Re: [Qemu-devel] [PATCH v3 17/22] virtio-pmem: prototype

[Pankaj Gupta]

> On 24/09/2018 07:45, David Gibson wrote:
> > On Thu, Sep 20, 2018 at 12:32:38PM +0200, David Hildenbrand wrote:
> >> From: Pankaj Gupta <address@hidden>
> >>
> >> This is the current protoype of virtio-pmem. Support will require
> >> machine changes for the architectures that will support it, so it will
> >> not yet be compiled.
> >>
> >> Signed-off-by: Pankaj Gupta <address@hidden>
> >> [ MemoryDevice/MemoryRegion changes, cleanups, addr property "memaddr",
> >>   split up patches ]
> >> Signed-off-by: David Hildenbrand <address@hidden>
> > 
> > Reviewed-by: David Gibson <address@hidden>
> > 
> > Seems generally sane.  Is there a (craft?) virtio-pmem spec around to
> > see what this is actually trying to implement?
> 
> I am not aware of any. The first goal is to get something implemented
> that works and doesn't break important concepts. E.g. hotplugging memory
> devices was one of these concepts that Igor saw as a potential problem
> for getting virtio-pmem upstream.
> 
> But maybe Pankaj already has some draft version of a spec lying around.

I have created document but no official spec. Project idea is shared here [1] 
and kernel part of series [2]. I will create a spec and share.

[1] https://www.spinics.net/lists/kvm/msg153095.html
[2] https://marc.info/?l=kvm&m=153572224719212&w=2

Thanks,
Pankaj
 
> 
> > 
> > And one nit..
> > 
> > [snip]
> >> +static int worker_cb(void *opaque)
> >> +{
> >> +    VirtIODeviceRequest *req = opaque;
> >> +    int err = 0;
> >> +
> >> +    printf("\n performing flush ...");
> > 
> > .. I assume the plan is to remove these debugging prints..
> > 
> 
> Yes, these are still in place for verifying that the machinery works.
> 
> --
> 
> Thanks,
> 
> David / dhildenb
> 
>