On Wed, Dec 6, 2017 at 1:04 PM, Casey Bodley <cbodley@xxxxxxxxxx> wrote: > Following up on today's CDM discussion about experimenting with seastar, > here is the branch I mentioned: > https://github.com/cbodley/ceph/commits/wip-seastar-msg > > The commit that integrates it into our cmake is > https://github.com/cbodley/ceph/commit/c897cd9b66977df633a4408f12c8d6b3fa4d45b4. > This doesn't add seastar as a submodule, but uses the pkg-config file > seastar.pc that's generated by their build process. When built with dpdk, > seastar includes all of the dpdk cflags and ldflags it uses, so we'll need > to reconcile those with our existing dpdk build. > > I also mentioned some work to integrate seastar with the c++ coroutines ts > (http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2017/n4680.pdf) - see > https://groups.google.com/forum/#!msg/seastar-dev/tLvS7sx3ss4/_-P93HtsBgAJ > for that discussion. Huh, I see a statement in that discussion "There's a conflict between the reactor's model of single-use tasks, and reusable continuations." That sounds to me like even if we wanted to modify seastar to use a more sophisticated memory-allocation/futures-grouped-as-tasks model, it would not go well with how the project currently works. Anybody have any sophisticated insight there before we open up a dialogue with them? -Greg > > Casey > > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
