Re: Patch: replace mutex with object synchronization

[Bryce McKinlay <mckinlay at redhat dot com>]

Anthony Green wrote:

> Since I've been looking at libgcj's use of posix threads I came across
> this bit of code which uses _Jv_Mutex_t to synchronize access to a data
> structure.  This is guaranteed to use the underlying system's mutex
> primitives.  We can do a little better by using object synchronization
> and our own lightweight locks.
>
> Perhaps this is a micro-optimization, but it also collapses the locking
> code down from three places to a single code fragment.

The reason I used the _Jv_Mutex_t here is due to the "low-level" nature 
of the class linking code. We need to create an instance of Object in 
order to use Java locks, so what happens if we're already in the middle 
of linking the Object class? I guess it doesn't matter right now, since 
libgcj itself isn't built with the BC-ABI, but eventually, using Java 
locks here could potentially lead to circularity problems.

Regards

Bryce