Hi Jan, On 7/9/19 7:36 AM, Jan Kiszka wrote: > Hi all, > > long-announced, I feel like it's finally in a good shape to publish it: The > new > version of something I would call "ivshmem 2.0" is now available in > wip/ivshmem2. Highlights of this:
Cool, thanks! Will try it. > > - vmexit-free peer state table in shared memory > - all key features of current QEMU's ivshmem, specifically > - multi-peer support > - multi-vector support > - unidirectional shared memory (optional) > - UIO-friendly one-shot interrupt mode (optional) I saw there's now a FEATURES register. Should we maybe also introduce a version/revision register for more invasive changes that go beyong adding a feature? Or will handling of different versions be done via vendor/device-ID (as I suggest that's the case at the moment). Ralf > > The corresponding kernel drivers, including a rewritten UIO driver, are > available under > http://git.kiszka.org/?p=linux.git;a=shortlog;h=refs/heads/queues/jailhouse-ivshmem2. > > The code hasn't been extensively tested yet, so any early feedback is welcome. > Just keep in mind that things are still in flux, specifically as the > discussion > with the QEMU and also virtio communities hasn't been started yet. > > One of the next steps is demonstrating the feasibility of stacking virtio on > top > of ivshmem2. I also plan to reach out to the QEMU community to ask for > feedback > and interest in enhancing their ivshmem model equivalently. In the ideal > world, > this model will not only used by Jailhouse in the end. > > Jan > -- You received this message because you are subscribed to the Google Groups "Jailhouse" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/jailhouse-dev/c89ecf4a-6287-34ac-b421-457c084d654a%40oth-regensburg.de. For more options, visit https://groups.google.com/d/optout.
