Received-SPF: pass (google.com: domain of linux-crypto-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 20 Apr 2021 11:47:13 +0200
From:   Borislav Petkov <bp@alien8.de>
To:     Brijesh Singh <brijesh.singh@amd.com>
Cc:     linux-kernel@vger.kernel.org, x86@kernel.org, kvm@vger.kernel.org,
        linux-crypto@vger.kernel.org, ak@linux.intel.com,
        herbert@gondor.apana.org.au, Thomas Gleixner <tglx@linutronix.de>,
        Ingo Molnar <mingo@redhat.com>, Joerg Roedel <jroedel@suse.de>,
        "H. Peter Anvin" <hpa@zytor.com>, Tony Luck <tony.luck@intel.com>,
        Dave Hansen <dave.hansen@intel.com>,
        "Peter Zijlstra (Intel)" <peterz@infradead.org>,
        Paolo Bonzini <pbonzini@redhat.com>,
        Tom Lendacky <thomas.lendacky@amd.com>,
        David Rientjes <rientjes@google.com>,
        Sean Christopherson <seanjc@google.com>,
        Vlastimil Babka <vbabka@suse.cz>
Subject: Re: [RFC Part2 PATCH 04/30] x86/mm: split the physmap when adding
 the page in RMP table
Message-ID: <20210420094713.GD5029@zn.tnic>
References: <20210324170436.31843-1-brijesh.singh@amd.com>
 <20210324170436.31843-5-brijesh.singh@amd.com>
 <20210419123226.GC9093@zn.tnic>
 <befbe586-1c45-ebf7-709a-00150365e7ec@amd.com>
 <20210419165214.GF9093@zn.tnic>
 <30bff969-e8cf-a991-7660-054ea136855a@amd.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <30bff969-e8cf-a991-7660-054ea136855a@amd.com>
Precedence: bulk

On Mon, Apr 19, 2021 at 12:46:53PM -0500, Brijesh Singh wrote:
> - KVM calls  alloc_page() to allocate a VMSA page. The allocator returns
> 0xffffc80000200000 (PFN 0x200, page-level=2M). The VMSA page is private
> page so KVM will call RMPUPDATE to add the page as a private page in the
> RMP table. While adding the RMP entry the KVM will use the page level=4K.

Right, and *here* we split the 2M page on the *host* so that there's no
discrepancy between the host pagetable and the RMP. I guess your patch
does exactly that. :)

And AFAIR, set_memory.c doesn't have the functionality to coalesce 4K
pages back into the corresponding 2M page. Which means, the RMP table is
in sync, more or less.

Thx and thanks for elaborating.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette