Message-ID: <50575db8-efb8-2d56-5dd9-fe4318db2af3@redhat.com>
Date:   Tue, 21 Feb 2023 08:45:56 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.7.1
Subject: Re: [PATCH v2 27/29] LoongArch: KVM: Implement vcpu world switch
Content-Language: en-US
To:     Tianrui Zhao <zhaotianrui@loongson.cn>
Cc:     Huacai Chen <chenhuacai@kernel.org>,
        WANG Xuerui <kernel@xen0n.name>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        loongarch@lists.linux.dev, linux-kernel@vger.kernel.org,
        kvm@vger.kernel.org, Jens Axboe <axboe@kernel.dk>,
        Mark Brown <broonie@kernel.org>,
        Alex Deucher <alexander.deucher@amd.com>,
        Oliver Upton <oliver.upton@linux.dev>, maobibo@loongson.cn
References: <20230220065735.1282809-1-zhaotianrui@loongson.cn>
 <20230220065735.1282809-28-zhaotianrui@loongson.cn>
From:   Paolo Bonzini <pbonzini@redhat.com>
In-Reply-To: <20230220065735.1282809-28-zhaotianrui@loongson.cn>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2/20/23 07:57, Tianrui Zhao wrote:
> +	/* Load Guest gprs */
> +	ld.d    $r1,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 1)
> +	ld.d    $r2,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 2)
> +	ld.d    $r3,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 3)
> +	ld.d    $r4,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 4)
> +	ld.d    $r5,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 5)
> +	ld.d    $r7,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 7)
> +	ld.d    $r8,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 8)
> +	ld.d    $r9,   \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 9)
> +	ld.d    $r10,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 10)
> +	ld.d    $r11,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 11)
> +	ld.d    $r12,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 12)
> +	ld.d    $r13,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 13)
> +	ld.d    $r14,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 14)
> +	ld.d    $r15,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 15)
> +	ld.d    $r16,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 16)
> +	ld.d    $r17,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 17)
> +	ld.d    $r18,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 18)
> +	ld.d    $r19,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 19)
> +	ld.d    $r20,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 20)
> +	ld.d    $r21,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 21)
> +	ld.d    $r22,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 22)
> +	ld.d    $r23,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 23)
> +	ld.d    $r24,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 24)
> +	ld.d    $r25,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 25)
> +	ld.d    $r26,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 26)
> +	ld.d    $r27,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 27)
> +	ld.d    $r28,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 28)
> +	ld.d    $r29,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 29)
> +	ld.d    $r30,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 30)
> +	ld.d    $r31,  \KVM_ARCH,  (KVM_ARCH_GGPR + 8 * 31)
> +	/* Load KVM_ARCH register */
> +	ld.d	\KVM_ARCH, \KVM_ARCH, (KVM_ARCH_GGPR + 8 * \GPRNUM)

This in practice relies on KVM_ARCH being a2 so please remove the 
KVM_ARCH and GPRNUM arguments from the macro; just replace \KVM_ARCH 
with a2 as needed.

Also, in these ld.d and st.d sequences you may want to use the ABI names 
instead of the rNN names, so it's clearer that you are skipping the 
KVM_ARCH register.

Paolo