Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Fixes | 2015-09-21 | |
| | |||
* | Call list progress | 2015-09-21 | |
| | |||
* | Call list progress | 2015-09-18 | |
| | |||
* | Cleanup | 2015-09-18 | |
| | |||
* | Introduce call lists for moving work outside locks | 2015-09-18 | |
| | |||
* | Poke fd workqueues on add - to unblock a potential starvation issue | 2015-09-17 | |
| | |||
* | Merge github.com:grpc/grpc into shindig | 2015-09-17 | |
|\ | | | | | | | | | Conflicts: src/core/iomgr/tcp_server_posix.c | ||
| * | Move in-flight callback check up | 2015-09-16 | |
| | | | | | | | | | | | | | | | | Fixes #3203 We entered a live lock race between the promotion callback and pollers, which was only resolved when the background executor got lucky enough to pick up the async work. | ||
* | | Tests compile | 2015-09-16 | |
| | | |||
* | | Progress towards workqueue transition | 2015-09-15 | |
| | | |||
* | | Merge github.com:grpc/grpc into shindig | 2015-09-11 | |
|\| | |||
| * | Enable -Wconversion | 2015-09-10 | |
| | | |||
| * | Sign conversion fixes | 2015-09-10 | |
| | | |||
* | | Starting to convert code to work queues | 2015-09-09 | |
|/ | |||
* | Add a test of non-blocking API behavior | 2015-08-06 | |
| | | | | ... also fix things that were broken :) | ||
* | Working towards a non-blocking API test | 2015-08-06 | |
| | |||
* | Fix TSAN reported lock-inversion in epoll fd addition | 2015-07-30 | |
| | |||
* | Allow specific pollers to be woken | 2015-07-29 | |
| | | | | | | | | | | | | | | | | | | | Currently, if two threads call grpc_completion_queue_pluck on the same completion queue for different tags, there is a 50% chance that we deliver the completion wakeup to the wrong poller - forcing the correct poller to wait until its polling times out before it can return an event up to the application. This change tweaks our polling interfaces so that we can indeed wake a specific poller. Nothing has been performance tuned yet. It's definitely sub-optimal in a number of places. Wakeup file-descriptors should be recycled. We should have a path that avoids calling poll() followed by epoll(). We can probably live without it right at the second though. This code will fail on Windows at least (I'll do that port when I'm in the office and have a Windows machine). | ||
* | Move alarm subsystem to monotonic time | 2015-07-19 | |
| | |||
* | Merge pull request #2470 from ctiller/i-want-to-wait-free | 2015-07-17 | |
|\ | | | | | Lock free requested call fulfillment path | ||
| * | Spelling correction | 2015-07-17 | |
| | | |||
| * | Expand comment | 2015-07-16 | |
| | | |||
| * | Fix a leaked lock, and make a debug-only detection for this class of bug | 2015-07-16 | |
| | | |||
| * | Hoist epoll_ctl outside of pollset lock | 2015-07-14 | |
| | | |||
* | | Updating wrapped languages to new time functions | 2015-07-13 | |
| | | |||
* | | Introduce a clock type field onto gpr_timespec. | 2015-07-13 | |
|/ | | | | | Use it to validate that arithmetic on time types makes even some vague kind of sense. | ||
* | Merge github.com:grpc/grpc into footprints-on-the-sands-of-time | 2015-07-07 | |
|\ | | | | | | | | | | | Conflicts: src/core/channel/client_setup.c src/core/surface/init.c | ||
| * | Guard against double pollset shutdown completion | 2015-07-06 | |
| | | |||
* | | clang-format changed files | 2015-07-06 | |
| | | |||
* | | Introduce multiple clocks to GPR | 2015-07-01 | |
|/ | |||
* | Remove return status from maybe_work. | 2015-06-22 | |
| | | | | | | | Since alarm checks may mutate work deadlines for pollsets, the value returned from maybe_work is meaningless. Instead, maybe pollset_work always return 1 if maybe_work is invoked, and then redo the deadline check _on the next call_ to pollset_work. | ||
* | Docs | 2015-06-18 | |
| | |||
* | clang-format affected files | 2015-06-17 | |
| | |||
* | Be better about timeout rounding | 2015-06-15 | |
| | |||
* | prevent double destroy of pollset | 2015-06-10 | |
| | |||
* | clang-format | 2015-06-05 | |
| | |||
* | Bandaid over pollset promotion | 2015-06-03 | |
| | |||
* | Fix MSAN reported error | 2015-06-02 | |
| | |||
* | Compile fix | 2015-06-01 | |
| | |||
* | Merge branch 'count-the-things' into we-dont-need-no-backup | 2015-06-01 | |
|\ | | | | | | | | | | | | | | | | | | | | | Conflicts: Makefile src/core/iomgr/pollset_posix.c src/core/surface/call.c src/core/surface/channel.c src/core/surface/server.c src/python/src/grpc/_adapter/_low_test.py tools/doxygen/Doxyfile.core.internal | ||
* | | Release all fd refs before signalling shutdown | 2015-06-01 | |
| | | |||
* | | fd refcount debugging | 2015-05-31 | |
| | | |||
| * | removed managed closures from server.c | 2015-05-31 | |
| | | |||
| * | Comments addressed. | 2015-05-31 | |
| | | |||
| * | Adapted the following to the new iomgr's cb API: | 2015-05-31 | |
| | | | | | | | | | | alarm_test, tcp_posix, fd_posix, pollset_posix, credentials, call, channel, server, child_channel | ||
* | | Spam cleanup | 2015-05-29 | |
| | | |||
* | | Reduce spam | 2015-05-29 | |
| | | |||
* | | Fix null deref | 2015-05-28 | |
| | | |||
* | | Refactor pollset_kick | 2015-05-28 | |
| | | | | | | | | | | Allow multiple threads to be polling Remove unnecessary windows stubs | ||
* | | Add debugging | 2015-05-28 | |
| | |