in which case the
limit of 100 was retained, even though it was never advertised
by the remote endpoint.
Signed-off-by: Piotr Sikora <piotrsikora@google.com>
nghttp2_option_no_closed_streams controls whether closed streams are
retained or not. If nonzero is passed to that function's parameter
val, a session does not retain closed streams. It may hurt the shape
of priority tree, but can save memory.
Previously, in server side, we used closed streams to detect the error
that the misbehaving client sends a frame on the incoming stream it
explicitly closed. With this commit, we make a further step, and
detect one more error case. Since we retain closed streams as long as
the sum of its size and the number of opened streams are equal or less
than max concurrent streams, we can safely say that if we get a frame
which is sent on the stream that is not found in either closed or
opened stream, it is already closed or has not existed. Then we can
send GOAWAY.
The previous code shrinks closed streams when we closed another
stream, but now it is removed. It is enough to adjust closed streams
when new incoming stream is created.
While creating this commit, we noticed that
NGHTTP2_INITIAL_MAX_CONCURRENT_STREAMS is defined as INT32_MAX. But
since SETTINGS can contain value up to UINT32_MAX, it is not enough.
However, since the stream ID space is limited to INT32_MAX, it is high
enough. We could keep this value, but this time we deprecate
NGHTTP2_INITIAL_MAX_CONCURRENT_STREAMS macro. While it is in public
header, the effect of deprecating it is negligible because of the
reason we wrote above, and usually application sets much smaller value
(say, 100) as SETTINGS_MAX_CONCURRENT_STREAMS.
This function sets the maximum length of header block (a set of header
fields per HEADERS frame) to send. The length of given set of header
fields is calculated using nghttp2_hd_deflate_bound(). Previously,
this is hard-coded, and is 64KiB.
This option prevents the nghttp2 library from sending PING frame with
ACK flag set in the reply to incoming PING frame. To allow the
application to send PING with ACK flag set, nghttp2_submit_ping() now
recognizes NGHTTP2_FLAG_PING in its flags parameter.
Previously, we use session->next_stream_id to detect that given stream
ID was idle or not. But this was suboptimal, since it was updated
when stream ID was assigned, and it did not necessarily mean that it
actually has been sent to the peer. Now we introduced
session->sent_stream_id, which only updated when HEADERS/PUSH_PROMISE
has sent. Using sent_stream_id instead of next_stream_id tightens
idle stream detection, and misbehaved peer which sends frame with
stream ID that has not been generated.
This commit also overhauls test code which involves opening streams.
Now we have some wrapper functions for nghttp2_session_open_stream()
which also take care of updating next_stream_id and
last_recv_stream_id. They are crucial for some tests.
Previously, stream object for pushed resource was not created during
nghttp2_submit_push_promise(). It was created just before
nghttp2_before_frame_send_callback was called for that PUSH_PROMISE
frame. This means that application could not call
nghttp2_submit_response for the pushed resource before
nghttp2_before_frame_send_callback was called. This could be solved
by callback chaining, but for web server with back pressure from
backend stream, it is a bit unnecessarily hard to use.
This commit changes nghttp2_submit_push_promise() behaviour so that
stream object is created during that call. It makes application call
nghttp2_submit_response right after successful
nghttp2_submit_push_promise call.
Previously, nghttp2_session_end_request_headers_received assumes
stream is still writable (in other words, local endpoint has not sent
END_STREAM). But this assumption is false, because application can
send response in nghttp2_on_begin_frame_callback. Probably, this
assumption was made before the callback was introduced. This commit
addresses this issue. Since all
nghttp2_session_end_*_headers_received functions are identical, we
refactored them into one function.
This will improve performance since we can avoid indirect call of
internal functions. The downside is we now require libnghttp2 static
library to run unit tests.
Application can utilize this framework to send/receive user defined
extension frames. These frames are expected not to change existing
protocol behaviour.
This change adds new return error code from nghttp2_session_mem_recv
and nghttp2_session_recv functions, namely NGHTTP2_ERR_FLOODED. It is
fatal error, and is returned when flooding was detected.
RFC 7540 does not enforce any limit on the number of incoming reserved
streams (in RFC 7540 terms, streams in reserved (remote) state). This
only affects client side, since only server can push streams.
Malicious server can push arbitrary number of streams, and make
client's memory exhausted. The new option,
nghttp2_set_max_reserved_remote_streams, can set the maximum number of
such incoming streams to avoid possible memory exhaustion. If this
option is set, and pushed streams are automatically closed on
reception, without calling user provided callback, if they exceed the
given limit. The default value is 200. If session is configured as
server side, this option has no effect. Server can control the number
of streams to push.
We now use priority queue per stream, which contains the stream which
has ready to send a frame, or one of its descendants have a frame to
send. We maintain invariant that if a stream is queued, then its
ancestors are also queued (except for root). When we re-schedule
stream after transmission, we re-schedule all ancestors, so that
streams on the other path can get a chance to send. This is basically
the same mechanism h2o project uses, but there are differences in the
details.
Previously, the number of stream in one dependency tree (not including
root) is limited to 120. This is due to the fact that we use
recursive calls to traverse trees. Now we replaced recursive calls
with loop, we can remove this limitation. Also now all streams are
descendant of root stream, rather than linked list of individual
subtree root.
Tatsuhiro Tsujikawa
Dimitris Apostolou
James M Snell
Piotr Sikora