ast_threadpool_serializer_group leaks a reference to ser when listener
is allocated but tps is not. Although listener takes the reference to
ser cleanup functions are not run without tps.
ASTERISK-26191 #close
Change-Id: Ie3ccf69a3f1e676c2ef62a77067c0cb57dc9a585
worker_start checked for ZOMBIE status without holding a lock. All
other read/write of worker status are performed with a lock, so this
check should do the same.
ASTERISK-25777 #close
Change-Id: I5e33685a5c26fdb300851989a3b82be8c4e03781
We have observed situations where the SIP threadpool may become
deadlocked. However, because incoming traffic is still arriving, the SIP
threadpool's queue can continue to grow, eventually running the system
out of memory.
This change makes it so that incoming traffic gets rejected with a 503
response if the queue is backed up too much.
Change-Id: I4e736d48a2ba79fd1f8056c0dcd330e38e6a3816
This change adds handling of dead worker threads when moving them
to be active. When this happens the worker thread is removed from
both the active and idle threads container. If no threads are able
to be moved to active then the pool grows as configured.
A unit test has also been added which thrashes the idle timeout
and thread activation to exploit any race conditions between the
two.
ASTERISK-25546 #close
Change-Id: I6c455f9a40de60d9e86458d447b548fb52ba1143
A module trying to unload needs to wait for all serializers it creates and
uses to complete processing before unloading.
ASTERISK-24907
Reported by: Kevin Harwell
Change-Id: I8c80b90f2f82754e8dbb02ddf3c9121e5e966059
All send/receive processing for a SIP transaction needs to be done under
the same threadpool serializer to prevent reentrancy problems inside
pjproject when using an external DNS resolver to process messages for the
transaction.
* Add threadpool API call to get the current serializer associated with
the worker thread.
* Pick a serializer from a pool of default serializers if the caller of
res_pjsip.c:ast_sip_push_task() does not provide one.
This is a simple way to ensure that all outgoing SIP request messages are
processed under a serializer. Otherwise, any place where a pushed task is
done that would result in an outgoing out-of-dialog request would need to
be modified to supply a serializer. Serializers from the default
serializer pool are picked in a round robin sequence for simplicity.
A side effect is that the default serializer pool will limit the growth of
the thread pool from random tasks. This is not necessarily a bad thing.
* Made pjsip_resolver.c use the requesting thread's serializer to execute
the async callback.
* Made pjsip_distributor.c save the thread's serializer name on the
outgoing request tdata struct so the response can be processed under the
same serializer.
ASTERISK-25115 #close
Reported by: John Bigelow
Change-Id: Iea71c16ce1132017b5791635e198b8c27973f40a
When a threadpool is set to autoincrement its threadcount, an issue
may arise when multiple tasks are queued at once into the threadpool. Since
threads start active, each new task would result in autoincrementing the
thread count. So if all threads were active, and a thread's autoincrement
value were 5, then 3 new tasks would result in 15 threads being created even
though the initial autoincrement was sufficient to handle the number of tasks.
This change introduces three behavior changes:
1) New threads in the threadpool start idle instead of active.
2) When a threadpool autoincrements, one thread is activated after the growth.
3) When a threadpool's size is incremented manually, all added threads are activated.
For a more detailed explanation about the changes, please see the Review Board link
at the bottom of this commit.
Review: https://reviewboard.asterisk.org/r/2629
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@392318 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The pimp_my_sip branch is being merged at this point because
it offers basic functionality, and from an API standpoint, things
are complete.
SIP work is *not* feature-complete; however, with the completion
of the SUBSCRIBE/NOTIFY API, all APIs (except a PUBLISH API) have
been created, and thus it is possible for developers to attempt
to create new SIP work.
API documentation can be found in the doxygen in the code, but
usability documentation is still lacking.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@386540 65c4cc65-6c06-0410-ace0-fbb531ad65f3
If ast_threadpool_set_size with a size equal to the current size, a
reference to a set_size_data structure would be leaked.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@382555 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Ownership of the listener reference is not transferred because the
listener is reffed when placed into the taskprocessor. Ensure that the
listener is dereffed properly.
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@382489 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This patch adds the ability to create a serializer from a thread pool. A
serializer is a ast_taskprocessor with the same contract as a default
taskprocessor (tasks execute serially) except instead of executing out
of a dedicated thread, execution occurs in a thread from a
ast_threadpool. Think of it as a lightweight thread.
While it guarantees that each task will complete before executing the
next, there is no guarantee as to which thread from the pool individual
tasks will execute. This normally only matters if your code relys on
thread specific information, such as thread locals.
This patch also fixes a bug in how the 'was_empty' parameter is computed
for the push callback, and gets rid of the unused 'shutting_down' field.
Review: https://reviewboard.asterisk.org/r/2323/
git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@381326 65c4cc65-6c06-0410-ace0-fbb531ad65f3
* Add a max_size option for threadpools. Also added a test for this option.
* Fixed comments to be more accurate and have fewer typos.
* Updated copyright dates on new files.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@379375 65c4cc65-6c06-0410-ace0-fbb531ad65f3
* Clarify some documentation
* Change copyright date of taskprocessor files
* Address potential issue of creating taskprocessor with listener if
taskprocessor with that name exists already
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@379124 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Now user data is allocated by the creator of the taskprocessor
listener and that user data is passed into ast_taskprocessor_listener_alloc().
Similarly, freeing of the user data is left up to the user himself. He can
free the data when the taskprocessor shuts down, or he can choose to hold
onto it if it makes sense to do so.
This, unsurprisingly, makes threadpool allocation a LOT cleaner now.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@379120 65c4cc65-6c06-0410-ace0-fbb531ad65f3
* Remove extraneous whitespace
* Bump up debug levels of messages and add identifying info to messages.
* Account for potential failures of ao2_link()
* Add additional test and some more test data
* Add some comments in places where they could be useful
* Make threadpool listeners and their callbacks optional
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@378652 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The documentation for taskprocessors was incorrect with
regards to when a listener's alloc callback was called.
I also made the names of queued function calls in the
threadpool more uniform.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377802 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Unfortunately, this required a taskprocessor listener change that makes listener allocation
utterly silly. I'm going to change the scheme so that allocation of taskprocessor listeners
is done internally within taskprocessor code. This will make it parallel with threadpool
code, which is a good thing.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377687 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The only test added so far is an idle thread timeout
option. This will greatly aid threadpool users who wish
to maintain a threadpool by allowing for idle threads to
die out as necessary.
Test passes.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377580 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This helps tests to pass more often than before.
They are far less likely to queue extra processes
into the control taskprocessor since they are prevented
once the threadpool begins to shut down.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377578 65c4cc65-6c06-0410-ace0-fbb531ad65f3
This one involves shrinking the threadpool in such
a way that both idle and active threads are affected.
This test made me re-realize why the zombie state exists,
so I re-added it. We don't want to clog up the control
taskprocessor by waiting on active threads to complete
what they are doing. Instead, we mark them as zombies so
that when they are done, they can clean themselves up
properly.
Without the zombie state available, the new test actually
will deadlock.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377474 65c4cc65-6c06-0410-ace0-fbb531ad65f3
The new thread creation test fails because Asterisk locks up
while trying to lock a taskprocessor.
While trying to debug that, I found a race condition during taskprocessor
creation where a default taskprocessor listener could try to operate on
a partially started taskprocessor. This was fixed by adding a new callback
to taskprocessor listeners.
Then while testing that change, I found some bugs in the taskprocessor
tests where I was not properly unlocking when done with a lock. Scoped
locks have spoiled me a bit.
I still have not figured out why the threadpool thread creation test
is locking up.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377368 65c4cc65-6c06-0410-ace0-fbb531ad65f3
After giving it some consideration, there's no real
use for zombie threads. Listeners can't really use the
current number of zombie threads as a way of gauging activity,
zombifying threads is just an extra step before they die that
really serves no purpose, and since there's no way to re-animate
zombies, the operation does not need to be around.
I also fixed up some miscellaneous compilation errors that
were lingering from some past revisions.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377211 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Since threadpool shutdown is very strictly controlled,
there is no need to be so precise with reference counts
in queued operations. Since the threadpool shuts down its
own control taskprocessor before doing anything else destructive,
it can be guaranteed that all queued tasks will have a valid
pointer to the pool. This meant that some destructor functions
for helper structs could be removed entirely.
git-svn-id: https://origsvn.digium.com/svn/asterisk/team/mmichelson/threadpool@377210 65c4cc65-6c06-0410-ace0-fbb531ad65f3
Corey Farrell
Mark Michelson
Joshua Colp
Richard Mudgett
David M. Lee
Kinsey Moore