This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [RFC]: New zone based garbage collector and thoughts on incrementalcollection
On Fri, 5 Sep 2003, Richard Henderson wrote:
> On Fri, Sep 05, 2003 at 10:53:00AM -0400, Daniel Berlin wrote:
> > 1. In order to be able to restart marking (or for that matter, to mark
> > single zones at a time, or single generations at a time), we need to be
> > able to determine the marking function to call on any object.
> > We actually know it right now for the roots, and after marking, we know
> > it for all marked objects. Someone might be able to come up with a way to
> > use this information to mark object types outside of ggc_alloc time, but it's
> > something i've not been able to come up with.
>
> When I was thinking about this in the spring, I convinced myself
> that knowing the type at ggc_alloc time was never necessary.
For generations, sure, for incremental, you do need to know it, because
you need to restart marking at any point.
>
> I think what you'll be wanting is to mark the new generation from
> the previous generation. If you never reach an object during that
> first collection, then it doesn't matter what type it was, cause
> it's dead.
>
>
> r~
>