Re: [Qemu-devel] [PATCH v1 00/21] RISC-V QEMU Port Submission v1

[Richard W.M. Jones]

Just a few small points:

(1) I've built Fedora RPMs from this patch set [approximately - I'm
using a very slightly different / slightly older version, but it's not
substantively different]:

  http://copr-fe.cloud.fedoraproject.org/coprs/rjones/riscv/

It works well for me building plenty of Fedora packages over the past
few weeks, except for a possible Linux kernel bug which interacts with
the patch set not handling EBREAK correctly, which I had to patch
around (in the kernel):

  
https://groups.google.com/a/groups.riscv.org/forum/#!msg/sw-dev/v05FjcGC1EI/atXXUAcsCgAJ

(2) I'm worried that this patch starts off using virtio-mmio instead
of virtio-pci.  virtio-pci is better in every respect than
virtio-mmio, and while it may be a good interim solution I think we
need to have a plan to get rid of it eventually, and should make it
clear that virtio-mmio is not a permanent ABI so we don't get into the
same situation that we did with -M virt on ARM.

(3) poweroff doesn't work if you use -M virt (and hence don't use
HTIF).  I wrote a dummy poweroff device to get around this, but I
think it could raise a bigger point: Why bother to support HTIF
machine types at all?  The reason given in the patch set is because
spike (the cycle-accurate RISC-V emulator) only supports HTIF but
is that a good reason?

Rich.

-- 
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html