Don't complete writes of messages until they make it to the transports outbound loop. Since payloads could be large this introduces just enough pushback that, once #37868 goes in also we should be able to sense when a transport is busy writing and stop sending at higher layers.
Closes#37894
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37894 from ctiller:send-acked 0cb3d7f8ad
PiperOrigin-RevId: 686689473
This is missing in v3 vs v2
- in v2 we had Pipe setup so that multiple Pipe stages could be chained and only complete when the last stage had passed flow control, whereas in v3 the top stage will start accepting requests as soon as the first stage in the pipeline takes the message.
Closes#37868
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37868 from ctiller:drizzling 69209da8a7
PiperOrigin-RevId: 686652402
Passed interop test:
- [x] [grpc/core/master/linux/psm-csm-python](https://source.cloud.google.com/results/invocations/b9ba256b-31a9-4002-bd59-b21817aa9978)
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37837
PiperOrigin-RevId: 686643728
Currently the destructive reclaimer single threaded cancels existing requests, but we admit new rpcs on every channel (to be eventually cancelled, probably).
We've got evidence that this (shockingly) doesn't scale and senders can easily overwhelm and oom a server.
Instead under this experiment now we'll always reject new work under very high load, and allow the reclaimer to mop up any remaining work to get back to within bounds.
Closes#37927
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37927 from ctiller:fast_reject 835726473a
PiperOrigin-RevId: 686553599
You would think `always_inline` implies inline, but GCC complains loudly (`-Wattribute`) if the `always_inline` attribute is not also accompanied by the normal `inline` keyword.
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37834
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37834 from benjaminp:always-inline 3ecfaa6d50
PiperOrigin-RevId: 686553476
`std::string_view` is not allowed to use per Google C++ style guide. This test will prevent us from accidentally using it after C++17 upgrade.
Closes#37930
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37930 from veblush:string-view 88f56c12ba
PiperOrigin-RevId: 686526151
Follow up from #37917 to add `noexcept` to multiple Cython functions using `nogil` with a `void` return type.
Below are the performance hints that were encountered as part of the Cython translation which are solved by this PR:
The below performance hint still exists, and is not resolved.
This is because `noexcept` expects that the function doesn't raise an exception, or a raised exception is just displayed as a warning and not propagated. But `_poll` raises an `AssertionError` at `QUEUE_TIMEOUT`, and hence cannot use `noexcept` with `_poll`.
As [PR 37917](https://github.com/grpc/grpc/pull/37917) and this PR now solves the Cython Asyncio test timeouts caused by Cython upgrade, this PR also reverts the Bazel Cython downgrade PR #37884
### Testing for timeout
Tested using `bazel test -c dbg --runs_per_test=3000 --test_timeout=10 "//examples/python/auth:_auth_example_test"`
```
//examples/python/auth:_auth_example_test PASSED in 4.6s
Stats over 3000 runs: max = 4.6s, min = 2.6s, avg = 3.2s, dev = 0.2s
```
Closes#37922
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37922 from sreenithi:fix_bazel_cython_asyncio_timeout 2162cd28b1
PiperOrigin-RevId: 686524195
To prepare for the upcoming upgrade to C++17, the following changes were made:
Increased minimum supported operating system versions:
- iOS: 11 (previously 10)
- macOS: 10.14 (previously 10.12)
- tvOS: 13.0 (previously 12.0)
In addition to this, version requirements across different projects were updated to use these for consistency.
Closes#37931
PiperOrigin-RevId: 686519641
Currently this is still causing log spam for ruby.
Though this is in core, I believe this code path is specific to ruby.
Closes#37633
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37633 from apolcyn:reduce_log ca58824fdd
PiperOrigin-RevId: 686280644
Corrects bazel run command in README.
It does not match the bazel command in the BUILD file.
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37664
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37664 from nolasconapoleao:fix_interceptor_readme 42d828f2ec
PiperOrigin-RevId: 686223940
This migrates all of the xDS unit tests except for the fuzzer, which I'll get in a subsequent PR.
This also does not include the xDS e2e tests, which I will also do separately.
Closes#37896
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37896 from markdroth:xds_tests_use_real_protos2 de568b4e53
PiperOrigin-RevId: 686197812
This aims to resolve segfaults encountered when rebuilding ARM64 PHP Docker images. This segfaults came from gcc during ARM64 builds using QEMU. To workaround this problem, the PHP Docker images have been restructured to use apt-get for installation, thereby avoiding the use of gcc. This also needed to upgrade Debian and PHP to their latest versions, aligning with our minimum supported PHP 8.
Closes#37895
PiperOrigin-RevId: 686162640
The existing code fails to enable vsock on Android because it tries to enable vsock on Android when:
a) a certain Linux version is detected
b) and a certain libc is present.
For a), it requires that `<linux/version.h>` is included.
For b), Android does not use glibc. Instead `AF_VSOCK` is included since NDK version 14.
This commit fixes both issues by including the needed Linux header and detecting the `__NDK_MAJOR__` version.
For context, AOSP, has it always enabled:
06fb97455f%5E%21/src/core/lib/iomgr/port.h
But this patch compiles with older NDK version, in case someone is still using them.
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37733
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37733 from ricardoquesadawork:android 9b27187337
PiperOrigin-RevId: 685868655
This eliminates the need for the `grpc_cc_proto_library` bazel BUILD rule introduced in #37863.
To make this work, I had to upgrade several bazel dependencies and apply a patch to rules_go to work around https://github.com/bazelbuild/bazel/issues/11636.
Closes#37902
PiperOrigin-RevId: 685868647
Instead of getting value of `csm_mesh_id` from the bootstrap file, get it from the env var `CSM_MESH_ID`
Closes#37801
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37801 from yashykt:CsmMeshIdChange d0f149e023
PiperOrigin-RevId: 685864223
### What's happening
Some of our asyncio tests began timing out following a Cython upgrade to 3.0. This issue occurs consistently across both our Bazel and setup.py test environments.
### Why the time out
After some investigation, we found that our code here:4ffcdd4ab7/src/python/grpcio/grpc/_cython/_cygrpc/aio/completion_queue.pyx.pxi (L115-L116)
Was translated to this in Cython 0.29:
```
__pyx_f_7_cython_6cygrpc__unified_socket_write(__pyx_v_self->_write_fd);
```
And it changed to this in Cython 3.0:
```
__pyx_f_7_cython_6cygrpc__unified_socket_write(__pyx_v_self->_write_fd); if (unlikely(__Pyx_ErrOccurredWithGIL())) __PYX_ERR(7, 136, __pyx_L1_error)
```
Which indicates that this `nogil` function `_unified_socket_write` now requires GIL.
### What's new in Cython 3
* Cython 3 `cdef` functions with `void` return type will default to use `except *` as exception specification.
* If function have `void` return type and defined as `nogil`, Cython will always re-acquire the GIL after the function call to check if an exception has been raised.
* In some cases, this will cause a deadlock, especially if the function was called inside another `nogil` function.
### What's the fix
* This PR changes those functions to use `noexcept` as exception specification since we don't expect them to throw any exception, and this is also the suggested workarounds in Cython documentation: https://cython.readthedocs.io/en/latest/src/userguide/language_basics.html#error-return-values
### Test
* Tested locally by running `bazel test`, time our rate decreased from 5% to 0.3% in 3000 runs and 10s test time out.
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37917
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37917 from XuanWang-Amos:fix_cython_aio 063d27aee9
PiperOrigin-RevId: 685851320
There's a timeout flake when running windows tests for the Static CRL Provider, let's see if increasing the timeout helps
Closes#37915
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37915 from gtcooke94:increase_test_deadline 0386aab786
PiperOrigin-RevId: 685752852
These look too large for the configured timeouts internally... will revisit later once the system is starting to be more ready.
Closes#37905
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37905 from ctiller:lol-nope 6d44e515b3
PiperOrigin-RevId: 685223450
We'll probably disable some next week :)
But I want to watch a good selection and refine criteria for acceptance.
Closes#37903
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37903 from ctiller:ALL-the-things 5f829db870
PiperOrigin-RevId: 685010911
This will allow latent_see to be used with different build configurations in the future (and eventually perhaps defaulted on for some CI runs to aid postmortem debugging of test failures)
Closes#37897
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37897 from ctiller:latent-peek 52cbf36ce5
PiperOrigin-RevId: 684981143
So far missing for HTTP/2 style flow control has been a primitive to query whether there's a receiver for flow control data at the other end of the message pipes.
Here I'm updating the state machine accessors to accommodate that functionality.
No new states were needed.
Whilst here, document the current member functions on `CallState`.
Closes#37867
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37867 from ctiller:like-the-river c9814c737d
PiperOrigin-RevId: 684972125
Previously if a request was cancelled whilst being matched the matcher would leak a ref to the call forever.
Note that the only change here is `s/SpawnGuarded/SpawnGuardedUntilCallCompletes` and clang-format noise.
Closes#37886
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37886 from ctiller:flake-fightas-18 edc045102b
PiperOrigin-RevId: 684955517
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37883
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37883 from yousukseung:work_serializer_dispatch_extend 04c4602b73
PiperOrigin-RevId: 684887827
This is a trial baloon to see if we can actually make this work. If it does, I'll change the remaining xDS tests to use the real xDS protos and completely remove our local copies.
Closes#37863
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37863 from markdroth:xds_tests_use_real_protos 3ad2fe12be
PiperOrigin-RevId: 684877750
<!--
If you know who should review your pull request, please assign it to that
person, otherwise the pull request would get assigned randomly.
If your pull request is for a specific language, please add the appropriate
lang label.
-->
Closes#37853
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37853 from yijiem:dns-migration-chttp2-server b830720b20
PiperOrigin-RevId: 684631881
Prior to this change events could conspire such that newly read streams got added after the AbortWithError() code ran, and so those calls would be orphaned in the transport forever - continuing to hold a ref.
Closes#37887
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37887 from ctiller:flake-fightas-20 2648d7f37f
PiperOrigin-RevId: 684609806
This reverts commit e8aa408bba.
It looks like this increased flakiness... I'm going to roll forward the separate pieces after this.
Closes#37882
COPYBARA_INTEGRATE_REVIEW=https://github.com/grpc/grpc/pull/37882 from ctiller:flake-fightas-17 6e666cd985
PiperOrigin-RevId: 684577945
Mark D. Roth
Yash Tibrewal
Craig Tiller
AJ Heller
Xuan Wang
Benjamin Peterson
Esun Kim
Sreenithi Sridharan
apolcyn
Nolasco Napoleão
dependabot[bot]
Nana Pang
Ricardo Quesada
Gregory Cooke
Yijie Ma
Yousuk Seung
Richard Belleville
Nathan Baulch
Kannan J