Re: [PATCH v8 2/8] KVM: Extend the memslot to support fd-based private memory

[Jarkko Sakkinen]

On Mon, Oct 10, 2022 at 04:25:07PM +0800, Chao Peng wrote:
> On Sat, Oct 08, 2022 at 08:35:47PM +0300, Jarkko Sakkinen wrote:
> > On Sat, Oct 08, 2022 at 07:15:17PM +0300, Jarkko Sakkinen wrote:
> > > On Sat, Oct 08, 2022 at 12:54:32AM +0300, Jarkko Sakkinen wrote:
> > > > On Fri, Oct 07, 2022 at 02:58:54PM +0000, Sean Christopherson wrote:
> > > > > On Fri, Oct 07, 2022, Jarkko Sakkinen wrote:
> > > > > > On Thu, Oct 06, 2022 at 03:34:58PM +0000, Sean Christopherson wrote:
> > > > > > > On Thu, Oct 06, 2022, Jarkko Sakkinen wrote:
> > > > > > > > On Thu, Oct 06, 2022 at 05:58:03PM +0300, Jarkko Sakkinen wrote:
> > > > > > > > > On Thu, Sep 15, 2022 at 10:29:07PM +0800, Chao Peng wrote:
> > > > > > > > > > This new extension, indicated by the new flag KVM_MEM_PRIVATE, adds two
> > > > > > > > > > additional KVM memslot fields private_fd/private_offset to allow
> > > > > > > > > > userspace to specify that guest private memory provided from the
> > > > > > > > > > private_fd and guest_phys_addr mapped at the private_offset of the
> > > > > > > > > > private_fd, spanning a range of memory_size.
> > > > > > > > > > 
> > > > > > > > > > The extended memslot can still have the userspace_addr(hva). When use, a
> > > > > > > > > > single memslot can maintain both private memory through private
> > > > > > > > > > fd(private_fd/private_offset) and shared memory through
> > > > > > > > > > hva(userspace_addr). Whether the private or shared part is visible to
> > > > > > > > > > guest is maintained by other KVM code.
> > > > > > > > > 
> > > > > > > > > What is anyway the appeal of private_offset field, instead of having just
> > > > > > > > > 1:1 association between regions and files, i.e. one memfd per region?
> > > > > > > 
> > > > > > > Modifying memslots is slow, both in KVM and in QEMU (not sure about Google's VMM).
> > > > > > > E.g. if a vCPU converts a single page, it will be forced to wait until all other
> > > > > > > vCPUs drop SRCU, which can have severe latency spikes, e.g. if KVM is faulting in
> > > > > > > memory.  KVM's memslot updates also hold a mutex for the entire duration of the
> > > > > > > update, i.e. conversions on different vCPUs would be fully serialized, exacerbating
> > > > > > > the SRCU problem.
> > > > > > > 
> > > > > > > KVM also has historical baggage where it "needs" to zap _all_ SPTEs when any
> > > > > > > memslot is deleted.
> > > > > > > 
> > > > > > > Taking both a private_fd and a shared userspace address allows userspace to convert
> > > > > > > between private and shared without having to manipulate memslots.
> > > > > > 
> > > > > > Right, this was really good explanation, thank you.
> > > > > > 
> > > > > > Still wondering could this possibly work (or not):
> > > > > > 
> > > > > > 1. Union userspace_addr and private_fd.
> > > > > 
> > > > > No, because userspace needs to be able to provide both userspace_addr (shared
> > > > > memory) and private_fd (private memory) for a single memslot.
> > > > 
> > > > Got it, thanks for clearing my misunderstandings on this topic, and it
> > > > is quite obviously visible in 5/8 and 7/8. I.e. if I got it right,
> > > > memblock can be partially private, and you dig the shared holes with
> > > > KVM_MEMORY_ENCRYPT_UNREG_REGION. We have (in Enarx) ATM have memblock
> > > > per host mmap, I was looking into this dilated by that mindset but makes
> > > > definitely sense to support that.
> > > 
> > > For me the most useful reference with this feature is kvm_set_phys_mem()
> > > implementation in privmem-v8 branch. Took while to find it because I did
> > > not have much experience with QEMU code base. I'd even recommend to mention
> > > that function in the cover letter because it is really good reference on
> > > how this feature is supposed to be used.
> 
> That's a good point, I can mention that if people find useful. 

Yeah, I did implementation for Enarx (https://www.enarx.dev/) using just
that part as a reference. It has all the essentials what you need to
consider when you are already using KVM API, and want to add private
regions.

BR, Jarkko