Re: [RFC v2 0/4] Introduce MMIO/PIO dispatch file descriptors (ioregionfd)

[Stefan Hajnoczi <stefanha@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1227 bytes --]

On Thu, Jan 28, 2021 at 09:32:19PM +0300, Elena Afanasova wrote:
> This patchset introduces a KVM dispatch mechanism which can be used 
> for handling MMIO/PIO accesses over file descriptors without returning 
> from ioctl(KVM_RUN). This allows device emulation to run in another task 
> separate from the vCPU task.
> 
> This is achieved through KVM vm ioctl for registering MMIO/PIO regions and 
> a wire protocol that KVM uses to communicate with a task handling an 
> MMIO/PIO access.
> 
> TODOs:
> * Implement KVM_EXIT_IOREGIONFD_FAILURE
> * Add non-x86 arch support

This is an interesting topic for discussion with the KVM maintainers.
The ioctl(KVM_RUN) exit code is arch-specific in the sense that there is
no standard approach for PIO/MMIO accesses to return to userspace and
resume processing when ioctl(KVM_RUN) is called again.

This RFC series is x86-specific, but part of it can be made cross-arch
by introducing a core KVM ->complete_user_exit() function pointer that
MMIO/PIO and other users (EINTR?) can use to resume processing when
ioctl(KVM_RUN) is called again.

Maybe the benefit for a generic ->complete_user_exit() function is too
small so each arch open codes this behavior?

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]