[WIP] benchmark: add basic QUIC benchmark #182

PR-URL: #145 Reviewed-By: James M Snell <[email protected]>

PR-URL: #150 Reviewed-By: James M Snell <[email protected]>

Allow using the handle more directly for I/O in other parts of the codebase. PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

Co-authored-by: Anna Henningsen <[email protected]> Co-authored-by: Daniel Bevenius <[email protected]> Co-authored-by: gengjiawen <[email protected]> Co-authored-by: James M Snell <[email protected]> Co-authored-by: Lucas Pardue <[email protected]> Co-authored-by: Ouyang Yadong <[email protected]> Co-authored-by: Juan José Arboleda <[email protected]> Co-authored-by: Trivikram Kamat <[email protected]>

Ported from tatsuhiro-t/openssl@920a331 PR-URL: #6 Reviewed-By: Daniel Bevenius <[email protected]>

Fixes: #59 PR-URL: #145 Reviewed-By: James M Snell <[email protected]>

PR-URL: #147 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

Introduced in aa99a6a. PR-URL: #149 Reviewed-By: James M Snell <[email protected]>

PR-URL: #149 Reviewed-By: James M Snell <[email protected]>

PR-URL: #154 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #153 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #152 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

Previously, if the expiry timestamp was in the future by less than 1ms, the retransmission timer would not have been scheduled. (This also removes an unused line from the test that this problem made flaky.) PR-URL: #157 Reviewed-By: James M Snell <[email protected]>

If we set `preferredAddressPolicy` to `accept` on `QuicClientSession` but receives an invalid preferred address, e.g. we don't define `preferredAddress` on the server, the `QuicClientSession` should still work. Before this PR, the included test will cause segmetation fault. PR-URL: #155 Reviewed-By: James M Snell <[email protected]>

It would be better to add a test to ensure that the timeout is greater than the triple PTO if we can get the PTO. PR-URL: #160 Reviewed-By: James M Snell <[email protected]>

PR-URL: #159 Reviewed-By: James M Snell <[email protected]>

This reverts commit 170c5d0. PR-URL: #138 Reviewed-By: Anna Henningsen <[email protected]>

This reverts commit ecda77c. PR-URL: #138 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #138 Reviewed-By: Anna Henningsen <[email protected]>

This commit will need to be submitted upstream then backed out once it lands and we can update PR-URL: #138 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #138 Reviewed-By: Anna Henningsen <[email protected]>

The recent commits have broken the build, tests, and linting. One cctest is not fixed but rather disabled here. PR-URL: #162

After ngtcp2 being updated, `ngtcp2_conn_get_idle_timeout` is renamed as `ngtcp2_conn_get_idle_expiry` which returns `ngtcp2_tstamp` instead of `ngtcp2_duration`. Refs: https://github.com/ngtcp2/ngtcp2/blob/6f40668cdce7db7c043d3a80c07f379841d8c51e/lib/ngtcp2_conn.c#L8604 PR-URL: #166 Reviewed-By: Anna Henningsen <[email protected]>

Fixes: #75 PR-URL: #150 Reviewed-By: James M Snell <[email protected]>

Also fix some lint issues. PR-URL: #171 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: James M Snell <[email protected]>

PR-URL: #167 Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #168 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

Eliminate the subclasses at the C++ level. Coalesce into a single QuicSession class to help keep things more manageable in preparation for the http3 enablement PR-URL: #168 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #173 Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

This simplifies the code quite a bit. PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

This means that we always need to use `DeleteFnPtr` to make sure that the timer is actually cleaned up, but I guess that’s an acceptable restriction for now. PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

Keeping the JS object alive is not enough, because the per-Environment cleanup may destroy C++ objects even if their JS counterparts are still reachable. PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

Provide facilities to test `QUIC` sockets without an actual underlying UDP handle for more resilient tests. PR-URL: #165 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>

PR-URL: #170 Reviewed-By: Anna Henningsen <[email protected]>

This addresses a TODO comment. PR-URL: #178 Reviewed-By: James M Snell <[email protected]>

This commit adds a suggestion for enabling the cctest test_quic_verifyhostnameidentity.cc. PR-URL: #169 Reviewed-By: Anna Henningsen <[email protected]>

Explicitly allow calling `.connect()` multiple times to generate separate client sessions on the same socket. This requires making `kMaybeBind` call its callback even if the socket is already bound. PR-URL: #176 Reviewed-By: James M Snell <[email protected]>

Once again, linter and tests have been failing because of recent changes. :) PR-URL: #179 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

`Shutdown()` is already used in our own `StreamBase` API, and in particular `QuicStream::Shutdown()` shadowed `StreamBase::Shutdown()`; the former cleanly shuts down the writable side of a stream, whereas the latter indicates an abrupt termination through a `RESET_STREAM` frame. PR-URL: #180 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

It seems that `Hash` class for `unorderred_map` is used to decide which bucket to put the key. But keys that are different pointers with the same hash will still be regarded as different keys so that we still need to specify a `std::equal_to` to compare the key correctly. However, when `maxConnectionsPerHost` is exceeded, the new session will still be emitted on the server side and won't be destroied automatically. We need to figure out the reason and fix it. PR-URL: #186 Reviewed-By: Anna Henningsen <[email protected]>

This resolves the situation in which `openStream()` is called before the underlying native `QuicSession` handle exists. Fixes: #181 PR-URL: #187 Reviewed-By: James M Snell <[email protected]>

PR-URL: #188 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

Currently the following compiler warnings are generated: In file included from ../src/node_quic_http3_application.cc:4: ../src/node_quic_http3_application.h:90:8: warning: 'StreamClose' overrides a member function but is not marked 'override' [-Winconsistent-missing-override] void StreamClose(int64_t stream_id, uint64_t app_error_code); ^ ../src/node_quic_session.h:339:16: note: overridden virtual function is here virtual void StreamClose( ^ In file included from ../src/node_quic_http3_application.cc:4: ../src/node_quic_http3_application.h:123:8:warning: 'MemoryInfo' overrides a member function but is not marked 'override' [-Winconsistent-missing-override] void MemoryInfo(MemoryTracker* tracker) const; ^ ../src/memory_tracker.h:121:16: note: overridden virtual function is here virtual void MemoryInfo(MemoryTracker* tracker) const = 0; ^ 2 warnings generated. In file included from ../src/node_quic_session.cc:19: ../src/node_quic_http3_application.h:90:8: warning: 'StreamClose' overrides a member function but is not marked 'override' [-Winconsistent-missing-override] void StreamClose(int64_t stream_id, uint64_t app_error_code); ^ ../src/node_quic_session.h:339:16: note: overridden virtual function is here virtual void StreamClose( ^ In file included from ../src/node_quic_session.cc:19: ../src/node_quic_http3_application.h:123:8: warning: 'MemoryInfo' overrides a member function but is not marked 'override' [-Winconsistent-missing-override] void MemoryInfo(MemoryTracker* tracker) const; ^ ../src/memory_tracker.h:121:16: note: overridden virtual function is here virtual void MemoryInfo(MemoryTracker* tracker) const = 0; ^ 2 warnings generated. This commit adds the missing overrides. PR-URL: #189 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

Currently, limits is included implicitly through node_quic_buffer.h which includes node_internals.h, which includes util.h, which includes limits. This commit makes explicitely include limits. PR-URL: #190 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

This commit removes the include of the functional header from node_quic_session.h and node_quic_session.cc as the are not currently used. PR-URL: #191 Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #195 Reviewed-By: Anna Henningsen <[email protected]>

This commit removes the include of the algorithm header as I cannot find any usage of it. PR-URL: #193 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

This commit removes the unused includes climits and algorithm, and adds memory (for std::unique_ptr) and utility (for std::move). PR-URL: #192 Reviewed-By: Jiawen Geng <[email protected]>

PR-URL: #196 Fixes: #161 Reviewed-By: Daniel Bevenius <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: James M Snell <[email protected]>

PR-URL: #194 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]> Reviewed-By: James M Snell <[email protected]>

PR-URL: #194 Reviewed-By: Anna Henningsen <[email protected]> Reviewed-By: Trivikram Kamat <[email protected]>

PR-URL: #177 Reviewed-By: Anna Henningsen <[email protected]>

Fixes: #185 PR-URL: #200 Reviewed-By: James M Snell <[email protected]>

PR-URL: #204 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #206 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #205 Reviewed-By: Anna Henningsen <[email protected]>

First step to reconciling shared code in node_crypto.cc PR-URL: #205 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #205 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #207 Reviewed-By: #207

Allow for better future reuse of sockaddr related code by moving SocketAddress into it's own header PR-URL: #207 Reviewed-By: #207

Fixes: #208 PR-URL: #207 Reviewed-By: #207

PR-URL: #207 Reviewed-By: #207

Fixes: #208 PR-URL: #207 Reviewed-By: #207

Fixes: #86 PR-URL: #207 Reviewed-By: #207

PR-URL: #207 Reviewed-By: #207

Allows monitoring the number of times packets have to be retransmitted because of loss or delayed ack. An implementation experiencing a high number of retransmits may need to be torn down or mitigated in some way. This allows the user code to keep track. Later, we may want to add a setting that enforces a maximum number of retransmissions before failing a connection. Fixes: #78 PR-URL: #207 Reviewed-By: #207

Fixes: #58 PR-URL: #207 Reviewed-By: #207

PR-URL: #207 Reviewed-By: #207

PR-URL: #212

We currently generate connection ID's randomly. There are, however, alternative schemes being defined. Later we may allow user code to select alternative strategies for generating connection ids. This adds the basic mechanism. PR-URL: #211 Reviewed-By: Anna Henningsen <[email protected]>

Count the number of times sending stream data is blocked by flow control PR-URL: #213 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #213 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #216 Reviewed-By: Anna Henningsen <[email protected]>

Generate stateless reset token cryptographically Fixes: #62 PR-URL: #215 Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #217 Reviewed-By: Anna Henningsen <[email protected]>

Currently the doc target fails and this commit attempts for fix these errors. PR-URL: #221 Reviewed-By: James M Snell <[email protected]> Reviewed-By: Anna Henningsen <[email protected]>

PR-URL: #224 Reviewed-By: Anna Henningsen <[email protected]>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[WIP] benchmark: add basic QUIC benchmark #182

[WIP] benchmark: add basic QUIC benchmark #182

Commits on Dec 11, 2019

[WIP] benchmark: add basic QUIC benchmark #182

Are you sure you want to change the base?

[WIP] benchmark: add basic QUIC benchmark #182

Commits on Dec 11, 2019