C++ in the kernel

Bakul Shah bakul at bitblocks.com
Wed Oct 31 08:58:09 PDT 2007


> > > >     critical_section(lock) {
> > > > 	...
> > > > 	bar:
> > > > 	...
> > > > 	if (cond1) break;
> > > > 	...
> > > > 	if (cond2) goto foo;
> > > > 	...
> > > > 	if (cond3) goto bar;
> > > > 	...
> > > > 	if (cond4) return; // from enclosing function
> > > > 	...
> > > >     }
> > > >     ...
> > > >     foo:
> > > 
> > > 
> > > do you mean like C++:
> > > 
> > > do {
> > > 	critical_object critical_instance();
> > > 
> > > 
> > > 
> > > 
> > > }
> > 
> > No idea but I can not see how that will do what I had in mind. A purely
> > lexical translation of the snippet I gave above would be something like:
> 
> You can create an object on the stack that locks the mutex given
> to it  like so:
> 
> do {
> 	mtx_lock_object mtx_locker(&lock);
> 
> }
> 
> When the object is destroyed by stack popping, the lock will be freed.
> 
> It's the same thing.

Yes indeed, thanks!  I am starting to forget all the C++
tricks I learned.  Mercifully.  Two points though. This was
an example of what is possible with macros that can inspect
their argument code + they can also do many other things that
don't fit so easily with C++'s initialization/finalization
trick.  For example what if you can't gain the lock and want
to do something else?  Two, while C++ gives you a way to
solve this problem, it does it in a "clever" way, not an
obvious way.

But I will acknowledge I am comparing vaporware with
something that works now!


More information about the freebsd-arch mailing list