Re: [PATCH] KVM: selftests: Create memslot 0 at GPA 0x100000000 on x86_64

From: Nico Boehr <nrb@linux.ibm.com>
To: David Matlack <dmatlack@google.com>,
	Sean Christopherson <seanjc@google.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>,
	Fuad Tabba <tabba@google.com>, Peter Gonda <pgonda@google.com>,
	Ackerley Tng <ackerleytng@google.com>,
	Chao Peng <chao.p.peng@linux.intel.com>,
	Vishal Annapurve <vannapurve@google.com>,
	Michael Roth <michael.roth@amd.com>,
	Andrew Jones <ajones@ventanamicro.com>,
	kvm@vger.kernel.org
Subject: Re: [PATCH] KVM: selftests: Create memslot 0 at GPA 0x100000000 on x86_64
Date: Mon, 08 Apr 2024 15:47:02 +0200	[thread overview]
Message-ID: <171258402203.38901.9542700853299444599@t14-nrb> (raw)
In-Reply-To: <CALzav=fGUnYHiEc40Ym2Yh-H6wMRdw6biYj4+e1vZ0xmBDAnsg@mail.gmail.com>

Quoting David Matlack (2024-03-14 22:11:57)
[...]
> >   7. Use the PT_AT_4GIB knob in s390's CMMA test?  I suspect it does memslot
> >      shenanigans purely so that a low gfn (4096 in the test) is guaranteed to
> >      be available.
> 
> +Nico
> 
> Hm, if this test _needs_ to use GFN 4096, then maybe the framework can
> give tests two regions 0..KVM_FRAMEWORK_GPA and TEST_GPA..MAX.
> 
> If the test just needs any GFN then it can use TEST_GPA instead of
> 4096 << page_shift.

Sorry for the late reply, this got burried at the bottom of my inbox :(

The test creates two memslots with a gap in between to be able to test that
gaps in memslots are correctly skipped by the CMMA-related ioctls.

The test doesn't need GFN 4096. It should not matter where TEST_GPA is, as
long as there's a gap after the main memslot.

Feel free to include me on a follow-up, then I can test and look at this.