Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Merge pull request #2117 from dgquintas/bb_from_bbreader | Craig Tiller | 2015-07-05 |
|\ | | | | | Introduced function to create byte buffers from the output of a byte buffer reader. | ||
* | | Fix refcount leak | Craig Tiller | 2015-07-01 |
| | | |||
* | | Merge pull request #2284 from ctiller/oops-i-split-it-again | David G. Quintas | 2015-07-01 |
|\ \ | | | | | | | Remove errant assert | ||
| * | | Properly reset counter | Craig Tiller | 2015-07-01 |
| | | | |||
| * | | Remove errant assert | Craig Tiller | 2015-07-01 |
| | | | |||
* | | | Merge pull request #2282 from ctiller/red-orange-yellow-green-blue-indigo-violet | Nicolas Noble | 2015-07-01 |
|\ \ \ | | | | | | | | | Fix a TSAN reported race | ||
| * | | | Fix a TSAN reported race | Craig Tiller | 2015-07-01 |
| | | | | | | | | | | | | | | | | | | | | close() could race with epoll_ctl(); pretend to be polling while adding to the epoll set to prevent this | ||
* | | | | Merge pull request #2277 from ↵ | Nicolas Noble | 2015-07-01 |
|\ \ \ \ | |/ / / |/| | | | | | | | | | | | ctiller/in-the-town-where-i-was-born-there-lived-a-man Fix a bug in the multipoll on poll path where we use a file descripto… | ||
* | | | | Don't unregister resolver object until callback complete | Craig Tiller | 2015-07-01 |
| | | | | | | | | | | | | | | | | Prevents TSAN races in iomgr shutdown code | ||
* | | | | Delay unregister of fd until freelisted | Craig Tiller | 2015-07-01 |
| | | | | | | | | | | | | | | | | Prevents a race whereby we start deleting the freelist before it's used | ||
* | | | | Try harder to return DEADLINE_EXCEEDED when we should | Craig Tiller | 2015-07-01 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Do this by ensuring that the alarm callback has had a chance to run on a call before returning status to the application. If we do not do this: - the server alarm could be scheduled and run - it will write a RST_STREAM with a status that loses the deadline exceededness (because that is unexpressable in HTTP2 error codes) - it will be received by the client and processed - the client will return an INTERNAL error (the lossy re-encoding of the server status), and then run its alarm handler to set the status to something else | ||
| * | | | Fix a bug in the multipoll on poll path where we use a file descriptor thats ↵ | Craig Tiller | 2015-07-01 |
|/ / / | | | | | | | | | | orphaned by mistake | ||
* | | | Merge pull request #2216 from jboeuf/unpadded_base64 | Nicolas Noble | 2015-07-01 |
|\ \ \ | | | | | | | | | Base64 decode improvements | ||
* \ \ \ | Merge pull request #2229 from ctiller/forever-is-a-long-time | David Klempner | 2015-06-30 |
|\ \ \ \ | | | | | | | | | | | Fix fd handling on posix | ||
* | | | | | Fix leak in chttp2_transport | Craig Tiller | 2015-06-30 |
| | | | | | |||
| * | | | | Merge github.com:grpc/grpc into forever-is-a-long-time | Craig Tiller | 2015-06-30 |
| |\ \ \ \ | |||
* | | | | | | Spam cleanup | Craig Tiller | 2015-06-30 |
| |/ / / / |/| | | | | |||
* | | | | | Remove the ability to specify a string on cancel: it was broken and unused; ↵ | Craig Tiller | 2015-06-30 |
| | | | | | | | | | | | | | | | | | | | | will restore later | ||
* | | | | | Fix a bug where transports can leak because they are not shut down correctly | Craig Tiller | 2015-06-30 |
| | | | | | |||
* | | | | | Fix leaking unclaimed incoming metadata | Craig Tiller | 2015-06-30 |
| | | | | | |||
* | | | | | SSL refcounting fixes | Craig Tiller | 2015-06-30 |
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Handle the case where we recreate an auth context. Add (opt-in) debugging for refcounts on auth contexts. | ||
* | | | | | Fix leak in chttp2_transport incoming metadata | Craig Tiller | 2015-06-30 |
| | | | | | |||
| | * | | | Fixing indentation. | Julien Boeuf | 2015-06-30 |
| | | | | | |||
| * | | | | Add an fd conservation test to verify that fds actually close | Craig Tiller | 2015-06-29 |
| | | | | | |||
* | | | | | Merge pull request #2149 from ctiller/oops-i-split-it-again | Vijay Pai | 2015-06-26 |
|\ \ \ \ \ | | |_|/ / | |/| | | | Initial transport lock split up | ||
| | * | | | Merge github.com:grpc/grpc into forever-is-a-long-time | Craig Tiller | 2015-06-26 |
| | |\ \ \ | |_|/ / / |/| | | | | |||
* | | | | | Merge pull request #2204 from nicolasnoble/completion-queue-refcount-debugging | Craig Tiller | 2015-06-26 |
|\ \ \ \ \ | | | | | | | | | | | | | Adding file and line number to cq debug refcount. | ||
| | | * \ \ | Merge github.com:grpc/grpc into forever-is-a-long-time | Craig Tiller | 2015-06-26 |
| | | |\ \ \ | |_|_|/ / / |/| | | | | | |||
* | | | | | | Merge pull request #2221 from yang-g/no_new_bbq_after_flush | Craig Tiller | 2015-06-26 |
|\ \ \ \ \ \ | | | | | | | | | | | | | | | Bug fix. Do not push to incoming_queue after it is flushed. | ||
| | | | * | | | Call fd_orphan callback a little earlier | Craig Tiller | 2015-06-26 |
| | | | | | | | |||
* | | | | | | | Merge pull request #2214 from ctiller/are-includes-sane | Nicolas Noble | 2015-06-25 |
|\ \ \ \ \ \ \ | |_|_|_|/ / / |/| | | | | | | Add a test that headers are properly in build.json | ||
| | * | | | | | Bug fix. Do not push to incoming_queue after it is flushed. | yang-g | 2015-06-25 |
| | | | | | | | |||
* | | | | | | | Reworking the cascade kick to be Windows-specific. | Nicolas "Pixel" Noble | 2015-06-25 |
| |/ / / / / |/| | | | | | | | | | | | | | | | | | This brings the Windows port more in-par with Linux. We're now making sure all of the pollsets are going to return immediately before calling the shutdown callback. | ||
| | | * | | | Addressing comments | Craig Tiller | 2015-06-25 |
| | | | | | | |||
| | | | * | | Base64 decode improvements. | Julien Boeuf | 2015-06-25 |
| |_|_|/ / |/| | | | | | | | | | | | | | | | | | | | - Allow ablity to specify the length for decoding. - Allow for non-padded encodings to be decoded properly. | ||
| | | * | | Simpler code | Craig Tiller | 2015-06-25 |
| | | | | | |||
| | | * | | Remove dead code | Craig Tiller | 2015-06-25 |
| | | | | | |||
| | | * | | Merge github.com:grpc/grpc into oops-i-split-it-again | Craig Tiller | 2015-06-25 |
| | | |\ \ | |_|_|/ / |/| | | | | |||
| * | | | | Merge github.com:grpc/grpc into are-includes-sane | Craig Tiller | 2015-06-25 |
| |\ \ \ \ | |/ / / / |/| | | | | | | | | | | | | | | Conflicts: tools/doxygen/Doxyfile.c++.internal | ||
| * | | | | Add a test that headers are properly in build.json | Craig Tiller | 2015-06-25 |
| | | | | | | | | | | | | | | | | | | | | Clean up anything that failed | ||
* | | | | | Merge pull request #2202 from nicolasnoble/don-t-recursive-lock-me | Craig Tiller | 2015-06-24 |
|\ \ \ \ \ | | | | | | | | | | | | | Let's inform the system of the lock to unlock... | ||
* \ \ \ \ \ | Merge pull request #2205 from nicolasnoble/threadpools-need-to-cascade-kick | Craig Tiller | 2015-06-24 |
|\ \ \ \ \ \ | | | | | | | | | | | | | | | Cascade-kick pollsets when shutting down. | ||
* | | | | | | | On pollset shutdown, we need to unconditionally start the kick cascade. | Nicolas "Pixel" Noble | 2015-06-24 |
| | | | | | | | |||
| * | | | | | | Cascade-kick pollsets when shutting down. | Nicolas "Pixel" Noble | 2015-06-24 |
|/ / / / / / | | | | | | | | | | | | | | | | | | | In the case we're using a threadpool, depending on where the threads are waiting in completion queue's next, they'll get stuck on shutdown because we're only kicking one of them. Instead, let's cascade-kick the pollsets when we shutdown so that we make sure all of them are exitting properly. | ||
| | | * / / | Adding file and line number to cq debug refcount. | Nicolas "Pixel" Noble | 2015-06-24 |
| |_|/ / / |/| | | | | |||
* | | | | | Merge pull request #2185 from ctiller/it-hides-them | Nicolas Noble | 2015-06-24 |
|\ \ \ \ \ | | | | | | | | | | | | | Move some code out of src/ into tools/ | ||
* \ \ \ \ \ | Merge pull request #2167 from ctiller/everyone-is-special | David G. Quintas | 2015-06-24 |
|\ \ \ \ \ \ | |_|_|/ / / |/| | | | | | Return dummy data for empty histograms | ||
| | | * | | | Let's inform the system of the lock to unlock... | Nicolas "Pixel" Noble | 2015-06-24 |
| |_|/ / / |/| | | | | | | | | | | | | | | Fixes #2190. In grpc_pollset_work we need to pass down which lock to unlock before actually doing something. | ||
| | | * | | Merge github.com:grpc/grpc into oops-i-split-it-again | Craig Tiller | 2015-06-24 |
| | | |\ \ | |_|_|/ / |/| | | | | |||
| | | * | | Added comments | Craig Tiller | 2015-06-24 |
| | | | | |