August 2010 - libdispatch-dev - lists.macosforge.org

Question about event handling in libdispatch
by cee1 02 Sep '10

02 Sep '10

Hi all, A kevent can associate with more than one source. What if: 1. A kevent of EVFILT_READ indicating that data is available to read. 2. A series of sources associating with this kevent are waken up. 3. One source handler fetches the data. 4. The rest source handlers will find there is no data available to read in monitored fd. It seems waking up more than one source doesn't make sense here? The rest source handlers will simply fail to read. -- Regards, - cee1

3 4

Re: [libdispatch-dev] libpthread_workqueue released
by Mark Heily 04 Aug '10

04 Aug '10

On 08/02/2010 09:17 AM, Chen Jie wrote: > Hi mark, > > I've read some code snippets of libpthread_workqueue 0.1. It doesn't > implement the over-commit feature, am I right? > Not yet. It also doesn't implement priority levels, and the algorithm for managing the size of the thread pool is very simplistic. > Does over-commit means when some threads in thread pool are blocked, a > new thread will be created even thread pool is full? When the blocked > threads are waken, the total number of active threads will exceed the > size of thread pool? > That sounds like the general idea, yes. The size of the thread pool is dynamic, and will be determined by the overall load on the system, so having "too many threads" is okay. As soon as the load exceeds the capacity of the system, the number of threads will be automatically reduced. - Mark

1 0

Question about "better assembly generation" in _dispatch_continuation_pop
by cee1 02 Aug '10

02 Aug '10

Hi all, I've read "_dispatch_continuation_pop" in src/queue.c(line 338). For async dispatch_continuation_t, it will first free, then execute it. The comment says this will generate better assembly, why? Regards, -- cee1

2 1