Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221213125538.81209-1-jiangshanlai@gmail.com>
In-Reply-To: <20221213125538.81209-1-jiangshanlai@gmail.com>
From:   Lai Jiangshan <jiangshanlai@gmail.com>
Date:   Fri, 6 Jan 2023 09:45:37 +0800
Message-ID: <CAJhGHyDr6WH+yREd4=ua--c1BA7s5HNHLGJoPSSWAOAwECCZ+g@mail.gmail.com>
Subject: Re: [PATCH] kvm: x86/mmu: Remove FNAME(is_self_change_mapping)
To:     linux-kernel@vger.kernel.org
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        Sean Christopherson <seanjc@google.com>,
        Lai Jiangshan <jiangshan.ljs@antgroup.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
        Dave Hansen <dave.hansen@linux.intel.com>, x86@kernel.org,
        "H. Peter Anvin" <hpa@zytor.com>, kvm@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Dec 13, 2022 at 8:54 PM Lai Jiangshan <jiangshanlai@gmail.com> wrote:
>
> From: Lai Jiangshan <jiangshan.ljs@antgroup.com>
>
> FNAME(is_self_change_mapping) has two functionalities.
>
>   If the fault is on a huge page but at least one of the pagetable on
>   the walk is also on the terminal huge page, disable the huge page
>   mapping for the fault.
>
>   If the fault is modifying at least one of the pagetable on the walk,
>   set something to tell the emulator.
>
> The first functionality is much better handled by kvm_mmu_hugepage_adjust()
> now, and it has a defect that it blindly disables the huge page mapping
> rather than trying to reduce the size of the huge page first.
>
> Huang Hang reported that when a guest is writing to a 1G page, but
> only a 4K page is mapped because of the first functionality in a case
> in which we think a 2M page should be mapped.  The 1G page includes
> a pagetable on the pagetable-walk, but the narrowed 2M page doesn't.
>
> To fix the problem, remove FNAME(is_self_change_mapping) for its first
> functionality is already and better handled by kvm_mmu_hugepage_adjust(),
> and re-implement the second functionality in FNAME(fetch).
>
> Reported-by: Huang Hang <hhuang@linux.alibaba.com>
> Signed-off-by: Lai Jiangshan <jiangshan.ljs@antgroup.com>
> ---


Hello,

Ping.

Thanks

Lai