Re: Any interest in a QEMU emulation BoF at KVM Forum?

["Cédric Le Goater" <clg@kaod.org>]

On 9/13/22 09:12, Alex Bennée wrote:
> The BoF session will be in Lifey A (the big hall) this afternoon. I thought being able to sit around tables while we discuss things would make things a bit easier. We can share note taking on the etherpad:
> 
> https://etherpad.opendev.org/p/qemu-emulation-bof%40kvmforum2022 <https://etherpad.opendev.org/p/qemu-emulation-bof%40kvmforum2022>
> 
> I'll run a HO at: https://meet.google.com/rac-axef-xvv <https://meet.google.com/rac-axef-xvv>

Thanks for that. I will try to join. What time approximately ?

C.


> 
> On Wed, 31 Aug 2022 at 16:19, Alex Bennée <alex.bennee@linaro.org <mailto:alex.bennee@linaro.org>> wrote:
> 
>     Hi,
> 
>     Given our slowly growing range of TCG emulations and the evident
>     interest in keeping up with modern processor architectures is it worth
>     having an emulation focused BoF at the up-coming KVM Forum?
> 
>     Some potential topics for discussion I could think of might include:
> 
>       * Progress towards heterogeneous vCPU emulation
> 
>       We've been making slow progress in removing assumptions from the
>       various front-ends about their global nature and adding accel:TCG
>       abstractions and support for the translator loop. We can already have
>       CPUs from the same architecture family in a model. What else do we need
>       to do so we can have those funky ARM+RiscV+Tricore heterogeneous
>       models? Is it library or something else?
> 
>       * External Device Models
> 
>       I know this is a contentious topic given the potential for GPL
>       end-runs. However there are also good arguments for enabling the
>       testing of open source designs without having forcing the
>       implementation of a separate C model to test software. For example if
>       we hypothetically modelled a Pi Pico would it make sense to model the
>       PIO in C if we could just compile the Verilog for it into a SystemC
>       model? Would a plethora of closed device models be the inevitable
>       consequence of such an approach? Would it matter if we just
>       concentrated on supporting useful open source solutions?
> 
>       * Dynamic Machine Models
> 
>       While we try and avoid modelling bespoke virtual HW in QEMU
>       (virt/goldfish not withstanding ;-) there is obviously a desire in the
>       EDA space to allow such experimentation. Is this something we can
>       provide so aspiring HW engineers can experiment with system
>       architectures without having to form QEMU and learn QOM. There have
>       been suggestions about consuming device trees or maybe translating to
>       QMP calls and adding support for wiring devices together. Given the
>       number of forks that exist is this something that could be better
>       supported upstream without degenerating into messy hacks?
> 
>       * A sense of time
> 
>       Currently we have the fairly limited support for -icount in QEMU. At
>       the same time we have no desire to start expanding frontends with
>       the details cost models required for a more realistic sense of time to
>       be presented. One suggestion is to expand the TCG plugin interface to
>       allow for the plugin to control time allowing as much or little logic
>       to be pushed there as we like and freeing up frontends from ever having
>       to consider it.
> 
>     Are any of these topics of interest? Are there any other emulation
>     topics people would like to discuss?
> 
>     -- 
>     Alex Bennée
> 
> 
> 
> -- 
> Alex Bennée
> KVM/QEMU Hacker for Linaro