Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20190729115544.17895-1-anup.patel@wdc.com> <20190729115544.17895-5-anup.patel@wdc.com>
 <ade614ae-fcfe-35f2-0519-1df71d035bcd@redhat.com> <2de10efc-56f8-ff47-ed69-7e471a099c80@redhat.com>
In-Reply-To: <2de10efc-56f8-ff47-ed69-7e471a099c80@redhat.com>
From:   Anup Patel <anup@brainfault.org>
Date:   Tue, 30 Jul 2019 17:15:53 +0530
Message-ID: <CAAhSdy0OH9h-R=2NxhhPs6jmFPNgZVSwFtCjtJrf++htu82ifA@mail.gmail.com>
Subject: Re: [RFC PATCH 04/16] RISC-V: KVM: Implement VCPU create, init and
 destroy functions
To:     Paolo Bonzini <pbonzini@redhat.com>
Cc:     Anup Patel <Anup.Patel@wdc.com>,
        Palmer Dabbelt <palmer@sifive.com>,
        Paul Walmsley <paul.walmsley@sifive.com>,
        Radim K <rkrcmar@redhat.com>,
        Daniel Lezcano <daniel.lezcano@linaro.org>,
        Thomas Gleixner <tglx@linutronix.de>,
        Atish Patra <Atish.Patra@wdc.com>,
        Alistair Francis <Alistair.Francis@wdc.com>,
        Damien Le Moal <Damien.LeMoal@wdc.com>,
        Christoph Hellwig <hch@infradead.org>,
        "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
        "linux-riscv@lists.infradead.org" <linux-riscv@lists.infradead.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, Jul 30, 2019 at 3:46 PM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> On 30/07/19 10:48, Paolo Bonzini wrote:
> > On 29/07/19 13:56, Anup Patel wrote:
> >> +    cntx->hstatus |= HSTATUS_SP2V;
> >> +    cntx->hstatus |= HSTATUS_SP2P;
> > IIUC, cntx->hstatus's SP2P bit contains the guest's sstatus.SPP bit?
>
> Nevermind, that was also a bit confused.  The guest's sstatus.SPP is in
> vsstatus.  The pseudocode for V-mode switch is
>
> SRET:
>   V = hstatus.SPV (1)
>   MODE = sstatus.SPP
>   hstatus.SPV = hstatus.SP2V
>   sstatus.SPP = hstatus.SP2P
>   hstatus.SP2V = 0
>   hstatus.SP2P = 0
>   ...
>
> trap:
>   hstatus.SP2V = hstatus.SPV
>   hstatus.SP2P = sstatus.SPP
>   hstatus.SPV = V (1)
>   sstatus.SPP = MODE
>   V = 0
>   MODE = 1
>

Yes, this kind of pseudo-code are not explicitly specified in the
RISC-V spec. The RISC-V formal model is supposed to cover
this kind of detailed HW state transition.

> so:
>
> 1) indeed we need SP2V=SPV=1 when entering guest mode
>
> 2) sstatus.SPP contains the guest mode
>
> 3) SP2P doesn't really matter for KVM since it never goes to VS-mode
> from an interrupt handler, so if my reasoning is correct I'd leave it
> clear, but I guess it's up to you whether to set it or not.

Yes, SP2P does not matter but we set it to 1 here so that from Guest
perspective it seems we were in S-mode previously.

Regards,
Anup