[FFmpeg-devel] [PATCH] Reimplementation of OS/2 threads model

[KO Myung-Hun]

M?ns Rullg?rd wrote:
> KO Myung-Hun <komh at chollian.net> writes:
> 
>> M?ns Rullg?rd wrote:
>>> KO Myung-Hun <komh at chollian.net> writes:
>>>>
>>>> +static Semaphore *CreateSemaphore(void *v1, LONG lCount, LONG lMax, void *v2){
>>>
>>> Is this emulating something?  What's wrong with semaphore calls in the
>>> old code?
>>>
>>
>> The (event) semaphore of the old code wakes up all the waiting threads.
>> And it resets the status to non-signaled regardless of its post count.
>>
>> So I decided to emulate the semaphore model of Win32. It can wake up
>> waiting threads one by one and decrease its post count.
> 
> Would it be possible to use the native semaphores, or is this
> emulation strictly required?  If we end up emulating win32threads, we
> might as well use the pthreads emulation instead and save ourselves
> some trouble.
> 

Frankly, I cannot find any differences between the native semaphores and
the emulated models. But I agree with you that it would be better to use
the pthreads emulation in order to follow up the changes of pthread.c in
the future.

I'll try this later.

-- 
KO Myung-Hun

Using Mozilla SeaMonkey 2.0.3
Under OS/2 Warp 4 for Korean with FixPak #15
On AMD ThunderBird 1GHz with 512 MB RAM

Korean OS/2 User Community : http://www.ecomstation.co.kr