Add 'requiredThreads' field to the ThreadPool structure, to avoid
a race condition when waiting for a new thread to be created. The
race condition occurs when a thread is destroyed while the master
thread is waiting for a new thread to be created.
Thanks to Chuck Thomason for pointing the problem.
Summary: Race condition can hang miniserver thread - ID: 3158591
Details:
Hello,
I have found a race condition in the thread pool handling of
libupnp-1.6.6 that periodically results in the miniserver thread
getting blocked infinitely.
In my setup, I have the miniserver thread pool configured with 1
job per thread, 2 threads minimum, and 50 threads maximum.
Just before the lockup occurs, the miniserver thread pool contains
2 threads: one worker thread hanging around from a previous HTTP
request job (let's call that thread "old_worker") and the
miniserver thread itself.
A new HTTP request comes in. Accordingly, the miniserver enters
schedule_request_job() and then ThreadPoolAdd(). In
ThreadPoolAdd(), the job gets added to the medium-priority queue,
and AddWorker() is called. In AddWorker(), jobs = 1 and threads =
1, so CreateWorker gets called.
When we enter CreateWorker(), tp->totalThreads is 2, so
currentThreads is 3. The function creates a new thread and then
blocks on tp->start_and_shutdown. The miniserver thread expects
the newly created thread to increment tp->totalThreads and then
signal the condition variable to wake up the miniserver thread and
let it proceed.
The newly created thread starts in the WorkerThread() function. It
increments tp->totalThreads to 3, does a broadcast on the
start_and_shutdown condition, and starts running its job. However,
before the miniserver thread wakes up, "old_worker" times out. It
sees that there are no jobs in any queue and that the total number
of threads (3) is more than the minimum (2). As a result, it
reduces tp->totalThreads to 2 and dies.
Now the miniserver thread finally wakes up. It checks
tp->totalThreads and sees that its value is 2, so it blocks on
tp->start_and_shutdown again. It has now "missed" seeing
tp->totalThreads get incremented to 3 and will never be unblocked
again.
When this issue does occur for a server device, the miniserver
port remains open, but becomes unresponsive since the miniserver
thread is stuck. SSDP alive messages keep getting sent out, as
they are handled by a separate thread. Reproducing the issue is
difficult due to the timing coincidence involved, but in my
environment I am presently seeing it at least once a day. I
figured out the sequence described above through addition of my
own debug logs.
The relevant code involved in this bug has not changed
substantially in libupnp-1.6.10, though I am planning to test
against 1.6.10 as well in the near future.
Do you have any input for an elegant fix for this issue?
Thanks,
Chuck Thomason
When a lot of notifications were generated by a device in a short
period of time then 100% of the CPU was used to reorder those
notifications by pushing back the thread in the job queue. This
mechanism has been modified so now thread sleep 1 ms before being
pushed back into the job queue.
Removing DEFAULT_SCHED_PARAM parameter and use
sched_get_priority_min(DEFAULT_POLICY) instead.
This patch allows a user to customize the stack size of the threads used by
pupnp through the new THREAD_STACK_SIZE variable. This is especially useful
on embedded systems with limited memory where the user can set THREAD_STACK_SIZE
to ITHREAD_STACK_MIN.
However, as this modification can have side effects, I set 0 as the default
value, so threads will continue to use the default stack size of the system
(which varies greatly as stated in
https://computing.llnl.gov/tutorials/pthreads/).
libupnp and multi-flows scenario patch
Submited by Carlo Parata from STMicroelectronics.
Hi Roberto and Nektarios,
after an analysis of the problem of libupnp with a multi-flows scenario, I
noticed that the only cause of the freezed system is the ThreadPool
management. There are not mutex problems. In practise, if all threads in the
thread pool are busy executing jobs, a new worker thread should be created if
a job is scheduled (I inspired to tombupnp library). So I solved the problem
with a little patch in threadutil library that you can find attached in this
e-mail. I hope to have helped you.
git-svn-id: https://pupnp.svn.sourceforge.net/svnroot/pupnp/branches/branch-1.6.x@515 119443c7-1b9e-41f8-b6fc-b9c35fce742c
defines. These were just aliases, no reason to keep them.
* Changed the comments of the include files that expose the UPnP API
to use only C89 comments and no C99 comments.
git-svn-id: https://pupnp.svn.sourceforge.net/svnroot/pupnp/trunk@198 119443c7-1b9e-41f8-b6fc-b9c35fce742c
befor calling va_end().
- Removed all uses of the DBGONLY(x) macro. A static inline empty
function now is used and the compiler takes care of optimizing it out.
git-svn-id: https://pupnp.svn.sourceforge.net/svnroot/pupnp/trunk@185 119443c7-1b9e-41f8-b6fc-b9c35fce742c
Submitted By:
Fredrik Svensson - svefredrik
Incremented the libray versions and included some comments in the file
configure.ac so that we do not bump the library version excessively,
only the necessary numbers on the next release.
git-svn-id: https://pupnp.svn.sourceforge.net/svnroot/pupnp/trunk@115 119443c7-1b9e-41f8-b6fc-b9c35fce742c
