This is a multipoller based on epoll rather than poll.
Note that this implementation is aimed at correctness rather than
performance, although it should immediately have better scalability to
large numbers of FDs, both due to epoll's O(1) sized API and due to not
needing to wake up polling threads to do interest set changes.
One notable difference here is that we directly attach a wakeup fd
rather than using the freelisting kick mechanism that the poll() based
implementations use, because modifying the epoll set to use a different
kick fd each time isn't free.
The problem is that for the typedef to work we need _POSIX_C_SOURCE to
be defined properly before any file that uses gpr_timespec includes
anything. This is extremely fragile unless we change CFLAGS, which
probably isn't worth doing for this.
- Had to add support for files and environment variables as well.
- I can't compile on windows so I'm sure there will be some issues.
- Tested end-to-end with the simple ssl fullstack test.
This introduces the wakeup fd interface, corresponding approximately to
the existing Google version, complete with a ported giant detailed usage
comment.
The implementation has two layers, "specialized" and "fallback". The
specialized layer is intended to be a suitable platform specific
implementation like eventfd, whereas "fallback" is probably pipe, with
runtime detection of whether the specialized version works on this
system (currently stubbed out).
- Changed the unsigned char * + size to NULL terminated strings which
makes sense for the PEM format. I may change TSI later (but the impact
will hopefully be much more limited).
- Added a way to pass multiple key/cert pairs to servers which is needed
when hosting more than one domain.
- Removed the C++ SSL credentials tests as we are going to have an
option to not specify the roots which will then be derived from the
environment (well-known platform dependent locations and/or
environment variable).
- Fixed the php build which is the only one added in the run_test.py.
This change will certainly break node, python and ruby.
Allows servers to expedite shutdown once all in-progress calls are
completed. We may want to eventually remove grpc_server_shutdown in
preference to this.
Checked that the opensource version still builds.
Just addressing core for now. We will do c++ later.
Change on 2015/01/12 by jboeuf <jboeuf@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83804014
Allows tests to bind to port 0 and still have clients connect to them.
Change on 2015/01/12 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83800669
This necessitated the removal of gpr_vlog to keep a clean interface.
Change on 2015/01/09 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83634996
Fixed any exposed bugs across the stack.
Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them.
Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway.
Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent.
Things that need attention still:
- adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not
use this for multipolling once platform specific implementations are added.
- we rely on the backup poller too often - especially for SSL handshakes and for client
connection establishment we should have a better mechanism ([] []
- Linux needs to use epoll for multiple fds, FreeBSD variants (including
Darwin) need to use kqueue. ([] []
- Linux needs to use eventfd for poll kicking. ([]
Change on 2015/01/07 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83461069
Fixed any exposed bugs across the stack.
Add a poll() based implementation. Heavily leverages pollset infrastructure to allow small polls to be the norm. Exposes a mechanism to plug in epoll/kqueue for platforms where we have them.
Simplify iomgr callbacks to return one bit of success or failure (instead of the multi valued result that was mostly unused previously). This will ease the burden on new implementations, and the previous system provided no real value anyway.
Removed timeouts on endpoint read/write routines. This simplifies porting burden by providing a more orthogonal interface, and the functionality can always be replicated when desired by using an alarm combined with endpoint_shutdown. I'm fairly certain we ended up with this interface because it was convenient to do from libevent.
Things that need attention still:
- adding an fd to a pollset is O(n^2) - but this is probably ok given that we'll not
use this for multipolling once platform specific implementations are added.
- we rely on the backup poller too often - especially for SSL handshakes and for client
connection establishment we should have a better mechanism ([] []
- Linux needs to use epoll for multiple fds, FreeBSD variants (including
Darwin) need to use kqueue. ([] []
- Linux needs to use eventfd for poll kicking. ([]
Change on 2015/01/07 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=83461069
Properly set this up in our build environment.
Change on 2014/12/23 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82725899
This change implements a platform independent alarm manager in alarm.c.
It's integrated with iomgr, and some tests are cleaned up.
The alarm implementation itself is a fairly direct port of LazyAlarmList from eventmanager.
SpinLock has been replaced for now with gpr_mu, and other atomic operations have been dropped (again, for now).
A majority of tests have been ported.
Change on 2014/12/19 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82551363
The old test_ssl_channel related code is deleted and the new create channel
call is used for interop tests.
Change on 2014/12/19 by yangg <yangg@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82540921
it in subsequent cl's.
Change on 2014/12/17 by yangg <yangg@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82372934
Actually exposing it from the C API.
Also removing grpc_status, since it's mostly useless.
Change on 2014/12/15 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82180546
Additionally add rotl, rotr which will be needed to implement some hashing
functions shortly.
Change on 2014/12/15 by ctiller <ctiller@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82149559
a disaster and does not support url_safe which we need for the JWT.
Change on 2014/12/12 by jboeuf <jboeuf@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=82020601
-) Introducing gpr_vlog so to spare a few vsprintf later (at least one for now)
-) Renaming statistics/log.* to statistics/census_log.* to avoid collisions.
Change on 2014/12/12 by nnoble <nnoble@google.com>
-------------
Created by MOE: http://code.google.com/p/moe-java
MOE_MIGRATED_REVID=81995756
Craig Tiller