The existing nghttp2_session_consume() affects both connection and
stream level flow control windows. The new functions only affects
either connection or stream. There is some interesting use cases.
For example, we may want to pause a stream by not sending
WINDOW_UPDATE, meanwhile we want to continue to process other streams.
In this case, we use nghttp2_session_consume_connection() to tell
library that only connection level window is recovered. The relevant
discussion: https://code.google.com/p/chromium/issues/detail?id=473259
This is same issue described in https://github.com/h2o/h2o/issues/268.
That is if SSL object has decrypted data buffered inside it, and
application does not read it for some reason (e.g., rate limit), we
have to check the existence of data using SSL_pending. This is
because buffered data inside SSL is not notified by io watcher. It is
obvious, but we totally missed it.
nghttpx code normally reads everything until SSL_read returns error
(want-read). But if rate limit is involved, we stop reading early.
Also in HTTP/1 code, while processing one request, we just read until
buffer is filled up. In these cases, we may suffer from this problem.
This commit fixes this problem, by performing SSL_pending() and if it
has buffered data and read io watcher is enabled, we feed event using
ev_feed_event().
To avoid buffer copy in nghttp2_data_source_read_callback, this commit
introduces NGHTTP2_DATA_FLAG_NO_COPY and nghttp2_send_data_callback.
By using NGHTTP2_DATA_FLAG_NO_COPY in
nghttp2_data_source_read_callback, application can avoid to copy
application data to given buffer. Instead, application has to
implement nghttp2_send_data_callback to send complete DATA frame by
itself. We see noticeable performance increase in nghttpd and
tiny-nghttpd using this new feature. On the other hand, nghttpx does
not show such difference, probably because buffer copy is not
bottleneck. Using nghttp2_send_data_callback adds complexity, so it
is recommended to measure the performance to see whether this extra
complexity worth it.
Previously API reference is gigantic one rst file and it is a bit hard
to use, especially when browsing similar functions. This commit
splits API reference into smaller fine grained files. The macros,
enums, types are now in its own file. Each API function has its own
file now. API reference doc is now index to above documentation
files. The apiref-header.rst is renamed as programmers-guide.rst and
becomes standalone document.
- Tell Vim this is a Dockerfile to enable syntax highlight;
- Explicity use Ubuntu "Trusty";
- Remove downloaded file to save space;
- Chain up some RUN commands to generate fewer layers.
Tatsuhiro Tsujikawa
Remo E
Alexis La Goutte
Zhuoyun Wei
Dingjun