Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
Date:   Wed, 8 Jun 2022 16:01:38 +0000
From:   Sean Christopherson <seanjc@google.com>
To:     Andrew Jones <drjones@redhat.com>
Cc:     Paolo Bonzini <pbonzini@redhat.com>, kvm@vger.kernel.org,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        David Matlack <dmatlack@google.com>,
        Ben Gardon <bgardon@google.com>,
        Oliver Upton <oupton@google.com>, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 048/144] KVM: selftests: Rename 'struct vcpu' to
 'struct kvm_vcpu'
Message-ID: <YqDH4m0TxLcK5Brw@google.com>
References: <20220603004331.1523888-1-seanjc@google.com>
 <20220603004331.1523888-49-seanjc@google.com>
 <20220608151815.7mwlj3eppwmujaow@gator>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220608151815.7mwlj3eppwmujaow@gator>
Precedence: bulk

On Wed, Jun 08, 2022, Andrew Jones wrote:
> On Fri, Jun 03, 2022 at 12:41:55AM +0000, Sean Christopherson wrote:
> > Rename 'struct vcpu' to 'struct kvm_vcpu' to align with 'struct kvm_vm'
> > in the selftest, and to give readers a hint that the struct is specific
> > to KVM.
> 
> I'm not completely sold on this change. I don't mind that the selftest
> vcpu struct isn't named the same as the KVM vcpu struct, since they're
> different structs.

I don't care about about matching KVM's internal naming exactly, but I do care
about not having a bare "vcpu", it makes searching for usage a pain because it's
impossible to differentiate between instances of the struct and variables of the
same name without additional qualifiers.

> I also don't mind avoiding 'kvm_' prefixes in "KVM selftests" (indeed I
> wonder if we really need the kvm_ prefix for the vm struct).

Same as above, "struct vm *vm" will drive me bonkers :-)

> If we do need prefixes for the kvm selftest framework code to avoid
> collisions with test code, then maybe we should invent something else, rather
> than use the somewhat ambiguous 'kvm', which could also collide with stuff in
> the kvm uapi.

Potential collisions with the KVM uAPI is a feature of sorts, e.g. tests shouldn't
be redefining kvm_* structures (I'd prefer _tests_ not use kvm_* at all, and only
use kvm_* in the library), and I gotta imagine KVM would break at least one real
world userspace if it defined "kvm_vcpu".

That said, I don't have a super strong preference for kvm_ versus something else,
though I think it will be difficult to come up with something that's unique,
intuitive, and doesn't look like a typo.