Chiebot-Mirror/grpc - grpc - Gitea: Git with a cup of tea

Commit Graph

Author	SHA1	Message	Date
Jan Tattermusch	e682b32340	fixup core uses-polling	5 years ago
Jan Tattermusch	068340381f	add uses_polling=False to selected core tests	5 years ago
Juanli Shen	6ced125bb3	xds failover locality handling	5 years ago
Mark D. Roth	ece8fd6642	Change transport connectivity watch API to not be lossy.	5 years ago
Hope Casey-Allen	673e0cec33	Fix	5 years ago
Ashitha Santhosh	8695c464dc	Expose max protected frame size value for testing.	5 years ago
Hope Casey-Allen	0f116eb282	Fix	5 years ago
Vijay Pai	5d7e4bb301	Stop using deprecated gtest macros	5 years ago
yang-g	49b796eeb1	Clean up the test	5 years ago
yang-g	76308c0c9a	Fix sanity_check and add a test	5 years ago
Jan Tattermusch	6ebc1c616f	set uses_polling=False for some bazel tests	5 years ago
Hope Casey-Allen	2af922e8c3	Skip service config tests for gcc4.8 and below	5 years ago
Lidi Zheng	781b12023a	Fix import failure due to 'visibility = ["//visibility:public"]'	5 years ago
Soheil Hassas Yeganeh	9b9764cb2a	Fix build error in SSL fuzzer.	5 years ago
Soheil Hassas Yeganeh	35e2760ffa	Add a fast path to allocate slices in the resource quota. Currently, we are always scheduling a callback to run the allocation. Instead, this patch adds a fastpath so that TCP can read the socket inline, instead of waiting for the exec context to flush. BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/64/2 [polls/iter:12 ] 65.3µs ± 0% 64.7µs ± 1% -0.92% (p=0.032 n=4+5) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/0/2 [polls/iter:12.0002 ] 61.4µs ± 4% 57.3µs ±13% -6.66% (p=0.030 n=5+7) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/1/2 [polls/iter:12.0002 ] 61.7µs ± 1% 60.8µs ± 1% -1.49% (p=0.003 n=9+4) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/32768/2 [polls/iter:12.0002 ] 114µs ± 2% 112µs ± 0% -2.09% (p=0.030 n=10+2) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/1/2 [polls/iter:12.0002 ] 62.9µs ± 0% 62.0µs ± 1% -1.32% (p=0.001 n=7+6) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/64/2 [polls/iter:12.0002 ] 66.5µs ± 3% 64.4µs ± 1% -3.20% (p=0.016 n=5+4) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/512/2 [polls/iter:12 ] 66.1µs ± 1% 65.2µs ± 1% -1.30% (p=0.003 n=8+5) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/4096/2 [polls/iter:12 ] 72.1µs ± 4% 70.8µs ± 1% -1.92% (p=0.004 n=9+5) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/0/2 [polls/iter:12.0001 ] 60.4µs ± 1% 59.4µs ± 0% -1.55% (p=0.004 n=6+5) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/8/2 [polls/iter:12 ] 63.6µs ± 1% 62.2µs ± 1% -2.13% (p=0.001 n=7+6) BM_StreamingPingPongMsgs<TCP, NoOpMutator, NoOpMutator>/4096 [polls/iter:4.00008 ] 20.9µs ± 1% 19.8µs ±13% -5.30% (p=0.029 n=4+4) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/0/1 [polls/iter:8.00014 ] 44.4µs ± 1% 41.9µs ±13% -5.51% (p=0.010 n=6+4) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/8/2 [polls/iter:12.0001 ] 62.1µs ± 0% 61.1µs ± 1% -1.56% (p=0.036 n=3+5) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/64/2 [polls/iter:12.0002 ] 64.0µs ± 0% 63.3µs ± 0% -1.14% (p=0.004 n=5+6) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/1/2 [polls/iter:12 ] 62.0µs ± 1% 61.0µs ± 1% -1.61% (p=0.032 n=4+5) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/0/2 [polls/iter:12.0002 ] 62.4µs ± 4% 60.6µs ± 1% -2.82% (p=0.003 n=7+5) BM_StreamingPingPongMsgs<TCP, NoOpMutator, NoOpMutator>/512 [polls/iter:4.00009 ] 17.9µs ± 1% 16.5µs ±11% -7.40% (p=0.032 n=4+5) BM_StreamingPingPong<TCP, NoOpMutator, NoOpMutator>/4096/1 [polls/iter:8.00016 ] 50.2µs ± 0% 49.5µs ± 1% -1.46% (p=0.024 n=3+6) BM_StreamingPingPong<MinTCP, NoOpMutator, NoOpMutator>/512/2 [polls/iter:12.0002 ] 65.5µs ± 3% 61.9µs ±13% -5.59% (p=0.048 n=5+7) BM_UnaryPingPong<MinTCP, NoOpMutator, NoOpMutator>/0/0 [polls/iter:3.00009 ] 23.4µs ± 2% 23.0µs ± 0% -1.88% (p=0.036 n=3+5) BM_UnaryPingPong<TCP, NoOpMutator, NoOpMutator>/512/0 [polls/iter:3.0001 ] 25.6µs ± 2% 25.0µs ± 0% -2.44% (p=0.017 n=7+3) BM_UnaryPingPong<TCP, NoOpMutator, NoOpMutator>/262144/0 [polls/iter:3.00022 ] 164µs ± 2% 161µs ± 1% -1.83% (p=0.048 n=3+6)	5 years ago
Soheil Hassas Yeganeh	05901eaaa3	Fix filter_latency.cc On Windows tests, clock monotonic doesn't progress. Switch to CLOCK_PRECISE to fix this.	6 years ago
Esun Kim	c965ca1494	Pull out configuration from ssl_utils	6 years ago
Soheil Hassas Yeganeh	c9376b4e0b	Use cycle clock instead of clock monotonic to measure call latency. This removes two more getttime syscalls from the hot path, when cycle clock is enabled.	6 years ago
Mark D. Roth	51d18ecc63	Convert mpscq API from C to C++.	6 years ago
Vijay Pai	fd9bdd9922	Avoid duplicate initialization of SSL (from within grpc and from test)	6 years ago
Mark D. Roth	d3f50ace39	Use mutex instead of combiner in client channel data plane.	6 years ago
Brian Zhao	8274502af8	Windows's STACKFRAME Frame Addr should be RBP, the base pointer, not RSP, the stack pointer. This is documented here: https://docs.microsoft.com/en-us/windows/win32/api/dbghelp/ns-dbghelp-stackframe with the comment "AddrFrame x64: The frame pointer is RBP or RDI." Note that this is also what StackWalker uses: https://github.com/JochenKalmbach/StackWalker#initializing-the-stackframe64 and what Chromium uses: https://codesearch.chromium.org/chromium/src/v8/src/base/debug/stack_trace_win.cc?l=200&rcl=69d20d247f62a3378d15ce0956ed8bf9665e6a44 release notes: no	6 years ago
Esun Kim	3c6bb96a1f	Add test for static elements	6 years ago
Vijay Pai	65eb9c9ddb	Revert "Test message size of 100MB"	6 years ago
Juanli Shen	d34f366337	Test message size of 100MB	6 years ago
Arjun Roy	b46e3668d3	s/branch/tail_call/ for CH2 on_hdr(). on_hdr() checks if a void-return function pointer is null before jumping to it. If it is null, it returns an error; else it executes that function and returns success. This change converts the void-returning function to one that returns a grpc_error* and thus saves a branch in on_hdr() (since we're branching once by following the function pointer anyways, we're effectively coalescing these two branches).	6 years ago
Mark D. Roth	21c5424477	Add API for accessing per-call backend metric data in LB policies.	6 years ago
Arjun Roy	2767accc1b	Fixed init-order-fiasco for static slice table. Fixes init-order bug affecting https://github.com/grpc/grpc/issues/19819 which was first exposed by this commit: `857375a142`	6 years ago
Mark D. Roth	18be57b4df	LB policy API changes suggested by Sanjay.	6 years ago
Esun Kim	2707fd0bd4	Added GRPC_USE_CPP_STD_LIB flag	6 years ago
yang-g	3728329033	return unavailable on transport closed	6 years ago
Esun Kim	a031f0ffd6	Remove all nanopb defines	6 years ago
Mark D. Roth	5d646ff9ff	Make Map<> copyable.	6 years ago
Esun Kim	e0b94db1b9	Enforce abstract class rule	6 years ago
Hope Casey-Allen	59564ebd96	Fix warnings to unblock gcc8 support	6 years ago
Matthew Stevenson	6189d34724	Revert "Forgot to autogenerate files and run clang_format/tidy_code" This reverts commit `9295e4ebb3`.	6 years ago
Matthew Stevenson	9295e4ebb3	Forgot to autogenerate files and run clang_format/tidy_code	6 years ago
Juanli Shen	a802b14be6	Add xds resolver	6 years ago
Qiancheng Zhao	1391c93a95	atomic client idle filter	6 years ago
Esun Kim	feb263ba6d	Add check_attrs & clean-up build.yaml	6 years ago
Yihua Zhang	ae863630d5	Add spiffe client-side credential reload	6 years ago
Juanli Shen	45dd8be442	Use LRS in xds policy	6 years ago
Esun Kim	be4e684a82	Migrated tsi/alts from nanopb to upb	6 years ago
Tony Lu	d2ce6e707c	Isolated test-specific build steps so that unrelated tasks don't run them Move interop making back to run_tests.py	6 years ago
Yihua Zhang	c414fe06f8	add cred data to BUILD files	6 years ago
Esun Kim	6f2234277a	Disable nanopb fuzzer test	6 years ago
Juanli Shen	79f191114e	Use EDS in xds LB policy	6 years ago
Yihua Zhang	8a301a438a	revision 1	6 years ago
Esun Kim	c6b0bd08c7	Updated grpclb to use upb	6 years ago
Soheil Hassas Yeganeh	bdd3fdddb2	Expose cycle counter and use it /channelz. We are using clock real time (which is very expensive) on every RPC to update channelz. We should simply use cycle clock. This patch exposes cycle clock, enables RDTSC on intel, and use it for channelz. This patch also changes the implementation of time_precise_init to that more accurately using the monotonic clock and shorter by looping at most a few hundred milliseconds. This is part of a larger performance series.	6 years ago

... 5 6 7 8 9 ...

6089 Commits (31d75a5f8de8381c42ec846303d48cc28bdecbbb)